diff --git a/.gitattributes b/.gitattributes index bce831a021e803bec9976b60d2ce4a79107f2f63..0ce4e929cd5569eec7c6ceabd189ab29a548de14 100644 --- a/.gitattributes +++ b/.gitattributes @@ -458,3 +458,7 @@ checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/check checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-600/tokenizer.json filter=lfs diff=lfs merge=lfs -text checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1500/tokenizer.json filter=lfs diff=lfs merge=lfs -text checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_control_10M_seed0/runs/checkpoint-1100/global_step1100/zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_control_10M_seed0/runs/checkpoint-1100/global_step1100/zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..e389793c3e5533dd5d964ad770d70adcb042404e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_control_10M_seed0/runs/checkpoint-1100/global_step1100/zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb5c2d7e79ae83a2bc8fb18c4b40d97196f8f3b91fbd9ddee5a8e0b2c9ae3d1d +size 25702104228 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_control_10M_seed0/runs/checkpoint-1100/global_step1100/zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_control_10M_seed0/runs/checkpoint-1100/global_step1100/zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..303e28cb3e4fec45ebc511788dabd03da4c05c1f --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_control_10M_seed0/runs/checkpoint-1100/global_step1100/zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e176b1bd82e450938018cde5ccabf8cce77b777879d8f704e1a2bf946ff0f3e9 +size 25702107428 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/global_step1300/zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/global_step1300/zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..20871b355a400f464f93c10d6c3a0380d227dc12 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/global_step1300/zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b15721d2bfac0f8542571e48dc0cc3b316b86d7958e9f9d480a413aba99d20db +size 25702107684 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/global_step1300/zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/global_step1300/zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..84c7b176aade9cb63b505f8380ddae24a35bca98 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/global_step1300/zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:36025bbffae89eb9872505337f5f27f9de5a0d57c728c4721477847d54a593e2 +size 25702107428 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/2d73a6863086ff9d491c28e49df9fb697cd92c2b.lock b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/2d73a6863086ff9d491c28e49df9fb697cd92c2b.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/4719a04514ec2f060240711b7c33ab21187cac730ecaba3040b7a0fd95a9cefb.lock b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/4719a04514ec2f060240711b7c33ab21187cac730ecaba3040b7a0fd95a9cefb.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/47d4a5aa69cdef91a53b77f5c5583647a578ca0e.lock b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/47d4a5aa69cdef91a53b77f5c5583647a578ca0e.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/584d8d3e3f82f7964955174dfe5e3b1cf117a9d859f022cfdf7fcb884856e002.lock b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/584d8d3e3f82f7964955174dfe5e3b1cf117a9d859f022cfdf7fcb884856e002.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58.lock b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/adapter_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/adapter_config.json new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b new file mode 100644 index 0000000000000000000000000000000000000000..2d73a6863086ff9d491c28e49df9fb697cd92c2b --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "eos_token_id": 128001, + "transformers_version": "4.45.0.dev0", + "do_sample": true, + "temperature": 0.6, + "top_p": 0.9 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e new file mode 100644 index 0000000000000000000000000000000000000000..47d4a5aa69cdef91a53b77f5c5583647a578ca0e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "bfloat16", + "transformers_version": "4.45.0.dev0", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 new file mode 100644 index 0000000000000000000000000000000000000000..d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 @@ -0,0 +1,261 @@ +{ + "metadata": { + "total_size": 6425499648 + }, + "weight_map": { + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main new file mode 100644 index 0000000000000000000000000000000000000000..940c46e96db631910d2cc5186ab76cd83e32b124 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main @@ -0,0 +1 @@ +13afe5124825b4f3751f836b40dafda64c1ed062 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json new file mode 100644 index 0000000000000000000000000000000000000000..47d4a5aa69cdef91a53b77f5c5583647a578ca0e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "bfloat16", + "transformers_version": "4.45.0.dev0", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..2d73a6863086ff9d491c28e49df9fb697cd92c2b --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "eos_token_id": 128001, + "transformers_version": "4.45.0.dev0", + "do_sample": true, + "temperature": 0.6, + "top_p": 0.9 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json @@ -0,0 +1,261 @@ +{ + "metadata": { + "total_size": 6425499648 + }, + "weight_map": { + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-100/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-100/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ad42cab050a2af79a90d7a808c8bb133d17148c0 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-100/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ebf769328bb9a45b3791a403a4b342a675e314e92b44467f975ab6804c080e51 +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1100/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1100/model-00002-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bf5d43a9e6ba87260ae9cb9ae8eb70e9051d92aa --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1100/model-00002-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:92841c92981f1640438d8224c9d859c05086492f80deb34151b6b8cc280c563d +size 2247734920 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..cbba424cb2109da26bffbb01960eb52b2dc47b0d --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f29ab0036e394c39267ad86bf247bf1558e2213add95383fce34a980f109b42e +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0c27d91076a4f01fc26da3f3eb57e643a5b17b77 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1300/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e484bef0de9dbc2a1854d7263015c073f363105ee8aeafa103cb8b7f04d74337 +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..466c494b53a5cf3030ac585b29d1563c49b789d4 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a5075060d7b0d18a80b592cad740c3383144e4b5ba0bc7e7b9ff8ebb2afad8e5 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..262d37671147896b5dfa4ab77f0e9578ea0d968c --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:46766e2820279a98976091f452887a7ad876d6c74fa04b8df6451b77352f4c07 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..b31f6e0c1e981c4c1dee9a4626e2a6c6ae5abb8e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:deab7068d9411eb964839c7901679767e090930bd8dfb3faa9e266fdf4d5aed9 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..4522370045cda1f3bb3b317cfdfc4dc52d354ce8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b68599908ccded14a10b45eb143dcdffa78e1ccb710aa016edad8d76a092d874 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0c27d91076a4f01fc26da3f3eb57e643a5b17b77 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1400/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e484bef0de9dbc2a1854d7263015c073f363105ee8aeafa103cb8b7f04d74337 +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1600/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1600/model-00002-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..76f3f96b45a9030bfa55a0a333e91c5440f4be48 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1600/model-00002-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:afab43052a5c2ce5a73b17bc25c85bdc6aa24f2ee11f05a645e88bde88d0ba8f +size 2247734920 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..01569c25f35e725b06596c870e65bcbb2dc1e047 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/trainer_state.json @@ -0,0 +1,19533 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.73972602739726, + "eval_steps": 10, + "global_step": 2500, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0501, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0238, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0346, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0108, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.131, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0488, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0905, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0493, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1038, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.0925064086914062, + "eval_runtime": 38.3718, + "eval_samples_per_second": 26.061, + "eval_steps_per_second": 1.095, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0804, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0053, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0607, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0539, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.420984745025635, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0941, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.420984745025635, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0249, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.420984745025635, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0391, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 12.758050918579102, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.0669, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 7.247162818908691, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0067, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 4.390945911407471, + "learning_rate": 7.299270072992701e-08, + "loss": 3.013, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 2.980123519897461, + "eval_runtime": 38.0353, + "eval_samples_per_second": 26.291, + "eval_steps_per_second": 1.104, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 5.657729625701904, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9558, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 4.256235122680664, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8243, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 3.3831233978271484, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8079, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 8.511351585388184, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8349, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 2.908385753631592, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8029, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.563490867614746, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.7748, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.2141411304473877, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.7669, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.7528369426727295, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.705, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.106175422668457, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.6777, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.909290075302124, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6251, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.7223422527313232, + "eval_runtime": 37.5889, + "eval_samples_per_second": 26.604, + "eval_steps_per_second": 1.117, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.222564458847046, + "learning_rate": 2.737226277372263e-07, + "loss": 2.6447, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.3698360919952393, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6186, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.0300192832946777, + "learning_rate": 3.102189781021898e-07, + "loss": 2.5635, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.126431941986084, + "learning_rate": 3.284671532846716e-07, + "loss": 2.5587, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 1.9875943660736084, + "learning_rate": 3.467153284671533e-07, + "loss": 2.567, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 1.9899641275405884, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.5784, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 2.323369026184082, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5282, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.0645694732666016, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.4822, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 2.0784924030303955, + "learning_rate": 4.197080291970803e-07, + "loss": 2.5627, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.2178752422332764, + "learning_rate": 4.379562043795621e-07, + "loss": 2.4601, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.5703861713409424, + "eval_runtime": 37.5777, + "eval_samples_per_second": 26.612, + "eval_steps_per_second": 1.118, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 2.0207555294036865, + "learning_rate": 4.562043795620438e-07, + "loss": 2.4722, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.3551583290100098, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5123, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.7226097583770752, + "learning_rate": 4.927007299270073e-07, + "loss": 2.4687, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.9984400272369385, + "learning_rate": 5.109489051094891e-07, + "loss": 2.4836, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.9084681272506714, + "learning_rate": 5.291970802919708e-07, + "loss": 2.456, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 2.1871354579925537, + "learning_rate": 5.474452554744526e-07, + "loss": 2.4154, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 2.275301933288574, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4036, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 2.1016786098480225, + "learning_rate": 5.839416058394161e-07, + "loss": 2.4243, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 2.1861119270324707, + "learning_rate": 6.021897810218979e-07, + "loss": 2.3234, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.9693622589111328, + "learning_rate": 6.204379562043796e-07, + "loss": 2.4753, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.4698286056518555, + "eval_runtime": 37.607, + "eval_samples_per_second": 26.591, + "eval_steps_per_second": 1.117, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 2.0734782218933105, + "learning_rate": 6.386861313868614e-07, + "loss": 2.3517, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 2.243452787399292, + "learning_rate": 6.569343065693432e-07, + "loss": 2.3887, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 2.584733486175537, + "learning_rate": 6.751824817518248e-07, + "loss": 2.3081, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.8046469688415527, + "learning_rate": 6.934306569343066e-07, + "loss": 2.3254, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 2.988283395767212, + "learning_rate": 7.116788321167883e-07, + "loss": 2.3335, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 2.5031774044036865, + "learning_rate": 7.299270072992701e-07, + "loss": 2.3519, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 2.251530408859253, + "learning_rate": 7.48175182481752e-07, + "loss": 2.3549, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 3.031409740447998, + "learning_rate": 7.664233576642337e-07, + "loss": 2.3339, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.9044957160949707, + "learning_rate": 7.846715328467154e-07, + "loss": 2.2923, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 2.586066484451294, + "learning_rate": 8.029197080291971e-07, + "loss": 2.2626, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.4082915782928467, + "eval_runtime": 37.5901, + "eval_samples_per_second": 26.603, + "eval_steps_per_second": 1.117, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 2.358807325363159, + "learning_rate": 8.211678832116789e-07, + "loss": 2.2692, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 2.388153314590454, + "learning_rate": 8.394160583941606e-07, + "loss": 2.2921, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 2.4358012676239014, + "learning_rate": 8.576642335766425e-07, + "loss": 2.304, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 2.1953372955322266, + "learning_rate": 8.759124087591242e-07, + "loss": 2.3179, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 2.904700994491577, + "learning_rate": 8.94160583941606e-07, + "loss": 2.2718, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.9871703386306763, + "learning_rate": 9.124087591240876e-07, + "loss": 2.2855, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 2.612363815307617, + "learning_rate": 9.306569343065694e-07, + "loss": 2.2709, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8143366575241089, + "learning_rate": 9.489051094890511e-07, + "loss": 2.3218, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.739389657974243, + "learning_rate": 9.67153284671533e-07, + "loss": 2.2681, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 2.006971836090088, + "learning_rate": 9.854014598540146e-07, + "loss": 2.2215, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.364126443862915, + "eval_runtime": 37.5857, + "eval_samples_per_second": 26.606, + "eval_steps_per_second": 1.117, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 2.215372323989868, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.3008, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 2.1057345867156982, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.1944, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.524104356765747, + "learning_rate": 1.04014598540146e-06, + "loss": 2.1942, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.0156443119049072, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.2043, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.118178606033325, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.2582, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 2.201267957687378, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.2621, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 2.098020076751709, + "learning_rate": 1.113138686131387e-06, + "loss": 2.2073, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 2.015281915664673, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.1799, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 2.1951286792755127, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.201, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8758258819580078, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.1996, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.332369565963745, + "eval_runtime": 37.5585, + "eval_samples_per_second": 26.625, + "eval_steps_per_second": 1.118, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 2.1257612705230713, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.212, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 2.3016891479492188, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.2549, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.9585005044937134, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.2066, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.3478572368621826, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.2265, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 1.8603413105010986, + "learning_rate": 1.259124087591241e-06, + "loss": 2.2154, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 2.222263813018799, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.1734, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9255578517913818, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.1999, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.921866774559021, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.1363, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.198625087738037, + "learning_rate": 1.332116788321168e-06, + "loss": 2.1825, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.944101333618164, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.1068, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.3024067878723145, + "eval_runtime": 37.5703, + "eval_samples_per_second": 26.617, + "eval_steps_per_second": 1.118, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 1.9478638172149658, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.2281, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8018860816955566, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.1809, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 1.9129096269607544, + "learning_rate": 1.405109489051095e-06, + "loss": 2.2166, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.8421601057052612, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.1688, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 1.9687496423721313, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.1541, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 1.833653450012207, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.23, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.8790069818496704, + "learning_rate": 1.478102189781022e-06, + "loss": 2.1938, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9399162530899048, + "learning_rate": 1.496350364963504e-06, + "loss": 2.2153, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.9903068542480469, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.1774, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 1.911264419555664, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.1287, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.288886308670044, + "eval_runtime": 37.5336, + "eval_samples_per_second": 26.643, + "eval_steps_per_second": 1.119, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9418243169784546, + "learning_rate": 1.551094890510949e-06, + "loss": 2.1267, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.270538568496704, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.1968, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 1.8941543102264404, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.2536, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.137468099594116, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.1629, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 1.9722516536712646, + "learning_rate": 1.624087591240876e-06, + "loss": 2.1064, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 1.8995206356048584, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.153, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.305126190185547, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.1732, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.9987198114395142, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.0946, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.2726221084594727, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.1543, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 2.1480541229248047, + "learning_rate": 1.715328467153285e-06, + "loss": 2.1618, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.273084878921509, + "eval_runtime": 37.0539, + "eval_samples_per_second": 26.988, + "eval_steps_per_second": 1.133, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9935969114303589, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.1888, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 2.442995071411133, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.1926, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 2.1095662117004395, + "learning_rate": 1.77007299270073e-06, + "loss": 2.229, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 2.2200515270233154, + "learning_rate": 1.788321167883212e-06, + "loss": 2.1683, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 2.148380994796753, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.1051, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.2374966144561768, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.1413, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.8815765380859375, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.1433, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.1499204635620117, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.1392, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.9436023235321045, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.1387, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 2.125568389892578, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.1681, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.2652690410614014, + "eval_runtime": 37.4032, + "eval_samples_per_second": 26.736, + "eval_steps_per_second": 1.123, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 2.500214099884033, + "learning_rate": 1.916058394160584e-06, + "loss": 2.073, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 2.260161876678467, + "learning_rate": 1.934306569343066e-06, + "loss": 2.1191, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 2.1456682682037354, + "learning_rate": 1.952554744525548e-06, + "loss": 2.1296, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 2.1852567195892334, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.1241, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 2.0373728275299072, + "learning_rate": 1.989051094890511e-06, + "loss": 2.1846, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 2.1237425804138184, + "learning_rate": 2.007299270072993e-06, + "loss": 2.1257, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 2.020529270172119, + "learning_rate": 2.025547445255475e-06, + "loss": 2.133, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.9947962760925293, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.1288, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 2.196021556854248, + "learning_rate": 2.062043795620438e-06, + "loss": 2.1954, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.8369503021240234, + "learning_rate": 2.08029197080292e-06, + "loss": 2.0767, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.2477104663848877, + "eval_runtime": 37.5191, + "eval_samples_per_second": 26.653, + "eval_steps_per_second": 1.119, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.973464846611023, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.1511, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.1027629375457764, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.1781, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.943640112876892, + "learning_rate": 2.135036496350365e-06, + "loss": 2.0813, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.980432391166687, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.099, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9553724527359009, + "learning_rate": 2.171532846715329e-06, + "loss": 2.1445, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.85471773147583, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.111, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8751261234283447, + "learning_rate": 2.208029197080292e-06, + "loss": 2.1227, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 2.027106523513794, + "learning_rate": 2.226277372262774e-06, + "loss": 2.1508, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.9612401723861694, + "learning_rate": 2.244525547445256e-06, + "loss": 2.1396, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.8721307516098022, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.1068, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.23884916305542, + "eval_runtime": 37.5607, + "eval_samples_per_second": 26.624, + "eval_steps_per_second": 1.118, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.173870801925659, + "learning_rate": 2.281021897810219e-06, + "loss": 2.0908, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.0046143531799316, + "learning_rate": 2.299270072992701e-06, + "loss": 2.117, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.9593181610107422, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.078, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7910178899765015, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.1099, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.9955250024795532, + "learning_rate": 2.354014598540146e-06, + "loss": 2.1319, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.9732416868209839, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.0583, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 2.2180569171905518, + "learning_rate": 2.39051094890511e-06, + "loss": 2.0848, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 1.9734926223754883, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.0876, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.9308476448059082, + "learning_rate": 2.427007299270073e-06, + "loss": 2.143, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 2.032986640930176, + "learning_rate": 2.445255474452555e-06, + "loss": 2.0522, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.2264482975006104, + "eval_runtime": 37.579, + "eval_samples_per_second": 26.611, + "eval_steps_per_second": 1.118, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 1.981833815574646, + "learning_rate": 2.463503649635037e-06, + "loss": 2.028, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.8561030626296997, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.1475, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 1.9615944623947144, + "learning_rate": 2.5e-06, + "loss": 2.0456, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 1.983773112297058, + "learning_rate": 2.518248175182482e-06, + "loss": 2.0531, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 1.9475226402282715, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.1072, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.040128707885742, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.0875, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 1.9370408058166504, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.0855, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.942474365234375, + "learning_rate": 2.591240875912409e-06, + "loss": 2.1098, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.035705089569092, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.1195, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.062436580657959, + "learning_rate": 2.627737226277373e-06, + "loss": 2.096, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.217604637145996, + "eval_runtime": 37.58, + "eval_samples_per_second": 26.61, + "eval_steps_per_second": 1.118, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.9673733711242676, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.0682, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 2.315821409225464, + "learning_rate": 2.664233576642336e-06, + "loss": 2.0944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.136030912399292, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.0678, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0755345821380615, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.1164, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 2.417750597000122, + "learning_rate": 2.718978102189781e-06, + "loss": 2.1133, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.046565532684326, + "learning_rate": 2.737226277372263e-06, + "loss": 2.0582, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.3266139030456543, + "learning_rate": 2.755474452554745e-06, + "loss": 2.1682, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.272815704345703, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.1198, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.091608762741089, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.0509, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 2.2504429817199707, + "learning_rate": 2.81021897810219e-06, + "loss": 2.0983, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.2180068492889404, + "eval_runtime": 37.5842, + "eval_samples_per_second": 26.607, + "eval_steps_per_second": 1.117, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.090161085128784, + "learning_rate": 2.828467153284672e-06, + "loss": 2.0169, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.9184086322784424, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.0678, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 2.148932933807373, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.0242, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 1.8502088785171509, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.115, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 2.200047016143799, + "learning_rate": 2.901459854014599e-06, + "loss": 2.105, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 1.8261909484863281, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.079, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8948200941085815, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.1259, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 2.0827932357788086, + "learning_rate": 2.956204379562044e-06, + "loss": 2.0964, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.91581392288208, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.0353, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.151932716369629, + "learning_rate": 2.992700729927008e-06, + "loss": 2.0361, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.206122875213623, + "eval_runtime": 37.5989, + "eval_samples_per_second": 26.597, + "eval_steps_per_second": 1.117, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.010537624359131, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.1117, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 2.162351608276367, + "learning_rate": 3.029197080291971e-06, + "loss": 2.11, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.8603689670562744, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.0921, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 2.0103940963745117, + "learning_rate": 3.065693430656935e-06, + "loss": 2.1099, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0573348999023438, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.1019, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.868481993675232, + "learning_rate": 3.102189781021898e-06, + "loss": 2.0255, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 2.1905722618103027, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.0149, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.0946364402770996, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.1231, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 2.2042884826660156, + "learning_rate": 3.156934306569343e-06, + "loss": 2.0749, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 1.900771975517273, + "learning_rate": 3.175182481751825e-06, + "loss": 2.0534, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.1958627700805664, + "eval_runtime": 37.5954, + "eval_samples_per_second": 26.599, + "eval_steps_per_second": 1.117, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 2.010354995727539, + "learning_rate": 3.193430656934307e-06, + "loss": 2.0711, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.057535409927368, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.0559, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 2.153372287750244, + "learning_rate": 3.2299270072992707e-06, + "loss": 1.9923, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.926892638206482, + "learning_rate": 3.248175182481752e-06, + "loss": 2.0876, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.8907747268676758, + "learning_rate": 3.266423357664234e-06, + "loss": 2.0641, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.9739303588867188, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.0867, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.081756830215454, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.0655, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 2.034151554107666, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.0682, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8720734119415283, + "learning_rate": 3.339416058394161e-06, + "loss": 2.064, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 2.07344126701355, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.0695, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.1921699047088623, + "eval_runtime": 37.6118, + "eval_samples_per_second": 26.587, + "eval_steps_per_second": 1.117, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 2.0494537353515625, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.0668, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 2.058220863342285, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.0716, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9059336185455322, + "learning_rate": 3.412408759124088e-06, + "loss": 2.0062, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.9359859228134155, + "learning_rate": 3.43065693430657e-06, + "loss": 2.0581, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.9826178550720215, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.0127, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.7971314191818237, + "learning_rate": 3.467153284671533e-06, + "loss": 2.0487, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7797099351882935, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.0448, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 2.057704448699951, + "learning_rate": 3.503649635036497e-06, + "loss": 2.0772, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 2.0499088764190674, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.0894, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.907564640045166, + "learning_rate": 3.54014598540146e-06, + "loss": 2.1403, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.1906583309173584, + "eval_runtime": 37.1067, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.848954677581787, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.1259, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 2.1200220584869385, + "learning_rate": 3.576642335766424e-06, + "loss": 1.9716, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9336203336715698, + "learning_rate": 3.5948905109489054e-06, + "loss": 1.9726, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 2.119415283203125, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.0123, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9229284524917603, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.0335, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 2.0724565982818604, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.0099, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 2.315553903579712, + "learning_rate": 3.667883211678833e-06, + "loss": 2.0053, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 2.002094268798828, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.0997, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 2.5517187118530273, + "learning_rate": 3.704379562043796e-06, + "loss": 2.066, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 2.189847946166992, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.0107, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.1816186904907227, + "eval_runtime": 37.5023, + "eval_samples_per_second": 26.665, + "eval_steps_per_second": 1.12, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 2.1692676544189453, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.048, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 2.105806589126587, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.0346, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 2.0867691040039062, + "learning_rate": 3.777372262773723e-06, + "loss": 2.016, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 2.0230326652526855, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.0626, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 2.2379798889160156, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.0226, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.8058878183364868, + "learning_rate": 3.832116788321168e-06, + "loss": 1.9884, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 2.065242290496826, + "learning_rate": 3.85036496350365e-06, + "loss": 2.0263, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 2.1474945545196533, + "learning_rate": 3.868613138686132e-06, + "loss": 2.0431, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.8931554555892944, + "learning_rate": 3.886861313868613e-06, + "loss": 2.0767, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 2.1712071895599365, + "learning_rate": 3.905109489051096e-06, + "loss": 2.026, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.17148494720459, + "eval_runtime": 37.6313, + "eval_samples_per_second": 26.574, + "eval_steps_per_second": 1.116, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.9246273040771484, + "learning_rate": 3.923357664233577e-06, + "loss": 2.0875, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.9903236627578735, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.1101, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.805240511894226, + "learning_rate": 3.95985401459854e-06, + "loss": 1.977, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.877760648727417, + "learning_rate": 3.978102189781022e-06, + "loss": 2.0619, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.7705204486846924, + "learning_rate": 3.996350364963504e-06, + "loss": 1.9576, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.7442058324813843, + "learning_rate": 4.014598540145986e-06, + "loss": 2.0199, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.8216336965560913, + "learning_rate": 4.032846715328467e-06, + "loss": 2.0451, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.9217220544815063, + "learning_rate": 4.05109489051095e-06, + "loss": 1.9983, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 1.7437387704849243, + "learning_rate": 4.069343065693431e-06, + "loss": 2.0227, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.7862118482589722, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.1007, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.167337656021118, + "eval_runtime": 37.6184, + "eval_samples_per_second": 26.583, + "eval_steps_per_second": 1.116, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 1.9299460649490356, + "learning_rate": 4.105839416058395e-06, + "loss": 2.0198, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.9095406532287598, + "learning_rate": 4.124087591240876e-06, + "loss": 2.0116, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 1.9675389528274536, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.0801, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.867546796798706, + "learning_rate": 4.16058394160584e-06, + "loss": 1.9584, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9578276872634888, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.0528, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.9662718772888184, + "learning_rate": 4.197080291970803e-06, + "loss": 1.9995, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 1.8949064016342163, + "learning_rate": 4.215328467153285e-06, + "loss": 2.0053, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 2.1320366859436035, + "learning_rate": 4.233576642335767e-06, + "loss": 2.0366, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.100839853286743, + "learning_rate": 4.251824817518249e-06, + "loss": 1.999, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.9846616983413696, + "learning_rate": 4.27007299270073e-06, + "loss": 1.9992, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.164961099624634, + "eval_runtime": 37.6477, + "eval_samples_per_second": 26.562, + "eval_steps_per_second": 1.116, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.8623476028442383, + "learning_rate": 4.288321167883212e-06, + "loss": 1.9679, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 2.4636693000793457, + "learning_rate": 4.306569343065693e-06, + "loss": 2.0329, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 1.9858465194702148, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.0281, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 2.143198013305664, + "learning_rate": 4.343065693430658e-06, + "loss": 2.0124, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.021104574203491, + "learning_rate": 4.361313868613139e-06, + "loss": 2.0541, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.126469135284424, + "learning_rate": 4.379562043795621e-06, + "loss": 2.0377, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.0688600540161133, + "learning_rate": 4.397810218978102e-06, + "loss": 2.0148, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.9566450119018555, + "learning_rate": 4.416058394160584e-06, + "loss": 1.9907, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 1.9788639545440674, + "learning_rate": 4.434306569343066e-06, + "loss": 2.0354, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.9555352926254272, + "learning_rate": 4.452554744525548e-06, + "loss": 2.0345, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.157013177871704, + "eval_runtime": 37.6347, + "eval_samples_per_second": 26.571, + "eval_steps_per_second": 1.116, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.8738861083984375, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.0687, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 1.8439863920211792, + "learning_rate": 4.489051094890512e-06, + "loss": 2.0168, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 2.15783953666687, + "learning_rate": 4.507299270072993e-06, + "loss": 2.0169, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 1.8610495328903198, + "learning_rate": 4.525547445255475e-06, + "loss": 1.979, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.9754467010498047, + "learning_rate": 4.543795620437956e-06, + "loss": 2.077, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.07448410987854, + "learning_rate": 4.562043795620438e-06, + "loss": 2.0871, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.891177773475647, + "learning_rate": 4.580291970802921e-06, + "loss": 2.0398, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.9210039377212524, + "learning_rate": 4.598540145985402e-06, + "loss": 1.9975, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 1.8238471746444702, + "learning_rate": 4.6167883211678835e-06, + "loss": 1.9907, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.9147629737854004, + "learning_rate": 4.635036496350365e-06, + "loss": 2.0471, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.156062126159668, + "eval_runtime": 37.6117, + "eval_samples_per_second": 26.587, + "eval_steps_per_second": 1.117, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 2.142754077911377, + "learning_rate": 4.653284671532847e-06, + "loss": 2.0514, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7838966846466064, + "learning_rate": 4.671532846715329e-06, + "loss": 2.0162, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 1.8565750122070312, + "learning_rate": 4.689781021897811e-06, + "loss": 2.0561, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.9184898138046265, + "learning_rate": 4.708029197080292e-06, + "loss": 2.0313, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.904491662979126, + "learning_rate": 4.726277372262774e-06, + "loss": 1.9587, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.8652993440628052, + "learning_rate": 4.744525547445255e-06, + "loss": 2.0292, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 2.383134603500366, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.0369, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 2.039017915725708, + "learning_rate": 4.78102189781022e-06, + "loss": 1.9985, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.8531032800674438, + "learning_rate": 4.799270072992701e-06, + "loss": 2.0272, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 2.0863301753997803, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.0175, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.1532094478607178, + "eval_runtime": 37.6217, + "eval_samples_per_second": 26.58, + "eval_steps_per_second": 1.116, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.9828007221221924, + "learning_rate": 4.835766423357665e-06, + "loss": 1.9632, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.8582276105880737, + "learning_rate": 4.854014598540146e-06, + "loss": 1.9776, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 2.195439577102661, + "learning_rate": 4.872262773722628e-06, + "loss": 1.9669, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 2.1100921630859375, + "learning_rate": 4.89051094890511e-06, + "loss": 2.0061, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.854738712310791, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.0106, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.9990417957305908, + "learning_rate": 4.927007299270074e-06, + "loss": 1.9881, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 2.0629403591156006, + "learning_rate": 4.945255474452555e-06, + "loss": 2.0775, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 2.0885183811187744, + "learning_rate": 4.963503649635037e-06, + "loss": 2.0695, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.9232656955718994, + "learning_rate": 4.981751824817518e-06, + "loss": 2.0147, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.9721243381500244, + "learning_rate": 5e-06, + "loss": 2.0383, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.1488351821899414, + "eval_runtime": 37.5858, + "eval_samples_per_second": 26.606, + "eval_steps_per_second": 1.117, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 2.0475854873657227, + "learning_rate": 4.997969130787978e-06, + "loss": 1.9502, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 2.035140037536621, + "learning_rate": 4.995938261575955e-06, + "loss": 2.0332, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.8700470924377441, + "learning_rate": 4.993907392363932e-06, + "loss": 1.9516, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 1.8098673820495605, + "learning_rate": 4.9918765231519095e-06, + "loss": 1.9985, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.8885918855667114, + "learning_rate": 4.989845653939887e-06, + "loss": 1.9968, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.8569021224975586, + "learning_rate": 4.987814784727864e-06, + "loss": 1.9906, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.9430416822433472, + "learning_rate": 4.985783915515841e-06, + "loss": 2.0023, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.858731746673584, + "learning_rate": 4.983753046303819e-06, + "loss": 1.9946, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.9349243640899658, + "learning_rate": 4.981722177091796e-06, + "loss": 2.0244, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.8817707300186157, + "learning_rate": 4.979691307879773e-06, + "loss": 2.0061, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.1406915187835693, + "eval_runtime": 37.5706, + "eval_samples_per_second": 26.617, + "eval_steps_per_second": 1.118, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.7764776945114136, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.057, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.9521714448928833, + "learning_rate": 4.975629569455728e-06, + "loss": 1.9728, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.9012629985809326, + "learning_rate": 4.973598700243705e-06, + "loss": 1.9789, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.9934824705123901, + "learning_rate": 4.971567831031681e-06, + "loss": 1.8584, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.0267302989959717, + "learning_rate": 4.9695369618196595e-06, + "loss": 1.9512, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.9126142263412476, + "learning_rate": 4.967506092607636e-06, + "loss": 1.9649, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.8503376245498657, + "learning_rate": 4.965475223395614e-06, + "loss": 1.9738, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 1.9278347492218018, + "learning_rate": 4.9634443541835905e-06, + "loss": 1.9882, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 2.1085689067840576, + "learning_rate": 4.961413484971569e-06, + "loss": 1.9989, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9515564441680908, + "learning_rate": 4.959382615759545e-06, + "loss": 1.972, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.143543004989624, + "eval_runtime": 37.0893, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.073535442352295, + "learning_rate": 4.957351746547523e-06, + "loss": 2.1054, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 2.052830934524536, + "learning_rate": 4.9553208773355e-06, + "loss": 1.9983, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.12587571144104, + "learning_rate": 4.953290008123478e-06, + "loss": 1.9291, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.880568265914917, + "learning_rate": 4.951259138911454e-06, + "loss": 2.0618, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.0818779468536377, + "learning_rate": 4.9492282696994315e-06, + "loss": 1.9863, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.165020704269409, + "learning_rate": 4.947197400487409e-06, + "loss": 1.9849, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 2.093449115753174, + "learning_rate": 4.945166531275386e-06, + "loss": 1.9828, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.1064043045043945, + "learning_rate": 4.943135662063363e-06, + "loss": 2.0343, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.8361998796463013, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.0288, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0235719680786133, + "learning_rate": 4.939073923639318e-06, + "loss": 1.9626, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.143022298812866, + "eval_runtime": 37.3847, + "eval_samples_per_second": 26.749, + "eval_steps_per_second": 1.123, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.94564688205719, + "learning_rate": 4.937043054427295e-06, + "loss": 1.9517, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 1.8787976503372192, + "learning_rate": 4.935012185215272e-06, + "loss": 2.0165, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.8160210847854614, + "learning_rate": 4.93298131600325e-06, + "loss": 1.9858, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 1.9638844728469849, + "learning_rate": 4.930950446791227e-06, + "loss": 1.9992, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.8461036682128906, + "learning_rate": 4.928919577579204e-06, + "loss": 1.9872, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 1.9151743650436401, + "learning_rate": 4.9268887083671815e-06, + "loss": 1.9867, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 2.3277945518493652, + "learning_rate": 4.924857839155159e-06, + "loss": 2.0063, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 2.0543594360351562, + "learning_rate": 4.922826969943136e-06, + "loss": 1.9356, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 1.8508672714233398, + "learning_rate": 4.920796100731113e-06, + "loss": 1.9961, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 2.2927565574645996, + "learning_rate": 4.918765231519091e-06, + "loss": 1.9457, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.1402831077575684, + "eval_runtime": 37.5042, + "eval_samples_per_second": 26.664, + "eval_steps_per_second": 1.12, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 1.970535397529602, + "learning_rate": 4.916734362307068e-06, + "loss": 1.9548, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 1.9456896781921387, + "learning_rate": 4.914703493095045e-06, + "loss": 1.9432, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2055933475494385, + "learning_rate": 4.912672623883022e-06, + "loss": 1.9743, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 2.117036819458008, + "learning_rate": 4.910641754671e-06, + "loss": 1.969, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.9397608041763306, + "learning_rate": 4.908610885458977e-06, + "loss": 2.0149, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 2.122511863708496, + "learning_rate": 4.906580016246954e-06, + "loss": 1.9924, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 2.316382884979248, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.0075, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.8495583534240723, + "learning_rate": 4.902518277822909e-06, + "loss": 2.0052, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.9815629720687866, + "learning_rate": 4.900487408610886e-06, + "loss": 2.0154, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.9617043733596802, + "learning_rate": 4.898456539398863e-06, + "loss": 2.0951, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.1340901851654053, + "eval_runtime": 37.5299, + "eval_samples_per_second": 26.645, + "eval_steps_per_second": 1.119, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.967801570892334, + "learning_rate": 4.89642567018684e-06, + "loss": 1.9314, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.9150222539901733, + "learning_rate": 4.894394800974818e-06, + "loss": 2.0229, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 2.201645851135254, + "learning_rate": 4.892363931762794e-06, + "loss": 1.9562, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.8285130262374878, + "learning_rate": 4.8903330625507725e-06, + "loss": 1.9285, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 2.0036182403564453, + "learning_rate": 4.888302193338749e-06, + "loss": 2.0364, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.8367486000061035, + "learning_rate": 4.886271324126727e-06, + "loss": 2.0177, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.8427860736846924, + "learning_rate": 4.8842404549147034e-06, + "loss": 1.9837, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.9150030612945557, + "learning_rate": 4.8822095857026816e-06, + "loss": 1.9313, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 1.8935890197753906, + "learning_rate": 4.880178716490658e-06, + "loss": 1.9457, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 2.0992579460144043, + "learning_rate": 4.878147847278636e-06, + "loss": 1.9143, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.132192850112915, + "eval_runtime": 37.5467, + "eval_samples_per_second": 26.634, + "eval_steps_per_second": 1.119, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.9094585180282593, + "learning_rate": 4.8761169780666125e-06, + "loss": 1.9392, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 2.0265562534332275, + "learning_rate": 4.874086108854591e-06, + "loss": 1.9122, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.8244231939315796, + "learning_rate": 4.872055239642567e-06, + "loss": 1.9737, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8601778745651245, + "learning_rate": 4.870024370430544e-06, + "loss": 2.007, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.8879059553146362, + "learning_rate": 4.867993501218522e-06, + "loss": 1.9333, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 2.1386871337890625, + "learning_rate": 4.865962632006499e-06, + "loss": 1.959, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 2.050854444503784, + "learning_rate": 4.863931762794476e-06, + "loss": 2.001, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0564911365509033, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.0198, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 2.267160654067993, + "learning_rate": 4.859870024370431e-06, + "loss": 2.0268, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 2.0987768173217773, + "learning_rate": 4.857839155158408e-06, + "loss": 2.0075, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.12359356880188, + "eval_runtime": 37.5755, + "eval_samples_per_second": 26.613, + "eval_steps_per_second": 1.118, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.9282751083374023, + "learning_rate": 4.855808285946385e-06, + "loss": 1.9877, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 2.250850200653076, + "learning_rate": 4.8537774167343626e-06, + "loss": 1.9795, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 2.102360725402832, + "learning_rate": 4.85174654752234e-06, + "loss": 1.9848, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8127622604370117, + "learning_rate": 4.849715678310317e-06, + "loss": 1.9695, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 2.092017412185669, + "learning_rate": 4.847684809098294e-06, + "loss": 1.9801, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 2.179563522338867, + "learning_rate": 4.845653939886272e-06, + "loss": 1.9669, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 2.0050835609436035, + "learning_rate": 4.843623070674249e-06, + "loss": 2.0142, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.997016191482544, + "learning_rate": 4.841592201462226e-06, + "loss": 1.9198, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.9335150718688965, + "learning_rate": 4.8395613322502035e-06, + "loss": 1.9702, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9742637872695923, + "learning_rate": 4.837530463038181e-06, + "loss": 2.011, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.1204769611358643, + "eval_runtime": 37.5422, + "eval_samples_per_second": 26.637, + "eval_steps_per_second": 1.119, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.767903208732605, + "learning_rate": 4.835499593826158e-06, + "loss": 1.9775, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 2.180586576461792, + "learning_rate": 4.833468724614135e-06, + "loss": 1.8754, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.8280287981033325, + "learning_rate": 4.831437855402113e-06, + "loss": 1.9467, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 2.026468515396118, + "learning_rate": 4.82940698619009e-06, + "loss": 2.0044, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.9072835445404053, + "learning_rate": 4.827376116978067e-06, + "loss": 1.9821, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.9736080169677734, + "learning_rate": 4.8253452477660444e-06, + "loss": 1.9805, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.8904128074645996, + "learning_rate": 4.823314378554022e-06, + "loss": 1.9493, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 2.245344877243042, + "learning_rate": 4.821283509341998e-06, + "loss": 1.9696, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 2.074371337890625, + "learning_rate": 4.819252640129976e-06, + "loss": 1.9432, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.04290771484375, + "learning_rate": 4.817221770917953e-06, + "loss": 1.9778, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.117993116378784, + "eval_runtime": 37.5736, + "eval_samples_per_second": 26.614, + "eval_steps_per_second": 1.118, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.9518523216247559, + "learning_rate": 4.815190901705931e-06, + "loss": 1.9593, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9472901821136475, + "learning_rate": 4.813160032493907e-06, + "loss": 1.9905, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 2.118521213531494, + "learning_rate": 4.811129163281885e-06, + "loss": 1.9825, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.8100426197052002, + "learning_rate": 4.809098294069862e-06, + "loss": 2.0358, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 2.121580123901367, + "learning_rate": 4.80706742485784e-06, + "loss": 1.952, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.941680669784546, + "learning_rate": 4.805036555645816e-06, + "loss": 2.0051, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 2.023226261138916, + "learning_rate": 4.8030056864337945e-06, + "loss": 1.9639, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.9312723875045776, + "learning_rate": 4.800974817221771e-06, + "loss": 1.9801, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 2.0159788131713867, + "learning_rate": 4.798943948009749e-06, + "loss": 1.9683, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 2.1787915229797363, + "learning_rate": 4.7969130787977254e-06, + "loss": 1.9986, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.111140727996826, + "eval_runtime": 37.5942, + "eval_samples_per_second": 26.6, + "eval_steps_per_second": 1.117, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 1.8020837306976318, + "learning_rate": 4.794882209585703e-06, + "loss": 1.9371, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 2.1437766551971436, + "learning_rate": 4.79285134037368e-06, + "loss": 1.9536, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 2.169379949569702, + "learning_rate": 4.790820471161657e-06, + "loss": 1.9615, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.9989044666290283, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.0055, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 2.2973854541778564, + "learning_rate": 4.786758732737612e-06, + "loss": 1.9199, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 2.104336977005005, + "learning_rate": 4.784727863525589e-06, + "loss": 1.9654, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 2.0384998321533203, + "learning_rate": 4.782696994313566e-06, + "loss": 1.9105, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 2.1892294883728027, + "learning_rate": 4.780666125101544e-06, + "loss": 1.9465, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.9194742441177368, + "learning_rate": 4.778635255889521e-06, + "loss": 1.9424, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 2.2846031188964844, + "learning_rate": 4.776604386677498e-06, + "loss": 2.0083, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.1106128692626953, + "eval_runtime": 37.6075, + "eval_samples_per_second": 26.59, + "eval_steps_per_second": 1.117, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 2.146458625793457, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.0048, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.9768487215042114, + "learning_rate": 4.772542648253453e-06, + "loss": 1.9368, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 2.140291690826416, + "learning_rate": 4.77051177904143e-06, + "loss": 1.9798, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 2.228400230407715, + "learning_rate": 4.768480909829407e-06, + "loss": 1.9534, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 2.103806495666504, + "learning_rate": 4.766450040617385e-06, + "loss": 1.993, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 2.2943012714385986, + "learning_rate": 4.764419171405362e-06, + "loss": 1.9745, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.8355703353881836, + "learning_rate": 4.762388302193339e-06, + "loss": 1.9483, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 2.018692970275879, + "learning_rate": 4.760357432981316e-06, + "loss": 1.9757, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 2.1856532096862793, + "learning_rate": 4.758326563769294e-06, + "loss": 1.9743, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.9867593050003052, + "learning_rate": 4.756295694557271e-06, + "loss": 1.9182, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.1105189323425293, + "eval_runtime": 37.3261, + "eval_samples_per_second": 26.791, + "eval_steps_per_second": 1.125, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.905422329902649, + "learning_rate": 4.754264825345248e-06, + "loss": 1.9294, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 2.2912685871124268, + "learning_rate": 4.7522339561332255e-06, + "loss": 1.8917, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.850225567817688, + "learning_rate": 4.750203086921203e-06, + "loss": 1.9691, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 2.086705446243286, + "learning_rate": 4.74817221770918e-06, + "loss": 2.0256, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 2.111921548843384, + "learning_rate": 4.7461413484971565e-06, + "loss": 1.9094, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8698170185089111, + "learning_rate": 4.744110479285135e-06, + "loss": 1.9378, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 2.0495657920837402, + "learning_rate": 4.742079610073111e-06, + "loss": 1.9511, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.903684139251709, + "learning_rate": 4.740048740861089e-06, + "loss": 1.9559, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 2.0199437141418457, + "learning_rate": 4.738017871649066e-06, + "loss": 2.0162, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 2.1070516109466553, + "learning_rate": 4.735987002437044e-06, + "loss": 2.0247, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.1090195178985596, + "eval_runtime": 37.4923, + "eval_samples_per_second": 26.672, + "eval_steps_per_second": 1.12, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 2.1662495136260986, + "learning_rate": 4.73395613322502e-06, + "loss": 1.9802, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 2.014979124069214, + "learning_rate": 4.731925264012998e-06, + "loss": 1.9256, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 2.1485586166381836, + "learning_rate": 4.729894394800975e-06, + "loss": 1.8788, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 2.037607431411743, + "learning_rate": 4.727863525588953e-06, + "loss": 1.9313, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 2.164149284362793, + "learning_rate": 4.725832656376929e-06, + "loss": 1.9853, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 1.8796825408935547, + "learning_rate": 4.723801787164907e-06, + "loss": 1.9223, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 2.1312918663024902, + "learning_rate": 4.721770917952885e-06, + "loss": 1.9491, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 2.2716944217681885, + "learning_rate": 4.719740048740861e-06, + "loss": 1.9764, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 2.16573429107666, + "learning_rate": 4.717709179528839e-06, + "loss": 1.9642, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.9317946434020996, + "learning_rate": 4.715678310316816e-06, + "loss": 1.9161, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.1036643981933594, + "eval_runtime": 37.6, + "eval_samples_per_second": 26.596, + "eval_steps_per_second": 1.117, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 2.2168219089508057, + "learning_rate": 4.713647441104794e-06, + "loss": 1.9073, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 2.3897385597229004, + "learning_rate": 4.71161657189277e-06, + "loss": 1.9865, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 2.0954325199127197, + "learning_rate": 4.709585702680748e-06, + "loss": 2.0407, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 2.3382396697998047, + "learning_rate": 4.707554833468725e-06, + "loss": 1.9308, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 2.2122690677642822, + "learning_rate": 4.705523964256703e-06, + "loss": 1.9459, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.8825701475143433, + "learning_rate": 4.703493095044679e-06, + "loss": 1.8684, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 2.1475086212158203, + "learning_rate": 4.701462225832657e-06, + "loss": 2.0259, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 2.3192358016967773, + "learning_rate": 4.699431356620634e-06, + "loss": 1.9729, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.7854373455047607, + "learning_rate": 4.697400487408612e-06, + "loss": 1.9489, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 2.2106571197509766, + "learning_rate": 4.695369618196588e-06, + "loss": 1.9406, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.0999557971954346, + "eval_runtime": 37.6178, + "eval_samples_per_second": 26.583, + "eval_steps_per_second": 1.116, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 2.172455072402954, + "learning_rate": 4.693338748984566e-06, + "loss": 1.8869, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 2.002659320831299, + "learning_rate": 4.691307879772543e-06, + "loss": 1.9934, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.9833447933197021, + "learning_rate": 4.68927701056052e-06, + "loss": 1.9586, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 2.026893138885498, + "learning_rate": 4.6872461413484975e-06, + "loss": 1.9572, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 2.099869966506958, + "learning_rate": 4.685215272136475e-06, + "loss": 1.95, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.9816949367523193, + "learning_rate": 4.683184402924452e-06, + "loss": 2.0217, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 2.0313880443573, + "learning_rate": 4.681153533712429e-06, + "loss": 1.9117, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 2.0120208263397217, + "learning_rate": 4.679122664500407e-06, + "loss": 1.8958, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.9500595331192017, + "learning_rate": 4.677091795288384e-06, + "loss": 1.9944, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.877383828163147, + "learning_rate": 4.675060926076361e-06, + "loss": 1.9012, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.0977699756622314, + "eval_runtime": 37.6299, + "eval_samples_per_second": 26.575, + "eval_steps_per_second": 1.116, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.87163507938385, + "learning_rate": 4.673030056864338e-06, + "loss": 1.9072, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.8714340925216675, + "learning_rate": 4.670999187652316e-06, + "loss": 1.9023, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.8581104278564453, + "learning_rate": 4.668968318440293e-06, + "loss": 1.9872, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 2.1439096927642822, + "learning_rate": 4.66693744922827e-06, + "loss": 1.8788, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.9220702648162842, + "learning_rate": 4.6649065800162475e-06, + "loss": 1.9577, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 2.014348268508911, + "learning_rate": 4.662875710804225e-06, + "loss": 1.8707, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 2.099696159362793, + "learning_rate": 4.660844841592202e-06, + "loss": 1.991, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.862485647201538, + "learning_rate": 4.658813972380179e-06, + "loss": 1.8599, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.9138902425765991, + "learning_rate": 4.656783103168157e-06, + "loss": 1.9472, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 2.049471139907837, + "learning_rate": 4.654752233956134e-06, + "loss": 1.9278, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.0987558364868164, + "eval_runtime": 37.6574, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.7767254114151, + "learning_rate": 4.652721364744111e-06, + "loss": 1.9671, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.926533818244934, + "learning_rate": 4.6506904955320885e-06, + "loss": 1.972, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.9930791854858398, + "learning_rate": 4.648659626320066e-06, + "loss": 1.9397, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.9732123613357544, + "learning_rate": 4.646628757108043e-06, + "loss": 1.9065, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9514505863189697, + "learning_rate": 4.6445978878960194e-06, + "loss": 1.9386, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 2.0712249279022217, + "learning_rate": 4.6425670186839976e-06, + "loss": 1.9711, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 2.0905566215515137, + "learning_rate": 4.640536149471974e-06, + "loss": 1.9873, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 2.2611820697784424, + "learning_rate": 4.638505280259952e-06, + "loss": 1.9478, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.9862644672393799, + "learning_rate": 4.6364744110479285e-06, + "loss": 1.9983, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.9715702533721924, + "learning_rate": 4.634443541835907e-06, + "loss": 1.9184, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.09867787361145, + "eval_runtime": 37.6179, + "eval_samples_per_second": 26.583, + "eval_steps_per_second": 1.116, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 2.0207207202911377, + "learning_rate": 4.632412672623883e-06, + "loss": 1.8232, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 2.1219518184661865, + "learning_rate": 4.630381803411861e-06, + "loss": 1.9322, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.8199553489685059, + "learning_rate": 4.628350934199838e-06, + "loss": 1.9603, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 2.093383550643921, + "learning_rate": 4.626320064987816e-06, + "loss": 1.9296, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 2.0259850025177, + "learning_rate": 4.624289195775792e-06, + "loss": 1.9377, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.9583853483200073, + "learning_rate": 4.62225832656377e-06, + "loss": 1.9352, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 2.0816123485565186, + "learning_rate": 4.620227457351747e-06, + "loss": 1.9515, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.9207980632781982, + "learning_rate": 4.618196588139724e-06, + "loss": 1.91, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.8325142860412598, + "learning_rate": 4.616165718927701e-06, + "loss": 1.9966, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.9117507934570312, + "learning_rate": 4.6141348497156786e-06, + "loss": 1.9712, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.0903143882751465, + "eval_runtime": 37.647, + "eval_samples_per_second": 26.563, + "eval_steps_per_second": 1.116, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.8322441577911377, + "learning_rate": 4.612103980503656e-06, + "loss": 1.9416, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.8933136463165283, + "learning_rate": 4.610073111291633e-06, + "loss": 1.9243, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.990027904510498, + "learning_rate": 4.60804224207961e-06, + "loss": 1.9608, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.9176336526870728, + "learning_rate": 4.606011372867588e-06, + "loss": 1.9299, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 2.018432378768921, + "learning_rate": 4.603980503655565e-06, + "loss": 1.8663, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 2.0408661365509033, + "learning_rate": 4.601949634443542e-06, + "loss": 1.8735, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.9252862930297852, + "learning_rate": 4.5999187652315195e-06, + "loss": 1.9216, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.8235865831375122, + "learning_rate": 4.597887896019497e-06, + "loss": 1.928, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.9012048244476318, + "learning_rate": 4.595857026807474e-06, + "loss": 1.913, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.8505122661590576, + "learning_rate": 4.593826157595451e-06, + "loss": 1.9935, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.0896096229553223, + "eval_runtime": 37.6478, + "eval_samples_per_second": 26.562, + "eval_steps_per_second": 1.116, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 1.9841936826705933, + "learning_rate": 4.591795288383429e-06, + "loss": 1.949, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.9749683141708374, + "learning_rate": 4.589764419171406e-06, + "loss": 1.9113, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.9591636657714844, + "learning_rate": 4.587733549959383e-06, + "loss": 1.8972, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 2.040715217590332, + "learning_rate": 4.5857026807473604e-06, + "loss": 1.8661, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 2.228426933288574, + "learning_rate": 4.583671811535338e-06, + "loss": 1.9762, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 2.06842041015625, + "learning_rate": 4.581640942323315e-06, + "loss": 1.9786, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 2.1064720153808594, + "learning_rate": 4.579610073111292e-06, + "loss": 1.9728, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 2.759824275970459, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.0051, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.9189305305480957, + "learning_rate": 4.575548334687247e-06, + "loss": 1.9066, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 2.3444809913635254, + "learning_rate": 4.573517465475224e-06, + "loss": 1.9501, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.0909998416900635, + "eval_runtime": 37.6353, + "eval_samples_per_second": 26.571, + "eval_steps_per_second": 1.116, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 2.5543978214263916, + "learning_rate": 4.571486596263201e-06, + "loss": 1.9634, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.9702211618423462, + "learning_rate": 4.569455727051178e-06, + "loss": 1.9362, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 2.3196160793304443, + "learning_rate": 4.567424857839156e-06, + "loss": 2.0069, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.0895776748657227, + "learning_rate": 4.565393988627132e-06, + "loss": 1.9347, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 2.0111513137817383, + "learning_rate": 4.5633631194151105e-06, + "loss": 1.9359, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 2.0169050693511963, + "learning_rate": 4.561332250203087e-06, + "loss": 1.8756, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 2.2852976322174072, + "learning_rate": 4.559301380991065e-06, + "loss": 1.9285, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 2.046302556991577, + "learning_rate": 4.5572705117790414e-06, + "loss": 1.966, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.0600104331970215, + "learning_rate": 4.5552396425670196e-06, + "loss": 1.8898, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 2.0704476833343506, + "learning_rate": 4.553208773354996e-06, + "loss": 1.9935, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.085967540740967, + "eval_runtime": 37.3159, + "eval_samples_per_second": 26.798, + "eval_steps_per_second": 1.126, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 1.943389892578125, + "learning_rate": 4.551177904142974e-06, + "loss": 1.9105, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.9895267486572266, + "learning_rate": 4.5491470349309505e-06, + "loss": 1.937, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9171580076217651, + "learning_rate": 4.547116165718929e-06, + "loss": 1.9054, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8117244243621826, + "learning_rate": 4.545085296506905e-06, + "loss": 1.8909, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.9584983587265015, + "learning_rate": 4.543054427294882e-06, + "loss": 1.8928, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.9802683591842651, + "learning_rate": 4.54102355808286e-06, + "loss": 1.9218, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.8491944074630737, + "learning_rate": 4.538992688870837e-06, + "loss": 1.9261, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.9530425071716309, + "learning_rate": 4.536961819658814e-06, + "loss": 1.9389, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.883504867553711, + "learning_rate": 4.5349309504467915e-06, + "loss": 1.909, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.8567957878112793, + "learning_rate": 4.532900081234769e-06, + "loss": 1.934, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.084246873855591, + "eval_runtime": 37.423, + "eval_samples_per_second": 26.722, + "eval_steps_per_second": 1.122, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 1.7832186222076416, + "learning_rate": 4.530869212022746e-06, + "loss": 1.9224, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.8867939710617065, + "learning_rate": 4.528838342810723e-06, + "loss": 1.9288, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9270094633102417, + "learning_rate": 4.526807473598701e-06, + "loss": 1.9035, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.988304615020752, + "learning_rate": 4.524776604386678e-06, + "loss": 1.9611, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 2.0739402770996094, + "learning_rate": 4.522745735174655e-06, + "loss": 1.9381, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.7896252870559692, + "learning_rate": 4.520714865962632e-06, + "loss": 1.944, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.9031213521957397, + "learning_rate": 4.51868399675061e-06, + "loss": 1.9017, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.9683704376220703, + "learning_rate": 4.516653127538587e-06, + "loss": 1.8923, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.9705106019973755, + "learning_rate": 4.514622258326564e-06, + "loss": 1.9189, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.866782546043396, + "learning_rate": 4.5125913891145415e-06, + "loss": 1.9592, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.078827142715454, + "eval_runtime": 37.5185, + "eval_samples_per_second": 26.654, + "eval_steps_per_second": 1.119, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.8817288875579834, + "learning_rate": 4.510560519902519e-06, + "loss": 1.8751, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9678497314453125, + "learning_rate": 4.508529650690496e-06, + "loss": 1.8702, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.8850630521774292, + "learning_rate": 4.506498781478473e-06, + "loss": 1.9202, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.8706049919128418, + "learning_rate": 4.504467912266451e-06, + "loss": 1.9906, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.9863978624343872, + "learning_rate": 4.502437043054428e-06, + "loss": 1.938, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8260631561279297, + "learning_rate": 4.500406173842405e-06, + "loss": 1.926, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.8155447244644165, + "learning_rate": 4.4983753046303824e-06, + "loss": 1.9084, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 1.8787226676940918, + "learning_rate": 4.49634443541836e-06, + "loss": 1.9753, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.769175410270691, + "learning_rate": 4.494313566206336e-06, + "loss": 1.9106, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.819743275642395, + "learning_rate": 4.492282696994314e-06, + "loss": 1.8738, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.0850396156311035, + "eval_runtime": 37.5396, + "eval_samples_per_second": 26.639, + "eval_steps_per_second": 1.119, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.8778479099273682, + "learning_rate": 4.490251827782291e-06, + "loss": 1.9032, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.9413594007492065, + "learning_rate": 4.488220958570269e-06, + "loss": 1.8525, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.940185546875, + "learning_rate": 4.486190089358245e-06, + "loss": 1.9583, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.846958041191101, + "learning_rate": 4.484159220146223e-06, + "loss": 1.9348, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.8270854949951172, + "learning_rate": 4.4821283509342e-06, + "loss": 1.9136, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8181536197662354, + "learning_rate": 4.480097481722178e-06, + "loss": 1.8867, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.8371113538742065, + "learning_rate": 4.478066612510154e-06, + "loss": 1.911, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 1.8989226818084717, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.0402, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.9969398975372314, + "learning_rate": 4.474004874086109e-06, + "loss": 1.8697, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.8979823589324951, + "learning_rate": 4.471974004874087e-06, + "loss": 1.9785, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.0822362899780273, + "eval_runtime": 37.5328, + "eval_samples_per_second": 26.643, + "eval_steps_per_second": 1.119, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.853217363357544, + "learning_rate": 4.4699431356620635e-06, + "loss": 1.9221, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8655697107315063, + "learning_rate": 4.467912266450041e-06, + "loss": 1.9166, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.91925048828125, + "learning_rate": 4.465881397238018e-06, + "loss": 1.9262, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.9345825910568237, + "learning_rate": 4.463850528025995e-06, + "loss": 1.9493, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9874762296676636, + "learning_rate": 4.4618196588139726e-06, + "loss": 1.9412, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.9364317655563354, + "learning_rate": 4.45978878960195e-06, + "loss": 1.8791, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 2.027653455734253, + "learning_rate": 4.457757920389927e-06, + "loss": 1.9147, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 1.9354783296585083, + "learning_rate": 4.455727051177904e-06, + "loss": 1.9664, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.8379180431365967, + "learning_rate": 4.453696181965882e-06, + "loss": 1.9329, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 1.9118332862854004, + "learning_rate": 4.451665312753859e-06, + "loss": 1.9863, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.078366756439209, + "eval_runtime": 37.5355, + "eval_samples_per_second": 26.641, + "eval_steps_per_second": 1.119, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.8729168176651, + "learning_rate": 4.449634443541836e-06, + "loss": 1.8839, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.8171452283859253, + "learning_rate": 4.4476035743298135e-06, + "loss": 1.9054, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.963930368423462, + "learning_rate": 4.445572705117791e-06, + "loss": 1.975, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 2.0567688941955566, + "learning_rate": 4.443541835905768e-06, + "loss": 1.912, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.9490222930908203, + "learning_rate": 4.441510966693745e-06, + "loss": 1.9701, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.1417038440704346, + "learning_rate": 4.439480097481723e-06, + "loss": 1.8404, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.8069504499435425, + "learning_rate": 4.4374492282697e-06, + "loss": 1.929, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 1.8261983394622803, + "learning_rate": 4.435418359057677e-06, + "loss": 1.9384, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.90203857421875, + "learning_rate": 4.433387489845654e-06, + "loss": 1.8897, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.9192901849746704, + "learning_rate": 4.431356620633632e-06, + "loss": 1.8997, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.0795819759368896, + "eval_runtime": 37.5379, + "eval_samples_per_second": 26.64, + "eval_steps_per_second": 1.119, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.9556365013122559, + "learning_rate": 4.429325751421609e-06, + "loss": 1.921, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 2.0469253063201904, + "learning_rate": 4.427294882209586e-06, + "loss": 1.855, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 1.903293490409851, + "learning_rate": 4.4252640129975635e-06, + "loss": 1.9228, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.8541758060455322, + "learning_rate": 4.423233143785541e-06, + "loss": 1.9238, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.8954806327819824, + "learning_rate": 4.421202274573518e-06, + "loss": 1.8204, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.9716240167617798, + "learning_rate": 4.4191714053614945e-06, + "loss": 1.8635, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.8913261890411377, + "learning_rate": 4.417140536149473e-06, + "loss": 1.9399, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 2.006896495819092, + "learning_rate": 4.415109666937449e-06, + "loss": 1.8792, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.720639944076538, + "learning_rate": 4.413078797725427e-06, + "loss": 1.9148, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.7996881008148193, + "learning_rate": 4.411047928513404e-06, + "loss": 1.923, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.0718770027160645, + "eval_runtime": 37.5467, + "eval_samples_per_second": 26.633, + "eval_steps_per_second": 1.119, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.8438975811004639, + "learning_rate": 4.409017059301382e-06, + "loss": 1.9436, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.8904563188552856, + "learning_rate": 4.406986190089358e-06, + "loss": 1.946, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.7897257804870605, + "learning_rate": 4.404955320877336e-06, + "loss": 1.8898, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.834946632385254, + "learning_rate": 4.402924451665313e-06, + "loss": 1.8722, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 2.132070779800415, + "learning_rate": 4.400893582453291e-06, + "loss": 1.9082, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.817094087600708, + "learning_rate": 4.398862713241267e-06, + "loss": 1.9222, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.83479642868042, + "learning_rate": 4.396831844029245e-06, + "loss": 1.9077, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 2.0423061847686768, + "learning_rate": 4.394800974817222e-06, + "loss": 1.925, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.9318578243255615, + "learning_rate": 4.392770105605199e-06, + "loss": 1.9363, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.903995156288147, + "learning_rate": 4.390739236393176e-06, + "loss": 1.9379, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.0708532333374023, + "eval_runtime": 37.5382, + "eval_samples_per_second": 26.64, + "eval_steps_per_second": 1.119, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.9046835899353027, + "learning_rate": 4.388708367181154e-06, + "loss": 1.9244, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 2.0306906700134277, + "learning_rate": 4.386677497969131e-06, + "loss": 1.9721, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.8097593784332275, + "learning_rate": 4.384646628757108e-06, + "loss": 1.8959, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.8585506677627563, + "learning_rate": 4.3826157595450855e-06, + "loss": 1.8603, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.8535593748092651, + "learning_rate": 4.380584890333063e-06, + "loss": 1.8164, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.8379638195037842, + "learning_rate": 4.37855402112104e-06, + "loss": 1.9738, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.7997084856033325, + "learning_rate": 4.376523151909017e-06, + "loss": 1.8875, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 2.067387104034424, + "learning_rate": 4.3744922826969946e-06, + "loss": 1.9186, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.9402276277542114, + "learning_rate": 4.372461413484972e-06, + "loss": 1.9545, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.986254096031189, + "learning_rate": 4.370430544272949e-06, + "loss": 1.9596, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.066056728363037, + "eval_runtime": 37.5327, + "eval_samples_per_second": 26.643, + "eval_steps_per_second": 1.119, + "step": 600 + }, + { + "epoch": 0.6586301369863014, + "grad_norm": 1.9289029836654663, + "learning_rate": 4.368399675060926e-06, + "loss": 1.9129, + "step": 601 + }, + { + "epoch": 0.6597260273972603, + "grad_norm": 2.1002073287963867, + "learning_rate": 4.366368805848904e-06, + "loss": 1.8595, + "step": 602 + }, + { + "epoch": 0.6608219178082192, + "grad_norm": 1.8755868673324585, + "learning_rate": 4.364337936636881e-06, + "loss": 1.9254, + "step": 603 + }, + { + "epoch": 0.6619178082191781, + "grad_norm": 2.0698325634002686, + "learning_rate": 4.362307067424858e-06, + "loss": 1.8866, + "step": 604 + }, + { + "epoch": 0.663013698630137, + "grad_norm": 1.7680374383926392, + "learning_rate": 4.3602761982128355e-06, + "loss": 1.9036, + "step": 605 + }, + { + "epoch": 0.6641095890410958, + "grad_norm": 1.9385933876037598, + "learning_rate": 4.358245329000813e-06, + "loss": 1.9139, + "step": 606 + }, + { + "epoch": 0.6652054794520548, + "grad_norm": 2.000523805618286, + "learning_rate": 4.35621445978879e-06, + "loss": 1.8924, + "step": 607 + }, + { + "epoch": 0.6663013698630137, + "grad_norm": 1.8936865329742432, + "learning_rate": 4.354183590576767e-06, + "loss": 1.8531, + "step": 608 + }, + { + "epoch": 0.6673972602739726, + "grad_norm": 1.9764833450317383, + "learning_rate": 4.352152721364745e-06, + "loss": 1.9255, + "step": 609 + }, + { + "epoch": 0.6684931506849315, + "grad_norm": 2.0126848220825195, + "learning_rate": 4.350121852152722e-06, + "loss": 1.9198, + "step": 610 + }, + { + "epoch": 0.6684931506849315, + "eval_loss": 2.0630650520324707, + "eval_runtime": 37.2724, + "eval_samples_per_second": 26.83, + "eval_steps_per_second": 1.127, + "step": 610 + }, + { + "epoch": 0.6695890410958905, + "grad_norm": 1.899513602256775, + "learning_rate": 4.348090982940699e-06, + "loss": 1.9298, + "step": 611 + }, + { + "epoch": 0.6706849315068493, + "grad_norm": 1.8928594589233398, + "learning_rate": 4.3460601137286764e-06, + "loss": 1.894, + "step": 612 + }, + { + "epoch": 0.6717808219178082, + "grad_norm": 1.7667078971862793, + "learning_rate": 4.344029244516654e-06, + "loss": 1.9788, + "step": 613 + }, + { + "epoch": 0.6728767123287671, + "grad_norm": 1.7920507192611694, + "learning_rate": 4.341998375304631e-06, + "loss": 1.9127, + "step": 614 + }, + { + "epoch": 0.673972602739726, + "grad_norm": 2.0444908142089844, + "learning_rate": 4.339967506092607e-06, + "loss": 1.9078, + "step": 615 + }, + { + "epoch": 0.675068493150685, + "grad_norm": 1.9189746379852295, + "learning_rate": 4.3379366368805855e-06, + "loss": 1.8892, + "step": 616 + }, + { + "epoch": 0.6761643835616439, + "grad_norm": 1.9108613729476929, + "learning_rate": 4.335905767668562e-06, + "loss": 1.927, + "step": 617 + }, + { + "epoch": 0.6772602739726027, + "grad_norm": 1.8147832155227661, + "learning_rate": 4.33387489845654e-06, + "loss": 1.9061, + "step": 618 + }, + { + "epoch": 0.6783561643835616, + "grad_norm": 1.8600728511810303, + "learning_rate": 4.3318440292445165e-06, + "loss": 1.8703, + "step": 619 + }, + { + "epoch": 0.6794520547945205, + "grad_norm": 1.8668869733810425, + "learning_rate": 4.329813160032495e-06, + "loss": 1.8604, + "step": 620 + }, + { + "epoch": 0.6794520547945205, + "eval_loss": 2.062499523162842, + "eval_runtime": 37.4372, + "eval_samples_per_second": 26.711, + "eval_steps_per_second": 1.122, + "step": 620 + }, + { + "epoch": 0.6805479452054795, + "grad_norm": 2.1097028255462646, + "learning_rate": 4.327782290820471e-06, + "loss": 1.9431, + "step": 621 + }, + { + "epoch": 0.6816438356164384, + "grad_norm": 1.8451250791549683, + "learning_rate": 4.325751421608449e-06, + "loss": 1.8864, + "step": 622 + }, + { + "epoch": 0.6827397260273973, + "grad_norm": 1.9259945154190063, + "learning_rate": 4.323720552396426e-06, + "loss": 1.8852, + "step": 623 + }, + { + "epoch": 0.6838356164383562, + "grad_norm": 1.9554214477539062, + "learning_rate": 4.321689683184404e-06, + "loss": 1.8893, + "step": 624 + }, + { + "epoch": 0.684931506849315, + "grad_norm": 1.8851567506790161, + "learning_rate": 4.31965881397238e-06, + "loss": 1.9295, + "step": 625 + }, + { + "epoch": 0.686027397260274, + "grad_norm": 1.954161524772644, + "learning_rate": 4.3176279447603574e-06, + "loss": 1.9227, + "step": 626 + }, + { + "epoch": 0.6871232876712329, + "grad_norm": 1.9994785785675049, + "learning_rate": 4.315597075548335e-06, + "loss": 1.9479, + "step": 627 + }, + { + "epoch": 0.6882191780821918, + "grad_norm": 2.0280768871307373, + "learning_rate": 4.313566206336312e-06, + "loss": 1.8915, + "step": 628 + }, + { + "epoch": 0.6893150684931507, + "grad_norm": 1.8648207187652588, + "learning_rate": 4.311535337124289e-06, + "loss": 1.8975, + "step": 629 + }, + { + "epoch": 0.6904109589041096, + "grad_norm": 2.0631797313690186, + "learning_rate": 4.3095044679122665e-06, + "loss": 1.9176, + "step": 630 + }, + { + "epoch": 0.6904109589041096, + "eval_loss": 2.061344623565674, + "eval_runtime": 37.5855, + "eval_samples_per_second": 26.606, + "eval_steps_per_second": 1.117, + "step": 630 + }, + { + "epoch": 0.6915068493150685, + "grad_norm": 1.9490165710449219, + "learning_rate": 4.307473598700244e-06, + "loss": 1.8358, + "step": 631 + }, + { + "epoch": 0.6926027397260274, + "grad_norm": 2.237065076828003, + "learning_rate": 4.305442729488221e-06, + "loss": 1.9377, + "step": 632 + }, + { + "epoch": 0.6936986301369863, + "grad_norm": 1.943978190422058, + "learning_rate": 4.303411860276198e-06, + "loss": 1.9207, + "step": 633 + }, + { + "epoch": 0.6947945205479452, + "grad_norm": 1.9008289575576782, + "learning_rate": 4.301380991064176e-06, + "loss": 1.9503, + "step": 634 + }, + { + "epoch": 0.6958904109589041, + "grad_norm": 2.0620996952056885, + "learning_rate": 4.299350121852153e-06, + "loss": 1.9309, + "step": 635 + }, + { + "epoch": 0.6969863013698631, + "grad_norm": 2.0431745052337646, + "learning_rate": 4.29731925264013e-06, + "loss": 1.8917, + "step": 636 + }, + { + "epoch": 0.6980821917808219, + "grad_norm": 2.0840272903442383, + "learning_rate": 4.2952883834281075e-06, + "loss": 1.8586, + "step": 637 + }, + { + "epoch": 0.6991780821917808, + "grad_norm": 2.059349775314331, + "learning_rate": 4.293257514216085e-06, + "loss": 1.9321, + "step": 638 + }, + { + "epoch": 0.7002739726027397, + "grad_norm": 2.2720837593078613, + "learning_rate": 4.291226645004062e-06, + "loss": 1.9549, + "step": 639 + }, + { + "epoch": 0.7013698630136986, + "grad_norm": 1.8631972074508667, + "learning_rate": 4.289195775792039e-06, + "loss": 1.8824, + "step": 640 + }, + { + "epoch": 0.7013698630136986, + "eval_loss": 2.0568759441375732, + "eval_runtime": 37.6065, + "eval_samples_per_second": 26.591, + "eval_steps_per_second": 1.117, + "step": 640 + }, + { + "epoch": 0.7024657534246576, + "grad_norm": 2.004605770111084, + "learning_rate": 4.287164906580017e-06, + "loss": 1.875, + "step": 641 + }, + { + "epoch": 0.7035616438356165, + "grad_norm": 1.9779096841812134, + "learning_rate": 4.285134037367994e-06, + "loss": 1.8617, + "step": 642 + }, + { + "epoch": 0.7046575342465753, + "grad_norm": 1.884163498878479, + "learning_rate": 4.283103168155971e-06, + "loss": 1.8514, + "step": 643 + }, + { + "epoch": 0.7057534246575342, + "grad_norm": 1.8833951950073242, + "learning_rate": 4.281072298943948e-06, + "loss": 1.8985, + "step": 644 + }, + { + "epoch": 0.7068493150684931, + "grad_norm": 2.054605484008789, + "learning_rate": 4.279041429731926e-06, + "loss": 1.8769, + "step": 645 + }, + { + "epoch": 0.707945205479452, + "grad_norm": 2.112792491912842, + "learning_rate": 4.277010560519903e-06, + "loss": 1.8872, + "step": 646 + }, + { + "epoch": 0.709041095890411, + "grad_norm": 1.8838117122650146, + "learning_rate": 4.27497969130788e-06, + "loss": 1.9302, + "step": 647 + }, + { + "epoch": 0.7101369863013699, + "grad_norm": 1.9498014450073242, + "learning_rate": 4.2729488220958575e-06, + "loss": 1.9433, + "step": 648 + }, + { + "epoch": 0.7112328767123288, + "grad_norm": 1.9905095100402832, + "learning_rate": 4.270917952883835e-06, + "loss": 1.8593, + "step": 649 + }, + { + "epoch": 0.7123287671232876, + "grad_norm": 1.9532862901687622, + "learning_rate": 4.268887083671812e-06, + "loss": 1.9785, + "step": 650 + }, + { + "epoch": 0.7123287671232876, + "eval_loss": 2.0571775436401367, + "eval_runtime": 37.6146, + "eval_samples_per_second": 26.585, + "eval_steps_per_second": 1.117, + "step": 650 + }, + { + "epoch": 0.7134246575342466, + "grad_norm": 1.8848406076431274, + "learning_rate": 4.266856214459789e-06, + "loss": 1.8521, + "step": 651 + }, + { + "epoch": 0.7145205479452055, + "grad_norm": 1.9215036630630493, + "learning_rate": 4.264825345247766e-06, + "loss": 1.8949, + "step": 652 + }, + { + "epoch": 0.7156164383561644, + "grad_norm": 1.9438517093658447, + "learning_rate": 4.262794476035744e-06, + "loss": 1.8104, + "step": 653 + }, + { + "epoch": 0.7167123287671233, + "grad_norm": 1.9203664064407349, + "learning_rate": 4.26076360682372e-06, + "loss": 1.8701, + "step": 654 + }, + { + "epoch": 0.7178082191780822, + "grad_norm": 1.9330025911331177, + "learning_rate": 4.2587327376116984e-06, + "loss": 1.9191, + "step": 655 + }, + { + "epoch": 0.718904109589041, + "grad_norm": 1.8935991525650024, + "learning_rate": 4.256701868399675e-06, + "loss": 1.9015, + "step": 656 + }, + { + "epoch": 0.72, + "grad_norm": 1.8104923963546753, + "learning_rate": 4.254670999187653e-06, + "loss": 1.9069, + "step": 657 + }, + { + "epoch": 0.7210958904109589, + "grad_norm": 1.912269949913025, + "learning_rate": 4.252640129975629e-06, + "loss": 1.8876, + "step": 658 + }, + { + "epoch": 0.7221917808219178, + "grad_norm": 1.94264554977417, + "learning_rate": 4.2506092607636075e-06, + "loss": 1.8255, + "step": 659 + }, + { + "epoch": 0.7232876712328767, + "grad_norm": 1.946187973022461, + "learning_rate": 4.248578391551584e-06, + "loss": 1.8762, + "step": 660 + }, + { + "epoch": 0.7232876712328767, + "eval_loss": 2.0591659545898438, + "eval_runtime": 37.6345, + "eval_samples_per_second": 26.571, + "eval_steps_per_second": 1.116, + "step": 660 + }, + { + "epoch": 0.7243835616438357, + "grad_norm": 1.9981677532196045, + "learning_rate": 4.246547522339562e-06, + "loss": 1.8869, + "step": 661 + }, + { + "epoch": 0.7254794520547945, + "grad_norm": 2.04217267036438, + "learning_rate": 4.2445166531275385e-06, + "loss": 1.8352, + "step": 662 + }, + { + "epoch": 0.7265753424657534, + "grad_norm": 2.1432833671569824, + "learning_rate": 4.242485783915517e-06, + "loss": 1.8529, + "step": 663 + }, + { + "epoch": 0.7276712328767123, + "grad_norm": 1.8703324794769287, + "learning_rate": 4.240454914703493e-06, + "loss": 1.8498, + "step": 664 + }, + { + "epoch": 0.7287671232876712, + "grad_norm": 1.8276578187942505, + "learning_rate": 4.23842404549147e-06, + "loss": 1.8327, + "step": 665 + }, + { + "epoch": 0.7298630136986302, + "grad_norm": 2.031381607055664, + "learning_rate": 4.236393176279448e-06, + "loss": 1.9141, + "step": 666 + }, + { + "epoch": 0.7309589041095891, + "grad_norm": 1.7438545227050781, + "learning_rate": 4.234362307067425e-06, + "loss": 1.9431, + "step": 667 + }, + { + "epoch": 0.7320547945205479, + "grad_norm": 1.8626387119293213, + "learning_rate": 4.232331437855402e-06, + "loss": 1.9316, + "step": 668 + }, + { + "epoch": 0.7331506849315068, + "grad_norm": 2.0319037437438965, + "learning_rate": 4.2303005686433795e-06, + "loss": 1.8908, + "step": 669 + }, + { + "epoch": 0.7342465753424657, + "grad_norm": 1.7567272186279297, + "learning_rate": 4.228269699431357e-06, + "loss": 1.9002, + "step": 670 + }, + { + "epoch": 0.7342465753424657, + "eval_loss": 2.0582168102264404, + "eval_runtime": 37.6169, + "eval_samples_per_second": 26.584, + "eval_steps_per_second": 1.117, + "step": 670 + }, + { + "epoch": 0.7353424657534247, + "grad_norm": 1.9140722751617432, + "learning_rate": 4.226238830219334e-06, + "loss": 1.8943, + "step": 671 + }, + { + "epoch": 0.7364383561643836, + "grad_norm": 1.8341933488845825, + "learning_rate": 4.224207961007311e-06, + "loss": 1.8898, + "step": 672 + }, + { + "epoch": 0.7375342465753425, + "grad_norm": 1.924918293952942, + "learning_rate": 4.2221770917952886e-06, + "loss": 1.9072, + "step": 673 + }, + { + "epoch": 0.7386301369863014, + "grad_norm": 1.9501385688781738, + "learning_rate": 4.220146222583266e-06, + "loss": 1.8955, + "step": 674 + }, + { + "epoch": 0.7397260273972602, + "grad_norm": 1.7961885929107666, + "learning_rate": 4.218115353371243e-06, + "loss": 1.8618, + "step": 675 + }, + { + "epoch": 0.7408219178082192, + "grad_norm": 1.8352770805358887, + "learning_rate": 4.21608448415922e-06, + "loss": 1.9458, + "step": 676 + }, + { + "epoch": 0.7419178082191781, + "grad_norm": 1.9617079496383667, + "learning_rate": 4.214053614947198e-06, + "loss": 1.8369, + "step": 677 + }, + { + "epoch": 0.743013698630137, + "grad_norm": 2.0143256187438965, + "learning_rate": 4.212022745735175e-06, + "loss": 1.9134, + "step": 678 + }, + { + "epoch": 0.7441095890410959, + "grad_norm": 1.9239531755447388, + "learning_rate": 4.209991876523152e-06, + "loss": 1.8393, + "step": 679 + }, + { + "epoch": 0.7452054794520548, + "grad_norm": 1.9146385192871094, + "learning_rate": 4.2079610073111295e-06, + "loss": 1.9378, + "step": 680 + }, + { + "epoch": 0.7452054794520548, + "eval_loss": 2.052194356918335, + "eval_runtime": 37.6157, + "eval_samples_per_second": 26.585, + "eval_steps_per_second": 1.117, + "step": 680 + }, + { + "epoch": 0.7463013698630137, + "grad_norm": 1.7900207042694092, + "learning_rate": 4.205930138099107e-06, + "loss": 1.8953, + "step": 681 + }, + { + "epoch": 0.7473972602739726, + "grad_norm": 1.9322775602340698, + "learning_rate": 4.203899268887084e-06, + "loss": 1.9326, + "step": 682 + }, + { + "epoch": 0.7484931506849315, + "grad_norm": 1.9659802913665771, + "learning_rate": 4.201868399675061e-06, + "loss": 1.8191, + "step": 683 + }, + { + "epoch": 0.7495890410958904, + "grad_norm": 1.9274431467056274, + "learning_rate": 4.199837530463039e-06, + "loss": 1.8868, + "step": 684 + }, + { + "epoch": 0.7506849315068493, + "grad_norm": 1.833921194076538, + "learning_rate": 4.197806661251016e-06, + "loss": 1.8673, + "step": 685 + }, + { + "epoch": 0.7517808219178083, + "grad_norm": 1.8480485677719116, + "learning_rate": 4.195775792038993e-06, + "loss": 1.8075, + "step": 686 + }, + { + "epoch": 0.7528767123287671, + "grad_norm": 2.027998208999634, + "learning_rate": 4.19374492282697e-06, + "loss": 1.911, + "step": 687 + }, + { + "epoch": 0.753972602739726, + "grad_norm": 1.9346990585327148, + "learning_rate": 4.191714053614948e-06, + "loss": 1.8635, + "step": 688 + }, + { + "epoch": 0.7550684931506849, + "grad_norm": 1.9366426467895508, + "learning_rate": 4.189683184402924e-06, + "loss": 1.9337, + "step": 689 + }, + { + "epoch": 0.7561643835616438, + "grad_norm": 1.9615753889083862, + "learning_rate": 4.187652315190902e-06, + "loss": 1.8613, + "step": 690 + }, + { + "epoch": 0.7561643835616438, + "eval_loss": 2.051565647125244, + "eval_runtime": 37.6504, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 690 + }, + { + "epoch": 0.7572602739726028, + "grad_norm": 1.8080062866210938, + "learning_rate": 4.185621445978879e-06, + "loss": 1.8821, + "step": 691 + }, + { + "epoch": 0.7583561643835617, + "grad_norm": 1.840960144996643, + "learning_rate": 4.183590576766857e-06, + "loss": 1.932, + "step": 692 + }, + { + "epoch": 0.7594520547945205, + "grad_norm": 1.9197412729263306, + "learning_rate": 4.181559707554833e-06, + "loss": 1.8485, + "step": 693 + }, + { + "epoch": 0.7605479452054794, + "grad_norm": 2.0478925704956055, + "learning_rate": 4.179528838342811e-06, + "loss": 1.868, + "step": 694 + }, + { + "epoch": 0.7616438356164383, + "grad_norm": 1.8427965641021729, + "learning_rate": 4.177497969130788e-06, + "loss": 1.9516, + "step": 695 + }, + { + "epoch": 0.7627397260273973, + "grad_norm": 1.9763332605361938, + "learning_rate": 4.175467099918766e-06, + "loss": 1.8769, + "step": 696 + }, + { + "epoch": 0.7638356164383562, + "grad_norm": 1.9944579601287842, + "learning_rate": 4.173436230706742e-06, + "loss": 1.8638, + "step": 697 + }, + { + "epoch": 0.7649315068493151, + "grad_norm": 1.910360336303711, + "learning_rate": 4.1714053614947205e-06, + "loss": 1.8682, + "step": 698 + }, + { + "epoch": 0.766027397260274, + "grad_norm": 2.013545036315918, + "learning_rate": 4.169374492282697e-06, + "loss": 1.9218, + "step": 699 + }, + { + "epoch": 0.7671232876712328, + "grad_norm": 2.092571496963501, + "learning_rate": 4.167343623070675e-06, + "loss": 1.8816, + "step": 700 + }, + { + "epoch": 0.7671232876712328, + "eval_loss": 2.050309419631958, + "eval_runtime": 37.6291, + "eval_samples_per_second": 26.575, + "eval_steps_per_second": 1.116, + "step": 700 + }, + { + "epoch": 0.7682191780821918, + "grad_norm": 2.042099714279175, + "learning_rate": 4.165312753858652e-06, + "loss": 1.868, + "step": 701 + }, + { + "epoch": 0.7693150684931507, + "grad_norm": 1.9263545274734497, + "learning_rate": 4.163281884646629e-06, + "loss": 1.9026, + "step": 702 + }, + { + "epoch": 0.7704109589041096, + "grad_norm": 1.8949450254440308, + "learning_rate": 4.161251015434607e-06, + "loss": 1.8497, + "step": 703 + }, + { + "epoch": 0.7715068493150685, + "grad_norm": 2.053208827972412, + "learning_rate": 4.159220146222583e-06, + "loss": 1.8784, + "step": 704 + }, + { + "epoch": 0.7726027397260274, + "grad_norm": 2.0786876678466797, + "learning_rate": 4.157189277010561e-06, + "loss": 1.9072, + "step": 705 + }, + { + "epoch": 0.7736986301369863, + "grad_norm": 2.010636568069458, + "learning_rate": 4.155158407798538e-06, + "loss": 1.8822, + "step": 706 + }, + { + "epoch": 0.7747945205479452, + "grad_norm": 1.9623888731002808, + "learning_rate": 4.153127538586516e-06, + "loss": 1.8957, + "step": 707 + }, + { + "epoch": 0.7758904109589041, + "grad_norm": 2.013491153717041, + "learning_rate": 4.151096669374492e-06, + "loss": 1.8766, + "step": 708 + }, + { + "epoch": 0.776986301369863, + "grad_norm": 1.9842627048492432, + "learning_rate": 4.1490658001624705e-06, + "loss": 1.8749, + "step": 709 + }, + { + "epoch": 0.7780821917808219, + "grad_norm": 1.9263224601745605, + "learning_rate": 4.147034930950447e-06, + "loss": 1.867, + "step": 710 + }, + { + "epoch": 0.7780821917808219, + "eval_loss": 2.0445501804351807, + "eval_runtime": 37.3189, + "eval_samples_per_second": 26.796, + "eval_steps_per_second": 1.125, + "step": 710 + }, + { + "epoch": 0.7791780821917809, + "grad_norm": 1.8684934377670288, + "learning_rate": 4.145004061738425e-06, + "loss": 1.8023, + "step": 711 + }, + { + "epoch": 0.7802739726027397, + "grad_norm": 1.9439146518707275, + "learning_rate": 4.1429731925264015e-06, + "loss": 1.8444, + "step": 712 + }, + { + "epoch": 0.7813698630136986, + "grad_norm": 1.9599238634109497, + "learning_rate": 4.140942323314379e-06, + "loss": 1.8742, + "step": 713 + }, + { + "epoch": 0.7824657534246575, + "grad_norm": 1.8391298055648804, + "learning_rate": 4.138911454102356e-06, + "loss": 1.9015, + "step": 714 + }, + { + "epoch": 0.7835616438356164, + "grad_norm": 1.920380711555481, + "learning_rate": 4.136880584890333e-06, + "loss": 1.8928, + "step": 715 + }, + { + "epoch": 0.7846575342465754, + "grad_norm": 1.859230637550354, + "learning_rate": 4.1348497156783106e-06, + "loss": 1.8917, + "step": 716 + }, + { + "epoch": 0.7857534246575343, + "grad_norm": 1.9734764099121094, + "learning_rate": 4.132818846466288e-06, + "loss": 1.912, + "step": 717 + }, + { + "epoch": 0.7868493150684932, + "grad_norm": 2.058546304702759, + "learning_rate": 4.130787977254265e-06, + "loss": 1.9114, + "step": 718 + }, + { + "epoch": 0.787945205479452, + "grad_norm": 1.9323111772537231, + "learning_rate": 4.128757108042242e-06, + "loss": 1.9622, + "step": 719 + }, + { + "epoch": 0.7890410958904109, + "grad_norm": 2.0416202545166016, + "learning_rate": 4.12672623883022e-06, + "loss": 1.8854, + "step": 720 + }, + { + "epoch": 0.7890410958904109, + "eval_loss": 2.049363374710083, + "eval_runtime": 37.4828, + "eval_samples_per_second": 26.679, + "eval_steps_per_second": 1.121, + "step": 720 + }, + { + "epoch": 0.7901369863013699, + "grad_norm": 2.270467758178711, + "learning_rate": 4.124695369618197e-06, + "loss": 1.8803, + "step": 721 + }, + { + "epoch": 0.7912328767123288, + "grad_norm": 2.138092041015625, + "learning_rate": 4.122664500406174e-06, + "loss": 1.9043, + "step": 722 + }, + { + "epoch": 0.7923287671232877, + "grad_norm": 2.6211955547332764, + "learning_rate": 4.1206336311941515e-06, + "loss": 1.8406, + "step": 723 + }, + { + "epoch": 0.7934246575342466, + "grad_norm": 2.020765781402588, + "learning_rate": 4.118602761982129e-06, + "loss": 1.8168, + "step": 724 + }, + { + "epoch": 0.7945205479452054, + "grad_norm": 2.3701295852661133, + "learning_rate": 4.116571892770106e-06, + "loss": 1.9483, + "step": 725 + }, + { + "epoch": 0.7956164383561644, + "grad_norm": 2.1038002967834473, + "learning_rate": 4.114541023558083e-06, + "loss": 1.8666, + "step": 726 + }, + { + "epoch": 0.7967123287671233, + "grad_norm": 1.922037959098816, + "learning_rate": 4.112510154346061e-06, + "loss": 1.889, + "step": 727 + }, + { + "epoch": 0.7978082191780822, + "grad_norm": 2.082170248031616, + "learning_rate": 4.110479285134038e-06, + "loss": 1.9161, + "step": 728 + }, + { + "epoch": 0.7989041095890411, + "grad_norm": 2.027831554412842, + "learning_rate": 4.108448415922015e-06, + "loss": 1.907, + "step": 729 + }, + { + "epoch": 0.8, + "grad_norm": 1.9608428478240967, + "learning_rate": 4.1064175467099924e-06, + "loss": 1.7652, + "step": 730 + }, + { + "epoch": 0.8, + "eval_loss": 2.049232244491577, + "eval_runtime": 37.5909, + "eval_samples_per_second": 26.602, + "eval_steps_per_second": 1.117, + "step": 730 + }, + { + "epoch": 0.8010958904109589, + "grad_norm": 2.2006771564483643, + "learning_rate": 4.10438667749797e-06, + "loss": 1.9587, + "step": 731 + }, + { + "epoch": 0.8021917808219178, + "grad_norm": 1.954101324081421, + "learning_rate": 4.102355808285947e-06, + "loss": 1.8671, + "step": 732 + }, + { + "epoch": 0.8032876712328767, + "grad_norm": 1.9427088499069214, + "learning_rate": 4.100324939073924e-06, + "loss": 1.8733, + "step": 733 + }, + { + "epoch": 0.8043835616438356, + "grad_norm": 2.1140923500061035, + "learning_rate": 4.0982940698619015e-06, + "loss": 1.8787, + "step": 734 + }, + { + "epoch": 0.8054794520547945, + "grad_norm": 1.9789644479751587, + "learning_rate": 4.096263200649879e-06, + "loss": 1.8499, + "step": 735 + }, + { + "epoch": 0.8065753424657535, + "grad_norm": 1.8676373958587646, + "learning_rate": 4.094232331437856e-06, + "loss": 1.8452, + "step": 736 + }, + { + "epoch": 0.8076712328767123, + "grad_norm": 2.123491048812866, + "learning_rate": 4.092201462225833e-06, + "loss": 1.8901, + "step": 737 + }, + { + "epoch": 0.8087671232876712, + "grad_norm": 1.9341871738433838, + "learning_rate": 4.090170593013811e-06, + "loss": 1.8333, + "step": 738 + }, + { + "epoch": 0.8098630136986301, + "grad_norm": 1.948962688446045, + "learning_rate": 4.088139723801787e-06, + "loss": 1.8564, + "step": 739 + }, + { + "epoch": 0.810958904109589, + "grad_norm": 1.8412338495254517, + "learning_rate": 4.086108854589765e-06, + "loss": 1.8626, + "step": 740 + }, + { + "epoch": 0.810958904109589, + "eval_loss": 2.0422890186309814, + "eval_runtime": 37.5883, + "eval_samples_per_second": 26.604, + "eval_steps_per_second": 1.117, + "step": 740 + }, + { + "epoch": 0.812054794520548, + "grad_norm": 1.9491910934448242, + "learning_rate": 4.084077985377742e-06, + "loss": 1.9331, + "step": 741 + }, + { + "epoch": 0.8131506849315069, + "grad_norm": 2.053166627883911, + "learning_rate": 4.08204711616572e-06, + "loss": 1.8327, + "step": 742 + }, + { + "epoch": 0.8142465753424658, + "grad_norm": 1.7721631526947021, + "learning_rate": 4.080016246953696e-06, + "loss": 1.9131, + "step": 743 + }, + { + "epoch": 0.8153424657534246, + "grad_norm": 1.9681544303894043, + "learning_rate": 4.077985377741674e-06, + "loss": 1.9174, + "step": 744 + }, + { + "epoch": 0.8164383561643835, + "grad_norm": 1.8427720069885254, + "learning_rate": 4.075954508529651e-06, + "loss": 1.937, + "step": 745 + }, + { + "epoch": 0.8175342465753425, + "grad_norm": 1.9430180788040161, + "learning_rate": 4.073923639317629e-06, + "loss": 1.8545, + "step": 746 + }, + { + "epoch": 0.8186301369863014, + "grad_norm": 1.9430502653121948, + "learning_rate": 4.071892770105605e-06, + "loss": 1.8194, + "step": 747 + }, + { + "epoch": 0.8197260273972603, + "grad_norm": 2.0127646923065186, + "learning_rate": 4.069861900893583e-06, + "loss": 1.8531, + "step": 748 + }, + { + "epoch": 0.8208219178082192, + "grad_norm": 2.1581966876983643, + "learning_rate": 4.06783103168156e-06, + "loss": 1.853, + "step": 749 + }, + { + "epoch": 0.821917808219178, + "grad_norm": 2.0918970108032227, + "learning_rate": 4.065800162469538e-06, + "loss": 1.8905, + "step": 750 + }, + { + "epoch": 0.821917808219178, + "eval_loss": 2.041360378265381, + "eval_runtime": 37.6073, + "eval_samples_per_second": 26.591, + "eval_steps_per_second": 1.117, + "step": 750 + }, + { + "epoch": 0.823013698630137, + "grad_norm": 1.972295880317688, + "learning_rate": 4.063769293257514e-06, + "loss": 1.8669, + "step": 751 + }, + { + "epoch": 0.8241095890410959, + "grad_norm": 1.9043899774551392, + "learning_rate": 4.061738424045492e-06, + "loss": 1.8855, + "step": 752 + }, + { + "epoch": 0.8252054794520548, + "grad_norm": 1.9742753505706787, + "learning_rate": 4.059707554833469e-06, + "loss": 1.9149, + "step": 753 + }, + { + "epoch": 0.8263013698630137, + "grad_norm": 1.8181955814361572, + "learning_rate": 4.057676685621446e-06, + "loss": 1.8869, + "step": 754 + }, + { + "epoch": 0.8273972602739726, + "grad_norm": 1.8858075141906738, + "learning_rate": 4.0556458164094235e-06, + "loss": 1.796, + "step": 755 + }, + { + "epoch": 0.8284931506849315, + "grad_norm": 1.9334889650344849, + "learning_rate": 4.053614947197401e-06, + "loss": 1.8746, + "step": 756 + }, + { + "epoch": 0.8295890410958904, + "grad_norm": 1.9887906312942505, + "learning_rate": 4.051584077985378e-06, + "loss": 1.9178, + "step": 757 + }, + { + "epoch": 0.8306849315068493, + "grad_norm": 1.9420117139816284, + "learning_rate": 4.049553208773355e-06, + "loss": 1.8664, + "step": 758 + }, + { + "epoch": 0.8317808219178082, + "grad_norm": 1.9738057851791382, + "learning_rate": 4.047522339561333e-06, + "loss": 1.8682, + "step": 759 + }, + { + "epoch": 0.8328767123287671, + "grad_norm": 1.9430989027023315, + "learning_rate": 4.04549147034931e-06, + "loss": 1.9812, + "step": 760 + }, + { + "epoch": 0.8328767123287671, + "eval_loss": 2.0377445220947266, + "eval_runtime": 37.5872, + "eval_samples_per_second": 26.605, + "eval_steps_per_second": 1.117, + "step": 760 + }, + { + "epoch": 0.8339726027397261, + "grad_norm": 1.9183671474456787, + "learning_rate": 4.043460601137287e-06, + "loss": 1.8738, + "step": 761 + }, + { + "epoch": 0.8350684931506849, + "grad_norm": 1.8702431917190552, + "learning_rate": 4.041429731925264e-06, + "loss": 1.8719, + "step": 762 + }, + { + "epoch": 0.8361643835616438, + "grad_norm": 1.9595237970352173, + "learning_rate": 4.039398862713242e-06, + "loss": 1.9436, + "step": 763 + }, + { + "epoch": 0.8372602739726027, + "grad_norm": 1.8879667520523071, + "learning_rate": 4.037367993501219e-06, + "loss": 1.8587, + "step": 764 + }, + { + "epoch": 0.8383561643835616, + "grad_norm": 2.185802459716797, + "learning_rate": 4.035337124289196e-06, + "loss": 1.8279, + "step": 765 + }, + { + "epoch": 0.8394520547945206, + "grad_norm": 1.8477201461791992, + "learning_rate": 4.0333062550771735e-06, + "loss": 1.872, + "step": 766 + }, + { + "epoch": 0.8405479452054795, + "grad_norm": 1.8418530225753784, + "learning_rate": 4.031275385865151e-06, + "loss": 1.9017, + "step": 767 + }, + { + "epoch": 0.8416438356164384, + "grad_norm": 1.946689486503601, + "learning_rate": 4.029244516653128e-06, + "loss": 1.8092, + "step": 768 + }, + { + "epoch": 0.8427397260273972, + "grad_norm": 1.9002788066864014, + "learning_rate": 4.027213647441105e-06, + "loss": 1.8461, + "step": 769 + }, + { + "epoch": 0.8438356164383561, + "grad_norm": 2.004373788833618, + "learning_rate": 4.025182778229083e-06, + "loss": 1.8772, + "step": 770 + }, + { + "epoch": 0.8438356164383561, + "eval_loss": 2.041391134262085, + "eval_runtime": 37.5643, + "eval_samples_per_second": 26.621, + "eval_steps_per_second": 1.118, + "step": 770 + }, + { + "epoch": 0.8449315068493151, + "grad_norm": 1.9067468643188477, + "learning_rate": 4.02315190901706e-06, + "loss": 1.8697, + "step": 771 + }, + { + "epoch": 0.846027397260274, + "grad_norm": 1.9826830625534058, + "learning_rate": 4.021121039805037e-06, + "loss": 1.9692, + "step": 772 + }, + { + "epoch": 0.8471232876712329, + "grad_norm": 1.8215473890304565, + "learning_rate": 4.0190901705930144e-06, + "loss": 1.8428, + "step": 773 + }, + { + "epoch": 0.8482191780821918, + "grad_norm": 1.8924756050109863, + "learning_rate": 4.017059301380992e-06, + "loss": 1.9199, + "step": 774 + }, + { + "epoch": 0.8493150684931506, + "grad_norm": 2.027142286300659, + "learning_rate": 4.015028432168969e-06, + "loss": 1.837, + "step": 775 + }, + { + "epoch": 0.8504109589041096, + "grad_norm": 1.832080602645874, + "learning_rate": 4.012997562956945e-06, + "loss": 1.8494, + "step": 776 + }, + { + "epoch": 0.8515068493150685, + "grad_norm": 2.0690033435821533, + "learning_rate": 4.0109666937449235e-06, + "loss": 1.859, + "step": 777 + }, + { + "epoch": 0.8526027397260274, + "grad_norm": 2.038332939147949, + "learning_rate": 4.0089358245329e-06, + "loss": 1.8259, + "step": 778 + }, + { + "epoch": 0.8536986301369863, + "grad_norm": 2.0900540351867676, + "learning_rate": 4.006904955320878e-06, + "loss": 1.9069, + "step": 779 + }, + { + "epoch": 0.8547945205479452, + "grad_norm": 1.8365615606307983, + "learning_rate": 4.0048740861088545e-06, + "loss": 1.8944, + "step": 780 + }, + { + "epoch": 0.8547945205479452, + "eval_loss": 2.0401525497436523, + "eval_runtime": 37.5775, + "eval_samples_per_second": 26.612, + "eval_steps_per_second": 1.118, + "step": 780 + }, + { + "epoch": 0.8558904109589041, + "grad_norm": 2.0301108360290527, + "learning_rate": 4.002843216896833e-06, + "loss": 1.8516, + "step": 781 + }, + { + "epoch": 0.856986301369863, + "grad_norm": 2.067864418029785, + "learning_rate": 4.000812347684809e-06, + "loss": 1.8579, + "step": 782 + }, + { + "epoch": 0.8580821917808219, + "grad_norm": 1.9325357675552368, + "learning_rate": 3.998781478472787e-06, + "loss": 1.8285, + "step": 783 + }, + { + "epoch": 0.8591780821917808, + "grad_norm": 2.0696609020233154, + "learning_rate": 3.996750609260764e-06, + "loss": 1.9801, + "step": 784 + }, + { + "epoch": 0.8602739726027397, + "grad_norm": 2.1624257564544678, + "learning_rate": 3.994719740048742e-06, + "loss": 1.8809, + "step": 785 + }, + { + "epoch": 0.8613698630136987, + "grad_norm": 2.0313401222229004, + "learning_rate": 3.992688870836718e-06, + "loss": 1.8311, + "step": 786 + }, + { + "epoch": 0.8624657534246576, + "grad_norm": 1.7534451484680176, + "learning_rate": 3.990658001624696e-06, + "loss": 1.9045, + "step": 787 + }, + { + "epoch": 0.8635616438356164, + "grad_norm": 1.995911717414856, + "learning_rate": 3.988627132412673e-06, + "loss": 1.8597, + "step": 788 + }, + { + "epoch": 0.8646575342465753, + "grad_norm": 1.8776710033416748, + "learning_rate": 3.98659626320065e-06, + "loss": 1.8732, + "step": 789 + }, + { + "epoch": 0.8657534246575342, + "grad_norm": 1.7650753259658813, + "learning_rate": 3.984565393988627e-06, + "loss": 1.8784, + "step": 790 + }, + { + "epoch": 0.8657534246575342, + "eval_loss": 2.038827896118164, + "eval_runtime": 37.55, + "eval_samples_per_second": 26.631, + "eval_steps_per_second": 1.119, + "step": 790 + }, + { + "epoch": 0.8668493150684932, + "grad_norm": 2.1379759311676025, + "learning_rate": 3.9825345247766046e-06, + "loss": 1.8933, + "step": 791 + }, + { + "epoch": 0.8679452054794521, + "grad_norm": 2.026540756225586, + "learning_rate": 3.980503655564582e-06, + "loss": 1.8884, + "step": 792 + }, + { + "epoch": 0.869041095890411, + "grad_norm": 1.8633400201797485, + "learning_rate": 3.978472786352559e-06, + "loss": 1.877, + "step": 793 + }, + { + "epoch": 0.8701369863013698, + "grad_norm": 1.9949527978897095, + "learning_rate": 3.976441917140536e-06, + "loss": 1.9322, + "step": 794 + }, + { + "epoch": 0.8712328767123287, + "grad_norm": 2.0043675899505615, + "learning_rate": 3.974411047928514e-06, + "loss": 1.9033, + "step": 795 + }, + { + "epoch": 0.8723287671232877, + "grad_norm": 1.9397774934768677, + "learning_rate": 3.972380178716491e-06, + "loss": 1.916, + "step": 796 + }, + { + "epoch": 0.8734246575342466, + "grad_norm": 2.0209808349609375, + "learning_rate": 3.970349309504468e-06, + "loss": 1.8239, + "step": 797 + }, + { + "epoch": 0.8745205479452055, + "grad_norm": 2.020632266998291, + "learning_rate": 3.9683184402924455e-06, + "loss": 1.8129, + "step": 798 + }, + { + "epoch": 0.8756164383561644, + "grad_norm": 2.088230848312378, + "learning_rate": 3.966287571080423e-06, + "loss": 1.8871, + "step": 799 + }, + { + "epoch": 0.8767123287671232, + "grad_norm": 1.9384211301803589, + "learning_rate": 3.9642567018684e-06, + "loss": 1.8719, + "step": 800 + }, + { + "epoch": 0.8767123287671232, + "eval_loss": 2.034461498260498, + "eval_runtime": 37.5552, + "eval_samples_per_second": 26.627, + "eval_steps_per_second": 1.118, + "step": 800 + }, + { + "epoch": 0.8778082191780822, + "grad_norm": 1.9446778297424316, + "learning_rate": 3.962225832656377e-06, + "loss": 1.793, + "step": 801 + }, + { + "epoch": 0.8789041095890411, + "grad_norm": 2.129412889480591, + "learning_rate": 3.960194963444355e-06, + "loss": 1.8726, + "step": 802 + }, + { + "epoch": 0.88, + "grad_norm": 1.9478081464767456, + "learning_rate": 3.958164094232332e-06, + "loss": 1.8721, + "step": 803 + }, + { + "epoch": 0.8810958904109589, + "grad_norm": 1.9816874265670776, + "learning_rate": 3.956133225020309e-06, + "loss": 1.8755, + "step": 804 + }, + { + "epoch": 0.8821917808219178, + "grad_norm": 2.174790143966675, + "learning_rate": 3.954102355808286e-06, + "loss": 1.8684, + "step": 805 + }, + { + "epoch": 0.8832876712328767, + "grad_norm": 2.125790596008301, + "learning_rate": 3.952071486596264e-06, + "loss": 1.8588, + "step": 806 + }, + { + "epoch": 0.8843835616438356, + "grad_norm": 2.0364959239959717, + "learning_rate": 3.950040617384241e-06, + "loss": 1.8508, + "step": 807 + }, + { + "epoch": 0.8854794520547945, + "grad_norm": 1.9383763074874878, + "learning_rate": 3.948009748172218e-06, + "loss": 1.9199, + "step": 808 + }, + { + "epoch": 0.8865753424657534, + "grad_norm": 2.0637025833129883, + "learning_rate": 3.9459788789601955e-06, + "loss": 1.8675, + "step": 809 + }, + { + "epoch": 0.8876712328767123, + "grad_norm": 2.1377110481262207, + "learning_rate": 3.943948009748173e-06, + "loss": 1.869, + "step": 810 + }, + { + "epoch": 0.8876712328767123, + "eval_loss": 2.031646251678467, + "eval_runtime": 37.2364, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 810 + }, + { + "epoch": 0.8887671232876713, + "grad_norm": 1.7988247871398926, + "learning_rate": 3.94191714053615e-06, + "loss": 1.9005, + "step": 811 + }, + { + "epoch": 0.8898630136986302, + "grad_norm": 1.9103655815124512, + "learning_rate": 3.939886271324127e-06, + "loss": 1.8245, + "step": 812 + }, + { + "epoch": 0.890958904109589, + "grad_norm": 2.0703070163726807, + "learning_rate": 3.937855402112104e-06, + "loss": 1.8138, + "step": 813 + }, + { + "epoch": 0.8920547945205479, + "grad_norm": 1.9377508163452148, + "learning_rate": 3.935824532900082e-06, + "loss": 1.8904, + "step": 814 + }, + { + "epoch": 0.8931506849315068, + "grad_norm": 1.9508097171783447, + "learning_rate": 3.933793663688058e-06, + "loss": 1.9031, + "step": 815 + }, + { + "epoch": 0.8942465753424658, + "grad_norm": 2.090543508529663, + "learning_rate": 3.9317627944760365e-06, + "loss": 1.8195, + "step": 816 + }, + { + "epoch": 0.8953424657534247, + "grad_norm": 2.041964530944824, + "learning_rate": 3.929731925264013e-06, + "loss": 1.8417, + "step": 817 + }, + { + "epoch": 0.8964383561643836, + "grad_norm": 2.008850574493408, + "learning_rate": 3.927701056051991e-06, + "loss": 1.8383, + "step": 818 + }, + { + "epoch": 0.8975342465753424, + "grad_norm": 1.9293237924575806, + "learning_rate": 3.9256701868399674e-06, + "loss": 1.8295, + "step": 819 + }, + { + "epoch": 0.8986301369863013, + "grad_norm": 2.1226978302001953, + "learning_rate": 3.9236393176279456e-06, + "loss": 1.8282, + "step": 820 + }, + { + "epoch": 0.8986301369863013, + "eval_loss": 2.0328848361968994, + "eval_runtime": 37.4394, + "eval_samples_per_second": 26.71, + "eval_steps_per_second": 1.122, + "step": 820 + }, + { + "epoch": 0.8997260273972603, + "grad_norm": 1.9973869323730469, + "learning_rate": 3.921608448415922e-06, + "loss": 1.8717, + "step": 821 + }, + { + "epoch": 0.9008219178082192, + "grad_norm": 1.8512996435165405, + "learning_rate": 3.9195775792039e-06, + "loss": 1.8462, + "step": 822 + }, + { + "epoch": 0.9019178082191781, + "grad_norm": 1.918342113494873, + "learning_rate": 3.9175467099918765e-06, + "loss": 1.9025, + "step": 823 + }, + { + "epoch": 0.903013698630137, + "grad_norm": 2.0757699012756348, + "learning_rate": 3.915515840779855e-06, + "loss": 1.8877, + "step": 824 + }, + { + "epoch": 0.9041095890410958, + "grad_norm": 1.96065092086792, + "learning_rate": 3.913484971567831e-06, + "loss": 1.7911, + "step": 825 + }, + { + "epoch": 0.9052054794520548, + "grad_norm": 1.9318811893463135, + "learning_rate": 3.911454102355808e-06, + "loss": 1.8832, + "step": 826 + }, + { + "epoch": 0.9063013698630137, + "grad_norm": 2.0639116764068604, + "learning_rate": 3.909423233143786e-06, + "loss": 1.8774, + "step": 827 + }, + { + "epoch": 0.9073972602739726, + "grad_norm": 1.9585458040237427, + "learning_rate": 3.907392363931763e-06, + "loss": 1.9441, + "step": 828 + }, + { + "epoch": 0.9084931506849315, + "grad_norm": 1.9185659885406494, + "learning_rate": 3.90536149471974e-06, + "loss": 1.9617, + "step": 829 + }, + { + "epoch": 0.9095890410958904, + "grad_norm": 1.918180227279663, + "learning_rate": 3.9033306255077175e-06, + "loss": 1.9888, + "step": 830 + }, + { + "epoch": 0.9095890410958904, + "eval_loss": 2.0299599170684814, + "eval_runtime": 37.53, + "eval_samples_per_second": 26.645, + "eval_steps_per_second": 1.119, + "step": 830 + }, + { + "epoch": 0.9106849315068493, + "grad_norm": 2.06184720993042, + "learning_rate": 3.901299756295695e-06, + "loss": 1.8555, + "step": 831 + }, + { + "epoch": 0.9117808219178082, + "grad_norm": 1.9892781972885132, + "learning_rate": 3.899268887083672e-06, + "loss": 1.8581, + "step": 832 + }, + { + "epoch": 0.9128767123287671, + "grad_norm": 2.0172080993652344, + "learning_rate": 3.897238017871649e-06, + "loss": 1.8722, + "step": 833 + }, + { + "epoch": 0.913972602739726, + "grad_norm": 1.8966069221496582, + "learning_rate": 3.8952071486596266e-06, + "loss": 1.8301, + "step": 834 + }, + { + "epoch": 0.915068493150685, + "grad_norm": 2.085993528366089, + "learning_rate": 3.893176279447604e-06, + "loss": 1.9399, + "step": 835 + }, + { + "epoch": 0.9161643835616439, + "grad_norm": 1.8834633827209473, + "learning_rate": 3.891145410235581e-06, + "loss": 1.8812, + "step": 836 + }, + { + "epoch": 0.9172602739726028, + "grad_norm": 2.4515256881713867, + "learning_rate": 3.889114541023558e-06, + "loss": 1.8493, + "step": 837 + }, + { + "epoch": 0.9183561643835616, + "grad_norm": 2.0050864219665527, + "learning_rate": 3.887083671811536e-06, + "loss": 1.7973, + "step": 838 + }, + { + "epoch": 0.9194520547945205, + "grad_norm": 2.533527374267578, + "learning_rate": 3.885052802599513e-06, + "loss": 1.9291, + "step": 839 + }, + { + "epoch": 0.9205479452054794, + "grad_norm": 2.0509860515594482, + "learning_rate": 3.88302193338749e-06, + "loss": 1.8748, + "step": 840 + }, + { + "epoch": 0.9205479452054794, + "eval_loss": 2.0356998443603516, + "eval_runtime": 37.5521, + "eval_samples_per_second": 26.63, + "eval_steps_per_second": 1.118, + "step": 840 + }, + { + "epoch": 0.9216438356164384, + "grad_norm": 3.1594011783599854, + "learning_rate": 3.8809910641754675e-06, + "loss": 1.8616, + "step": 841 + }, + { + "epoch": 0.9227397260273973, + "grad_norm": 2.0838372707366943, + "learning_rate": 3.878960194963445e-06, + "loss": 1.9105, + "step": 842 + }, + { + "epoch": 0.9238356164383562, + "grad_norm": 1.9577261209487915, + "learning_rate": 3.876929325751422e-06, + "loss": 1.8079, + "step": 843 + }, + { + "epoch": 0.924931506849315, + "grad_norm": 1.8972842693328857, + "learning_rate": 3.874898456539399e-06, + "loss": 1.8835, + "step": 844 + }, + { + "epoch": 0.9260273972602739, + "grad_norm": 1.9245659112930298, + "learning_rate": 3.872867587327377e-06, + "loss": 1.8447, + "step": 845 + }, + { + "epoch": 0.9271232876712329, + "grad_norm": 2.0328168869018555, + "learning_rate": 3.870836718115354e-06, + "loss": 1.8458, + "step": 846 + }, + { + "epoch": 0.9282191780821918, + "grad_norm": 2.036590099334717, + "learning_rate": 3.868805848903331e-06, + "loss": 1.8734, + "step": 847 + }, + { + "epoch": 0.9293150684931507, + "grad_norm": 1.8761703968048096, + "learning_rate": 3.8667749796913084e-06, + "loss": 1.8703, + "step": 848 + }, + { + "epoch": 0.9304109589041096, + "grad_norm": 2.226569890975952, + "learning_rate": 3.864744110479286e-06, + "loss": 1.9069, + "step": 849 + }, + { + "epoch": 0.9315068493150684, + "grad_norm": 1.9963139295578003, + "learning_rate": 3.862713241267262e-06, + "loss": 1.912, + "step": 850 + }, + { + "epoch": 0.9315068493150684, + "eval_loss": 2.032756805419922, + "eval_runtime": 37.588, + "eval_samples_per_second": 26.604, + "eval_steps_per_second": 1.117, + "step": 850 + }, + { + "epoch": 0.9326027397260274, + "grad_norm": 2.1044211387634277, + "learning_rate": 3.86068237205524e-06, + "loss": 1.8752, + "step": 851 + }, + { + "epoch": 0.9336986301369863, + "grad_norm": 1.8839032649993896, + "learning_rate": 3.858651502843217e-06, + "loss": 1.7958, + "step": 852 + }, + { + "epoch": 0.9347945205479452, + "grad_norm": 1.9661914110183716, + "learning_rate": 3.856620633631195e-06, + "loss": 1.8598, + "step": 853 + }, + { + "epoch": 0.9358904109589041, + "grad_norm": 1.90200936794281, + "learning_rate": 3.854589764419171e-06, + "loss": 1.9464, + "step": 854 + }, + { + "epoch": 0.936986301369863, + "grad_norm": 1.9577614068984985, + "learning_rate": 3.852558895207149e-06, + "loss": 1.8936, + "step": 855 + }, + { + "epoch": 0.938082191780822, + "grad_norm": 2.0410702228546143, + "learning_rate": 3.850528025995126e-06, + "loss": 1.8381, + "step": 856 + }, + { + "epoch": 0.9391780821917808, + "grad_norm": 1.9985895156860352, + "learning_rate": 3.848497156783104e-06, + "loss": 1.851, + "step": 857 + }, + { + "epoch": 0.9402739726027397, + "grad_norm": 1.8565195798873901, + "learning_rate": 3.84646628757108e-06, + "loss": 1.8452, + "step": 858 + }, + { + "epoch": 0.9413698630136986, + "grad_norm": 2.1560182571411133, + "learning_rate": 3.8444354183590585e-06, + "loss": 1.8482, + "step": 859 + }, + { + "epoch": 0.9424657534246575, + "grad_norm": 2.146472454071045, + "learning_rate": 3.842404549147035e-06, + "loss": 1.816, + "step": 860 + }, + { + "epoch": 0.9424657534246575, + "eval_loss": 2.0297842025756836, + "eval_runtime": 37.561, + "eval_samples_per_second": 26.623, + "eval_steps_per_second": 1.118, + "step": 860 + }, + { + "epoch": 0.9435616438356165, + "grad_norm": 1.9534471035003662, + "learning_rate": 3.840373679935013e-06, + "loss": 1.8583, + "step": 861 + }, + { + "epoch": 0.9446575342465754, + "grad_norm": 1.8210909366607666, + "learning_rate": 3.8383428107229894e-06, + "loss": 1.8276, + "step": 862 + }, + { + "epoch": 0.9457534246575342, + "grad_norm": 2.042600154876709, + "learning_rate": 3.836311941510967e-06, + "loss": 1.879, + "step": 863 + }, + { + "epoch": 0.9468493150684931, + "grad_norm": 1.9915212392807007, + "learning_rate": 3.834281072298944e-06, + "loss": 1.8322, + "step": 864 + }, + { + "epoch": 0.947945205479452, + "grad_norm": 1.9704413414001465, + "learning_rate": 3.832250203086921e-06, + "loss": 1.912, + "step": 865 + }, + { + "epoch": 0.949041095890411, + "grad_norm": 1.9846084117889404, + "learning_rate": 3.8302193338748985e-06, + "loss": 1.9523, + "step": 866 + }, + { + "epoch": 0.9501369863013699, + "grad_norm": 2.2544002532958984, + "learning_rate": 3.828188464662876e-06, + "loss": 1.8813, + "step": 867 + }, + { + "epoch": 0.9512328767123288, + "grad_norm": 1.8418059349060059, + "learning_rate": 3.826157595450853e-06, + "loss": 1.8115, + "step": 868 + }, + { + "epoch": 0.9523287671232876, + "grad_norm": 2.0984606742858887, + "learning_rate": 3.82412672623883e-06, + "loss": 1.8846, + "step": 869 + }, + { + "epoch": 0.9534246575342465, + "grad_norm": 1.8945512771606445, + "learning_rate": 3.822095857026808e-06, + "loss": 1.8322, + "step": 870 + }, + { + "epoch": 0.9534246575342465, + "eval_loss": 2.0224313735961914, + "eval_runtime": 37.5658, + "eval_samples_per_second": 26.62, + "eval_steps_per_second": 1.118, + "step": 870 + }, + { + "epoch": 0.9545205479452055, + "grad_norm": 1.853250503540039, + "learning_rate": 3.820064987814785e-06, + "loss": 1.8334, + "step": 871 + }, + { + "epoch": 0.9556164383561644, + "grad_norm": 1.9714808464050293, + "learning_rate": 3.818034118602762e-06, + "loss": 1.8407, + "step": 872 + }, + { + "epoch": 0.9567123287671233, + "grad_norm": 1.8742942810058594, + "learning_rate": 3.8160032493907395e-06, + "loss": 1.842, + "step": 873 + }, + { + "epoch": 0.9578082191780822, + "grad_norm": 1.8271540403366089, + "learning_rate": 3.813972380178717e-06, + "loss": 1.8828, + "step": 874 + }, + { + "epoch": 0.958904109589041, + "grad_norm": 1.8894344568252563, + "learning_rate": 3.811941510966694e-06, + "loss": 1.8194, + "step": 875 + }, + { + "epoch": 0.96, + "grad_norm": 2.003514051437378, + "learning_rate": 3.809910641754671e-06, + "loss": 1.7662, + "step": 876 + }, + { + "epoch": 0.9610958904109589, + "grad_norm": 1.9021556377410889, + "learning_rate": 3.8078797725426486e-06, + "loss": 1.8613, + "step": 877 + }, + { + "epoch": 0.9621917808219178, + "grad_norm": 2.1169629096984863, + "learning_rate": 3.8058489033306254e-06, + "loss": 1.8362, + "step": 878 + }, + { + "epoch": 0.9632876712328767, + "grad_norm": 1.8580940961837769, + "learning_rate": 3.803818034118603e-06, + "loss": 1.909, + "step": 879 + }, + { + "epoch": 0.9643835616438357, + "grad_norm": 1.8422772884368896, + "learning_rate": 3.80178716490658e-06, + "loss": 1.8899, + "step": 880 + }, + { + "epoch": 0.9643835616438357, + "eval_loss": 2.0180108547210693, + "eval_runtime": 37.6112, + "eval_samples_per_second": 26.588, + "eval_steps_per_second": 1.117, + "step": 880 + }, + { + "epoch": 0.9654794520547946, + "grad_norm": 2.0347087383270264, + "learning_rate": 3.7997562956945577e-06, + "loss": 1.8454, + "step": 881 + }, + { + "epoch": 0.9665753424657534, + "grad_norm": 1.9369441270828247, + "learning_rate": 3.7977254264825345e-06, + "loss": 1.8273, + "step": 882 + }, + { + "epoch": 0.9676712328767123, + "grad_norm": 1.8486833572387695, + "learning_rate": 3.7956945572705122e-06, + "loss": 1.916, + "step": 883 + }, + { + "epoch": 0.9687671232876712, + "grad_norm": 1.9250234365463257, + "learning_rate": 3.793663688058489e-06, + "loss": 1.9197, + "step": 884 + }, + { + "epoch": 0.9698630136986301, + "grad_norm": 1.8732974529266357, + "learning_rate": 3.791632818846467e-06, + "loss": 1.9094, + "step": 885 + }, + { + "epoch": 0.9709589041095891, + "grad_norm": 1.963775634765625, + "learning_rate": 3.7896019496344436e-06, + "loss": 1.9058, + "step": 886 + }, + { + "epoch": 0.972054794520548, + "grad_norm": 1.9569355249404907, + "learning_rate": 3.787571080422421e-06, + "loss": 1.8822, + "step": 887 + }, + { + "epoch": 0.9731506849315068, + "grad_norm": 1.8828372955322266, + "learning_rate": 3.785540211210398e-06, + "loss": 1.8666, + "step": 888 + }, + { + "epoch": 0.9742465753424657, + "grad_norm": 1.8864827156066895, + "learning_rate": 3.7835093419983755e-06, + "loss": 1.8477, + "step": 889 + }, + { + "epoch": 0.9753424657534246, + "grad_norm": 1.8801312446594238, + "learning_rate": 3.7814784727863527e-06, + "loss": 1.9648, + "step": 890 + }, + { + "epoch": 0.9753424657534246, + "eval_loss": 2.019768476486206, + "eval_runtime": 37.5977, + "eval_samples_per_second": 26.597, + "eval_steps_per_second": 1.117, + "step": 890 + }, + { + "epoch": 0.9764383561643836, + "grad_norm": 1.944361925125122, + "learning_rate": 3.77944760357433e-06, + "loss": 1.9068, + "step": 891 + }, + { + "epoch": 0.9775342465753425, + "grad_norm": 1.8992615938186646, + "learning_rate": 3.7774167343623073e-06, + "loss": 1.8328, + "step": 892 + }, + { + "epoch": 0.9786301369863014, + "grad_norm": 1.8111578226089478, + "learning_rate": 3.7753858651502846e-06, + "loss": 1.9014, + "step": 893 + }, + { + "epoch": 0.9797260273972602, + "grad_norm": 2.143831968307495, + "learning_rate": 3.773354995938262e-06, + "loss": 1.8704, + "step": 894 + }, + { + "epoch": 0.9808219178082191, + "grad_norm": 1.8355118036270142, + "learning_rate": 3.771324126726239e-06, + "loss": 1.907, + "step": 895 + }, + { + "epoch": 0.9819178082191781, + "grad_norm": 1.8565425872802734, + "learning_rate": 3.7692932575142164e-06, + "loss": 1.8407, + "step": 896 + }, + { + "epoch": 0.983013698630137, + "grad_norm": 1.9198476076126099, + "learning_rate": 3.7672623883021937e-06, + "loss": 1.7668, + "step": 897 + }, + { + "epoch": 0.9841095890410959, + "grad_norm": 1.8843393325805664, + "learning_rate": 3.765231519090171e-06, + "loss": 1.8721, + "step": 898 + }, + { + "epoch": 0.9852054794520548, + "grad_norm": 2.0561683177948, + "learning_rate": 3.7632006498781482e-06, + "loss": 1.8124, + "step": 899 + }, + { + "epoch": 0.9863013698630136, + "grad_norm": 1.8568141460418701, + "learning_rate": 3.761169780666125e-06, + "loss": 1.8655, + "step": 900 + }, + { + "epoch": 0.9863013698630136, + "eval_loss": 2.022801160812378, + "eval_runtime": 37.639, + "eval_samples_per_second": 26.568, + "eval_steps_per_second": 1.116, + "step": 900 + }, + { + "epoch": 0.9873972602739726, + "grad_norm": 1.8860610723495483, + "learning_rate": 3.7591389114541028e-06, + "loss": 1.8913, + "step": 901 + }, + { + "epoch": 0.9884931506849315, + "grad_norm": 1.9264308214187622, + "learning_rate": 3.7571080422420796e-06, + "loss": 1.8346, + "step": 902 + }, + { + "epoch": 0.9895890410958904, + "grad_norm": 1.8701164722442627, + "learning_rate": 3.7550771730300573e-06, + "loss": 1.9085, + "step": 903 + }, + { + "epoch": 0.9906849315068493, + "grad_norm": 1.794785737991333, + "learning_rate": 3.753046303818034e-06, + "loss": 1.8809, + "step": 904 + }, + { + "epoch": 0.9917808219178083, + "grad_norm": 2.028759241104126, + "learning_rate": 3.751015434606012e-06, + "loss": 1.9283, + "step": 905 + }, + { + "epoch": 0.9928767123287672, + "grad_norm": 1.9132484197616577, + "learning_rate": 3.7489845653939887e-06, + "loss": 1.8221, + "step": 906 + }, + { + "epoch": 0.993972602739726, + "grad_norm": 2.0630099773406982, + "learning_rate": 3.7469536961819664e-06, + "loss": 1.8499, + "step": 907 + }, + { + "epoch": 0.9950684931506849, + "grad_norm": 1.8041224479675293, + "learning_rate": 3.7449228269699433e-06, + "loss": 1.8763, + "step": 908 + }, + { + "epoch": 0.9961643835616438, + "grad_norm": 1.9102509021759033, + "learning_rate": 3.742891957757921e-06, + "loss": 1.8677, + "step": 909 + }, + { + "epoch": 0.9972602739726028, + "grad_norm": 1.8909201622009277, + "learning_rate": 3.740861088545898e-06, + "loss": 1.8998, + "step": 910 + }, + { + "epoch": 0.9972602739726028, + "eval_loss": 2.0248031616210938, + "eval_runtime": 37.2988, + "eval_samples_per_second": 26.811, + "eval_steps_per_second": 1.126, + "step": 910 + }, + { + "epoch": 0.9983561643835617, + "grad_norm": 1.927571415901184, + "learning_rate": 3.7388302193338755e-06, + "loss": 1.8308, + "step": 911 + }, + { + "epoch": 0.9994520547945206, + "grad_norm": 1.9914634227752686, + "learning_rate": 3.7367993501218524e-06, + "loss": 1.8336, + "step": 912 + }, + { + "epoch": 1.0005479452054795, + "grad_norm": 1.9946560859680176, + "learning_rate": 3.7347684809098297e-06, + "loss": 1.8051, + "step": 913 + }, + { + "epoch": 1.0016438356164383, + "grad_norm": 1.988524079322815, + "learning_rate": 3.732737611697807e-06, + "loss": 1.7504, + "step": 914 + }, + { + "epoch": 1.0027397260273974, + "grad_norm": 1.9622360467910767, + "learning_rate": 3.7307067424857842e-06, + "loss": 1.7998, + "step": 915 + }, + { + "epoch": 1.0038356164383562, + "grad_norm": 1.9427080154418945, + "learning_rate": 3.7286758732737615e-06, + "loss": 1.7888, + "step": 916 + }, + { + "epoch": 1.004931506849315, + "grad_norm": 2.2243752479553223, + "learning_rate": 3.7266450040617388e-06, + "loss": 1.9293, + "step": 917 + }, + { + "epoch": 1.006027397260274, + "grad_norm": 2.104738235473633, + "learning_rate": 3.724614134849716e-06, + "loss": 1.769, + "step": 918 + }, + { + "epoch": 1.0071232876712328, + "grad_norm": 2.187030076980591, + "learning_rate": 3.7225832656376933e-06, + "loss": 1.8897, + "step": 919 + }, + { + "epoch": 1.0082191780821919, + "grad_norm": 2.1332106590270996, + "learning_rate": 3.7205523964256706e-06, + "loss": 1.8558, + "step": 920 + }, + { + "epoch": 1.0082191780821919, + "eval_loss": 2.021728992462158, + "eval_runtime": 37.495, + "eval_samples_per_second": 26.67, + "eval_steps_per_second": 1.12, + "step": 920 + }, + { + "epoch": 1.0093150684931507, + "grad_norm": 2.212008237838745, + "learning_rate": 3.718521527213648e-06, + "loss": 1.8444, + "step": 921 + }, + { + "epoch": 1.0104109589041095, + "grad_norm": 2.130812883377075, + "learning_rate": 3.716490658001625e-06, + "loss": 1.7314, + "step": 922 + }, + { + "epoch": 1.0115068493150685, + "grad_norm": 2.1988723278045654, + "learning_rate": 3.7144597887896024e-06, + "loss": 1.859, + "step": 923 + }, + { + "epoch": 1.0126027397260273, + "grad_norm": 2.16463565826416, + "learning_rate": 3.7124289195775797e-06, + "loss": 1.8411, + "step": 924 + }, + { + "epoch": 1.0136986301369864, + "grad_norm": 2.253486156463623, + "learning_rate": 3.710398050365557e-06, + "loss": 1.8142, + "step": 925 + }, + { + "epoch": 1.0147945205479452, + "grad_norm": 2.402863025665283, + "learning_rate": 3.708367181153534e-06, + "loss": 1.8589, + "step": 926 + }, + { + "epoch": 1.0158904109589042, + "grad_norm": 2.0467870235443115, + "learning_rate": 3.7063363119415115e-06, + "loss": 1.804, + "step": 927 + }, + { + "epoch": 1.016986301369863, + "grad_norm": 2.186394214630127, + "learning_rate": 3.7043054427294884e-06, + "loss": 1.7816, + "step": 928 + }, + { + "epoch": 1.0180821917808218, + "grad_norm": 2.056140184402466, + "learning_rate": 3.702274573517466e-06, + "loss": 1.8076, + "step": 929 + }, + { + "epoch": 1.0191780821917809, + "grad_norm": 2.055403709411621, + "learning_rate": 3.700243704305443e-06, + "loss": 1.7881, + "step": 930 + }, + { + "epoch": 1.0191780821917809, + "eval_loss": 2.0193421840667725, + "eval_runtime": 37.612, + "eval_samples_per_second": 26.587, + "eval_steps_per_second": 1.117, + "step": 930 + }, + { + "epoch": 1.0202739726027397, + "grad_norm": 2.1633987426757812, + "learning_rate": 3.6982128350934206e-06, + "loss": 1.7707, + "step": 931 + }, + { + "epoch": 1.0213698630136987, + "grad_norm": 2.2342350482940674, + "learning_rate": 3.6961819658813975e-06, + "loss": 1.8996, + "step": 932 + }, + { + "epoch": 1.0224657534246575, + "grad_norm": 1.9219639301300049, + "learning_rate": 3.694151096669375e-06, + "loss": 1.8224, + "step": 933 + }, + { + "epoch": 1.0235616438356165, + "grad_norm": 2.1321310997009277, + "learning_rate": 3.692120227457352e-06, + "loss": 1.8446, + "step": 934 + }, + { + "epoch": 1.0246575342465754, + "grad_norm": 2.174525499343872, + "learning_rate": 3.6900893582453297e-06, + "loss": 1.7404, + "step": 935 + }, + { + "epoch": 1.0257534246575342, + "grad_norm": 2.165740966796875, + "learning_rate": 3.6880584890333066e-06, + "loss": 1.8301, + "step": 936 + }, + { + "epoch": 1.0268493150684932, + "grad_norm": 2.0366575717926025, + "learning_rate": 3.6860276198212834e-06, + "loss": 1.8451, + "step": 937 + }, + { + "epoch": 1.027945205479452, + "grad_norm": 2.0515666007995605, + "learning_rate": 3.683996750609261e-06, + "loss": 1.8416, + "step": 938 + }, + { + "epoch": 1.029041095890411, + "grad_norm": 2.1161012649536133, + "learning_rate": 3.681965881397238e-06, + "loss": 1.857, + "step": 939 + }, + { + "epoch": 1.0301369863013699, + "grad_norm": 2.1115732192993164, + "learning_rate": 3.6799350121852157e-06, + "loss": 1.8088, + "step": 940 + }, + { + "epoch": 1.0301369863013699, + "eval_loss": 2.013059139251709, + "eval_runtime": 37.6392, + "eval_samples_per_second": 26.568, + "eval_steps_per_second": 1.116, + "step": 940 + }, + { + "epoch": 1.0312328767123287, + "grad_norm": 1.8958100080490112, + "learning_rate": 3.6779041429731925e-06, + "loss": 1.8267, + "step": 941 + }, + { + "epoch": 1.0323287671232877, + "grad_norm": 2.2424497604370117, + "learning_rate": 3.6758732737611702e-06, + "loss": 1.8073, + "step": 942 + }, + { + "epoch": 1.0334246575342465, + "grad_norm": 2.0759358406066895, + "learning_rate": 3.673842404549147e-06, + "loss": 1.8155, + "step": 943 + }, + { + "epoch": 1.0345205479452055, + "grad_norm": 2.242558002471924, + "learning_rate": 3.671811535337125e-06, + "loss": 1.8218, + "step": 944 + }, + { + "epoch": 1.0356164383561643, + "grad_norm": 2.074090003967285, + "learning_rate": 3.6697806661251016e-06, + "loss": 1.9521, + "step": 945 + }, + { + "epoch": 1.0367123287671234, + "grad_norm": 2.0505194664001465, + "learning_rate": 3.6677497969130793e-06, + "loss": 1.826, + "step": 946 + }, + { + "epoch": 1.0378082191780822, + "grad_norm": 2.181986093521118, + "learning_rate": 3.665718927701056e-06, + "loss": 1.8917, + "step": 947 + }, + { + "epoch": 1.038904109589041, + "grad_norm": 2.147963762283325, + "learning_rate": 3.663688058489034e-06, + "loss": 1.8162, + "step": 948 + }, + { + "epoch": 1.04, + "grad_norm": 1.943077564239502, + "learning_rate": 3.6616571892770107e-06, + "loss": 1.7558, + "step": 949 + }, + { + "epoch": 1.0410958904109588, + "grad_norm": 2.0034351348876953, + "learning_rate": 3.659626320064988e-06, + "loss": 1.8177, + "step": 950 + }, + { + "epoch": 1.0410958904109588, + "eval_loss": 2.014491558074951, + "eval_runtime": 37.6354, + "eval_samples_per_second": 26.571, + "eval_steps_per_second": 1.116, + "step": 950 + }, + { + "epoch": 1.0421917808219179, + "grad_norm": 2.1404898166656494, + "learning_rate": 3.6575954508529653e-06, + "loss": 1.8071, + "step": 951 + }, + { + "epoch": 1.0432876712328767, + "grad_norm": 2.12056303024292, + "learning_rate": 3.6555645816409426e-06, + "loss": 1.838, + "step": 952 + }, + { + "epoch": 1.0443835616438357, + "grad_norm": 1.998156189918518, + "learning_rate": 3.65353371242892e-06, + "loss": 1.6913, + "step": 953 + }, + { + "epoch": 1.0454794520547945, + "grad_norm": 2.0770769119262695, + "learning_rate": 3.651502843216897e-06, + "loss": 1.8411, + "step": 954 + }, + { + "epoch": 1.0465753424657533, + "grad_norm": 2.126035690307617, + "learning_rate": 3.6494719740048744e-06, + "loss": 1.7859, + "step": 955 + }, + { + "epoch": 1.0476712328767124, + "grad_norm": 1.9138388633728027, + "learning_rate": 3.6474411047928517e-06, + "loss": 1.8313, + "step": 956 + }, + { + "epoch": 1.0487671232876712, + "grad_norm": 2.0706698894500732, + "learning_rate": 3.645410235580829e-06, + "loss": 1.8133, + "step": 957 + }, + { + "epoch": 1.0498630136986302, + "grad_norm": 2.0428574085235596, + "learning_rate": 3.6433793663688062e-06, + "loss": 1.8009, + "step": 958 + }, + { + "epoch": 1.050958904109589, + "grad_norm": 2.0169312953948975, + "learning_rate": 3.6413484971567835e-06, + "loss": 1.8403, + "step": 959 + }, + { + "epoch": 1.0520547945205478, + "grad_norm": 1.980191946029663, + "learning_rate": 3.6393176279447608e-06, + "loss": 1.8465, + "step": 960 + }, + { + "epoch": 1.0520547945205478, + "eval_loss": 2.0130155086517334, + "eval_runtime": 37.6583, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 960 + }, + { + "epoch": 1.0531506849315069, + "grad_norm": 2.1955149173736572, + "learning_rate": 3.637286758732738e-06, + "loss": 1.8665, + "step": 961 + }, + { + "epoch": 1.0542465753424657, + "grad_norm": 2.0228099822998047, + "learning_rate": 3.6352558895207153e-06, + "loss": 1.8595, + "step": 962 + }, + { + "epoch": 1.0553424657534247, + "grad_norm": 1.9874252080917358, + "learning_rate": 3.633225020308692e-06, + "loss": 1.8214, + "step": 963 + }, + { + "epoch": 1.0564383561643835, + "grad_norm": 2.0241470336914062, + "learning_rate": 3.63119415109667e-06, + "loss": 1.8524, + "step": 964 + }, + { + "epoch": 1.0575342465753426, + "grad_norm": 1.9674946069717407, + "learning_rate": 3.6291632818846467e-06, + "loss": 1.7547, + "step": 965 + }, + { + "epoch": 1.0586301369863014, + "grad_norm": 1.9529986381530762, + "learning_rate": 3.6271324126726244e-06, + "loss": 1.7633, + "step": 966 + }, + { + "epoch": 1.0597260273972602, + "grad_norm": 2.122602701187134, + "learning_rate": 3.6251015434606013e-06, + "loss": 1.8139, + "step": 967 + }, + { + "epoch": 1.0608219178082192, + "grad_norm": 1.9844366312026978, + "learning_rate": 3.623070674248579e-06, + "loss": 1.7898, + "step": 968 + }, + { + "epoch": 1.061917808219178, + "grad_norm": 2.007964611053467, + "learning_rate": 3.621039805036556e-06, + "loss": 1.8564, + "step": 969 + }, + { + "epoch": 1.063013698630137, + "grad_norm": 2.070485830307007, + "learning_rate": 3.6190089358245335e-06, + "loss": 1.8292, + "step": 970 + }, + { + "epoch": 1.063013698630137, + "eval_loss": 2.012418270111084, + "eval_runtime": 37.6482, + "eval_samples_per_second": 26.562, + "eval_steps_per_second": 1.116, + "step": 970 + }, + { + "epoch": 1.0641095890410959, + "grad_norm": 2.1049320697784424, + "learning_rate": 3.6169780666125104e-06, + "loss": 1.8458, + "step": 971 + }, + { + "epoch": 1.0652054794520547, + "grad_norm": 2.1623475551605225, + "learning_rate": 3.614947197400488e-06, + "loss": 1.7463, + "step": 972 + }, + { + "epoch": 1.0663013698630137, + "grad_norm": 1.9398975372314453, + "learning_rate": 3.612916328188465e-06, + "loss": 1.8493, + "step": 973 + }, + { + "epoch": 1.0673972602739725, + "grad_norm": 2.1070353984832764, + "learning_rate": 3.610885458976442e-06, + "loss": 1.8378, + "step": 974 + }, + { + "epoch": 1.0684931506849316, + "grad_norm": 2.0092310905456543, + "learning_rate": 3.6088545897644195e-06, + "loss": 1.7752, + "step": 975 + }, + { + "epoch": 1.0695890410958904, + "grad_norm": 1.9634876251220703, + "learning_rate": 3.6068237205523963e-06, + "loss": 1.7765, + "step": 976 + }, + { + "epoch": 1.0706849315068494, + "grad_norm": 2.0078225135803223, + "learning_rate": 3.604792851340374e-06, + "loss": 1.8677, + "step": 977 + }, + { + "epoch": 1.0717808219178082, + "grad_norm": 2.190459728240967, + "learning_rate": 3.602761982128351e-06, + "loss": 1.8411, + "step": 978 + }, + { + "epoch": 1.072876712328767, + "grad_norm": 2.105957269668579, + "learning_rate": 3.6007311129163286e-06, + "loss": 1.7646, + "step": 979 + }, + { + "epoch": 1.073972602739726, + "grad_norm": 2.068345308303833, + "learning_rate": 3.5987002437043054e-06, + "loss": 1.8007, + "step": 980 + }, + { + "epoch": 1.073972602739726, + "eval_loss": 2.013199806213379, + "eval_runtime": 37.6279, + "eval_samples_per_second": 26.576, + "eval_steps_per_second": 1.116, + "step": 980 + }, + { + "epoch": 1.0750684931506849, + "grad_norm": 2.067021369934082, + "learning_rate": 3.596669374492283e-06, + "loss": 1.891, + "step": 981 + }, + { + "epoch": 1.076164383561644, + "grad_norm": 2.040861129760742, + "learning_rate": 3.59463850528026e-06, + "loss": 1.7647, + "step": 982 + }, + { + "epoch": 1.0772602739726027, + "grad_norm": 1.9669679403305054, + "learning_rate": 3.5926076360682377e-06, + "loss": 1.8588, + "step": 983 + }, + { + "epoch": 1.0783561643835617, + "grad_norm": 2.008066177368164, + "learning_rate": 3.5905767668562145e-06, + "loss": 1.8265, + "step": 984 + }, + { + "epoch": 1.0794520547945206, + "grad_norm": 1.9229918718338013, + "learning_rate": 3.5885458976441922e-06, + "loss": 1.8558, + "step": 985 + }, + { + "epoch": 1.0805479452054794, + "grad_norm": 2.1782007217407227, + "learning_rate": 3.586515028432169e-06, + "loss": 1.8692, + "step": 986 + }, + { + "epoch": 1.0816438356164384, + "grad_norm": 2.159749984741211, + "learning_rate": 3.5844841592201464e-06, + "loss": 1.7497, + "step": 987 + }, + { + "epoch": 1.0827397260273972, + "grad_norm": 2.0980350971221924, + "learning_rate": 3.5824532900081237e-06, + "loss": 1.8191, + "step": 988 + }, + { + "epoch": 1.0838356164383562, + "grad_norm": 1.987552285194397, + "learning_rate": 3.580422420796101e-06, + "loss": 1.8278, + "step": 989 + }, + { + "epoch": 1.084931506849315, + "grad_norm": 2.080035448074341, + "learning_rate": 3.578391551584078e-06, + "loss": 1.7288, + "step": 990 + }, + { + "epoch": 1.084931506849315, + "eval_loss": 2.011539936065674, + "eval_runtime": 37.6358, + "eval_samples_per_second": 26.57, + "eval_steps_per_second": 1.116, + "step": 990 + }, + { + "epoch": 1.086027397260274, + "grad_norm": 1.911167860031128, + "learning_rate": 3.5763606823720555e-06, + "loss": 1.829, + "step": 991 + }, + { + "epoch": 1.087123287671233, + "grad_norm": 2.0227794647216797, + "learning_rate": 3.5743298131600328e-06, + "loss": 1.8537, + "step": 992 + }, + { + "epoch": 1.0882191780821917, + "grad_norm": 2.2071900367736816, + "learning_rate": 3.57229894394801e-06, + "loss": 1.8351, + "step": 993 + }, + { + "epoch": 1.0893150684931507, + "grad_norm": 2.071361780166626, + "learning_rate": 3.5702680747359873e-06, + "loss": 1.7783, + "step": 994 + }, + { + "epoch": 1.0904109589041096, + "grad_norm": 2.1811583042144775, + "learning_rate": 3.5682372055239646e-06, + "loss": 1.8279, + "step": 995 + }, + { + "epoch": 1.0915068493150686, + "grad_norm": 2.2056329250335693, + "learning_rate": 3.566206336311942e-06, + "loss": 1.7787, + "step": 996 + }, + { + "epoch": 1.0926027397260274, + "grad_norm": 2.1159322261810303, + "learning_rate": 3.564175467099919e-06, + "loss": 1.7528, + "step": 997 + }, + { + "epoch": 1.0936986301369862, + "grad_norm": 2.1371147632598877, + "learning_rate": 3.5621445978878964e-06, + "loss": 1.8399, + "step": 998 + }, + { + "epoch": 1.0947945205479452, + "grad_norm": 1.974733829498291, + "learning_rate": 3.5601137286758737e-06, + "loss": 1.7965, + "step": 999 + }, + { + "epoch": 1.095890410958904, + "grad_norm": 1.998443365097046, + "learning_rate": 3.5580828594638505e-06, + "loss": 1.8306, + "step": 1000 + }, + { + "epoch": 1.095890410958904, + "eval_loss": 2.007416248321533, + "eval_runtime": 37.6113, + "eval_samples_per_second": 26.588, + "eval_steps_per_second": 1.117, + "step": 1000 + }, + { + "epoch": 1.096986301369863, + "grad_norm": 2.071329116821289, + "learning_rate": 3.5560519902518282e-06, + "loss": 1.8539, + "step": 1001 + }, + { + "epoch": 1.098082191780822, + "grad_norm": 1.9598201513290405, + "learning_rate": 3.554021121039805e-06, + "loss": 1.8115, + "step": 1002 + }, + { + "epoch": 1.099178082191781, + "grad_norm": 2.10245943069458, + "learning_rate": 3.551990251827783e-06, + "loss": 1.7842, + "step": 1003 + }, + { + "epoch": 1.1002739726027397, + "grad_norm": 2.047262191772461, + "learning_rate": 3.5499593826157596e-06, + "loss": 1.8059, + "step": 1004 + }, + { + "epoch": 1.1013698630136985, + "grad_norm": 2.0314691066741943, + "learning_rate": 3.5479285134037373e-06, + "loss": 1.7853, + "step": 1005 + }, + { + "epoch": 1.1024657534246576, + "grad_norm": 2.030719757080078, + "learning_rate": 3.545897644191714e-06, + "loss": 1.7767, + "step": 1006 + }, + { + "epoch": 1.1035616438356164, + "grad_norm": 2.0710442066192627, + "learning_rate": 3.543866774979692e-06, + "loss": 1.8205, + "step": 1007 + }, + { + "epoch": 1.1046575342465754, + "grad_norm": 2.068744659423828, + "learning_rate": 3.5418359057676687e-06, + "loss": 1.7831, + "step": 1008 + }, + { + "epoch": 1.1057534246575342, + "grad_norm": 2.0644986629486084, + "learning_rate": 3.5398050365556464e-06, + "loss": 1.8345, + "step": 1009 + }, + { + "epoch": 1.106849315068493, + "grad_norm": 1.9313405752182007, + "learning_rate": 3.5377741673436233e-06, + "loss": 1.7872, + "step": 1010 + }, + { + "epoch": 1.106849315068493, + "eval_loss": 2.00844144821167, + "eval_runtime": 37.2341, + "eval_samples_per_second": 26.857, + "eval_steps_per_second": 1.128, + "step": 1010 + }, + { + "epoch": 1.107945205479452, + "grad_norm": 1.923250675201416, + "learning_rate": 3.535743298131601e-06, + "loss": 1.8396, + "step": 1011 + }, + { + "epoch": 1.1090410958904109, + "grad_norm": 1.8591458797454834, + "learning_rate": 3.533712428919578e-06, + "loss": 1.8578, + "step": 1012 + }, + { + "epoch": 1.11013698630137, + "grad_norm": 2.010024070739746, + "learning_rate": 3.5316815597075547e-06, + "loss": 1.7891, + "step": 1013 + }, + { + "epoch": 1.1112328767123287, + "grad_norm": 2.099515438079834, + "learning_rate": 3.5296506904955324e-06, + "loss": 1.8163, + "step": 1014 + }, + { + "epoch": 1.1123287671232878, + "grad_norm": 1.9067803621292114, + "learning_rate": 3.5276198212835093e-06, + "loss": 1.8487, + "step": 1015 + }, + { + "epoch": 1.1134246575342466, + "grad_norm": 1.9962882995605469, + "learning_rate": 3.525588952071487e-06, + "loss": 1.8669, + "step": 1016 + }, + { + "epoch": 1.1145205479452054, + "grad_norm": 2.0068392753601074, + "learning_rate": 3.523558082859464e-06, + "loss": 1.8311, + "step": 1017 + }, + { + "epoch": 1.1156164383561644, + "grad_norm": 2.0002613067626953, + "learning_rate": 3.5215272136474415e-06, + "loss": 1.7967, + "step": 1018 + }, + { + "epoch": 1.1167123287671232, + "grad_norm": 2.0696845054626465, + "learning_rate": 3.5194963444354184e-06, + "loss": 1.7983, + "step": 1019 + }, + { + "epoch": 1.1178082191780823, + "grad_norm": 2.061471939086914, + "learning_rate": 3.517465475223396e-06, + "loss": 1.7636, + "step": 1020 + }, + { + "epoch": 1.1178082191780823, + "eval_loss": 2.012301445007324, + "eval_runtime": 37.416, + "eval_samples_per_second": 26.727, + "eval_steps_per_second": 1.123, + "step": 1020 + }, + { + "epoch": 1.118904109589041, + "grad_norm": 2.104631185531616, + "learning_rate": 3.515434606011373e-06, + "loss": 1.8414, + "step": 1021 + }, + { + "epoch": 1.12, + "grad_norm": 2.0233333110809326, + "learning_rate": 3.5134037367993506e-06, + "loss": 1.9032, + "step": 1022 + }, + { + "epoch": 1.121095890410959, + "grad_norm": 2.0639472007751465, + "learning_rate": 3.5113728675873275e-06, + "loss": 1.8281, + "step": 1023 + }, + { + "epoch": 1.1221917808219177, + "grad_norm": 2.187289237976074, + "learning_rate": 3.5093419983753047e-06, + "loss": 1.8235, + "step": 1024 + }, + { + "epoch": 1.1232876712328768, + "grad_norm": 2.0636637210845947, + "learning_rate": 3.507311129163282e-06, + "loss": 1.7585, + "step": 1025 + }, + { + "epoch": 1.1243835616438356, + "grad_norm": 2.222682237625122, + "learning_rate": 3.5052802599512593e-06, + "loss": 1.8042, + "step": 1026 + }, + { + "epoch": 1.1254794520547946, + "grad_norm": 2.0362460613250732, + "learning_rate": 3.5032493907392366e-06, + "loss": 1.8237, + "step": 1027 + }, + { + "epoch": 1.1265753424657534, + "grad_norm": 2.0469627380371094, + "learning_rate": 3.501218521527214e-06, + "loss": 1.7503, + "step": 1028 + }, + { + "epoch": 1.1276712328767124, + "grad_norm": 2.098461866378784, + "learning_rate": 3.499187652315191e-06, + "loss": 1.8115, + "step": 1029 + }, + { + "epoch": 1.1287671232876713, + "grad_norm": 2.0707895755767822, + "learning_rate": 3.4971567831031684e-06, + "loss": 1.8468, + "step": 1030 + }, + { + "epoch": 1.1287671232876713, + "eval_loss": 2.0071005821228027, + "eval_runtime": 37.5016, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 1030 + }, + { + "epoch": 1.12986301369863, + "grad_norm": 2.03425669670105, + "learning_rate": 3.4951259138911457e-06, + "loss": 1.755, + "step": 1031 + }, + { + "epoch": 1.130958904109589, + "grad_norm": 1.9484025239944458, + "learning_rate": 3.493095044679123e-06, + "loss": 1.8724, + "step": 1032 + }, + { + "epoch": 1.132054794520548, + "grad_norm": 2.1967852115631104, + "learning_rate": 3.4910641754671002e-06, + "loss": 1.7106, + "step": 1033 + }, + { + "epoch": 1.133150684931507, + "grad_norm": 2.200246810913086, + "learning_rate": 3.4890333062550775e-06, + "loss": 1.8387, + "step": 1034 + }, + { + "epoch": 1.1342465753424658, + "grad_norm": 2.132086753845215, + "learning_rate": 3.4870024370430548e-06, + "loss": 1.7348, + "step": 1035 + }, + { + "epoch": 1.1353424657534246, + "grad_norm": 2.009600877761841, + "learning_rate": 3.484971567831032e-06, + "loss": 1.7827, + "step": 1036 + }, + { + "epoch": 1.1364383561643836, + "grad_norm": 1.9694634675979614, + "learning_rate": 3.482940698619009e-06, + "loss": 1.8486, + "step": 1037 + }, + { + "epoch": 1.1375342465753424, + "grad_norm": 2.0271873474121094, + "learning_rate": 3.4809098294069866e-06, + "loss": 1.8069, + "step": 1038 + }, + { + "epoch": 1.1386301369863014, + "grad_norm": 2.1308176517486572, + "learning_rate": 3.4788789601949634e-06, + "loss": 1.8254, + "step": 1039 + }, + { + "epoch": 1.1397260273972603, + "grad_norm": 1.9240211248397827, + "learning_rate": 3.476848090982941e-06, + "loss": 1.7893, + "step": 1040 + }, + { + "epoch": 1.1397260273972603, + "eval_loss": 2.0071732997894287, + "eval_runtime": 37.5209, + "eval_samples_per_second": 26.652, + "eval_steps_per_second": 1.119, + "step": 1040 + }, + { + "epoch": 1.140821917808219, + "grad_norm": 2.200873613357544, + "learning_rate": 3.474817221770918e-06, + "loss": 1.9016, + "step": 1041 + }, + { + "epoch": 1.141917808219178, + "grad_norm": 2.1640889644622803, + "learning_rate": 3.4727863525588957e-06, + "loss": 1.8472, + "step": 1042 + }, + { + "epoch": 1.143013698630137, + "grad_norm": 2.2426443099975586, + "learning_rate": 3.4707554833468725e-06, + "loss": 1.8088, + "step": 1043 + }, + { + "epoch": 1.144109589041096, + "grad_norm": 2.2531626224517822, + "learning_rate": 3.4687246141348502e-06, + "loss": 1.7773, + "step": 1044 + }, + { + "epoch": 1.1452054794520548, + "grad_norm": 2.047778606414795, + "learning_rate": 3.466693744922827e-06, + "loss": 1.7964, + "step": 1045 + }, + { + "epoch": 1.1463013698630138, + "grad_norm": 2.282496690750122, + "learning_rate": 3.464662875710805e-06, + "loss": 1.8217, + "step": 1046 + }, + { + "epoch": 1.1473972602739726, + "grad_norm": 2.09999418258667, + "learning_rate": 3.4626320064987817e-06, + "loss": 1.8071, + "step": 1047 + }, + { + "epoch": 1.1484931506849314, + "grad_norm": 2.093827247619629, + "learning_rate": 3.4606011372867594e-06, + "loss": 1.8257, + "step": 1048 + }, + { + "epoch": 1.1495890410958904, + "grad_norm": 2.7829785346984863, + "learning_rate": 3.458570268074736e-06, + "loss": 1.8017, + "step": 1049 + }, + { + "epoch": 1.1506849315068493, + "grad_norm": 1.8894433975219727, + "learning_rate": 3.4565393988627135e-06, + "loss": 1.7817, + "step": 1050 + }, + { + "epoch": 1.1506849315068493, + "eval_loss": 2.004091262817383, + "eval_runtime": 37.5142, + "eval_samples_per_second": 26.657, + "eval_steps_per_second": 1.12, + "step": 1050 + }, + { + "epoch": 1.1517808219178083, + "grad_norm": 2.240211009979248, + "learning_rate": 3.4545085296506908e-06, + "loss": 1.7534, + "step": 1051 + }, + { + "epoch": 1.152876712328767, + "grad_norm": 2.59191632270813, + "learning_rate": 3.452477660438668e-06, + "loss": 1.8491, + "step": 1052 + }, + { + "epoch": 1.1539726027397261, + "grad_norm": 2.0091731548309326, + "learning_rate": 3.4504467912266453e-06, + "loss": 1.7488, + "step": 1053 + }, + { + "epoch": 1.155068493150685, + "grad_norm": 2.162797689437866, + "learning_rate": 3.4484159220146226e-06, + "loss": 1.8128, + "step": 1054 + }, + { + "epoch": 1.1561643835616437, + "grad_norm": 2.260213851928711, + "learning_rate": 3.4463850528026e-06, + "loss": 1.7934, + "step": 1055 + }, + { + "epoch": 1.1572602739726028, + "grad_norm": 1.951391339302063, + "learning_rate": 3.444354183590577e-06, + "loss": 1.8248, + "step": 1056 + }, + { + "epoch": 1.1583561643835616, + "grad_norm": 2.2070186138153076, + "learning_rate": 3.4423233143785544e-06, + "loss": 1.7604, + "step": 1057 + }, + { + "epoch": 1.1594520547945206, + "grad_norm": 2.104912042617798, + "learning_rate": 3.4402924451665317e-06, + "loss": 1.7313, + "step": 1058 + }, + { + "epoch": 1.1605479452054794, + "grad_norm": 2.0583512783050537, + "learning_rate": 3.438261575954509e-06, + "loss": 1.8195, + "step": 1059 + }, + { + "epoch": 1.1616438356164385, + "grad_norm": 2.061549663543701, + "learning_rate": 3.4362307067424862e-06, + "loss": 1.8554, + "step": 1060 + }, + { + "epoch": 1.1616438356164385, + "eval_loss": 2.007450819015503, + "eval_runtime": 37.5191, + "eval_samples_per_second": 26.653, + "eval_steps_per_second": 1.119, + "step": 1060 + }, + { + "epoch": 1.1627397260273973, + "grad_norm": 2.5224766731262207, + "learning_rate": 3.434199837530463e-06, + "loss": 1.823, + "step": 1061 + }, + { + "epoch": 1.163835616438356, + "grad_norm": 2.2069733142852783, + "learning_rate": 3.432168968318441e-06, + "loss": 1.7493, + "step": 1062 + }, + { + "epoch": 1.1649315068493151, + "grad_norm": 2.0238699913024902, + "learning_rate": 3.4301380991064176e-06, + "loss": 1.8297, + "step": 1063 + }, + { + "epoch": 1.166027397260274, + "grad_norm": 2.5067358016967773, + "learning_rate": 3.4281072298943953e-06, + "loss": 1.8648, + "step": 1064 + }, + { + "epoch": 1.167123287671233, + "grad_norm": 1.9712344408035278, + "learning_rate": 3.426076360682372e-06, + "loss": 1.865, + "step": 1065 + }, + { + "epoch": 1.1682191780821918, + "grad_norm": 2.183479070663452, + "learning_rate": 3.42404549147035e-06, + "loss": 1.8033, + "step": 1066 + }, + { + "epoch": 1.1693150684931506, + "grad_norm": 2.1122567653656006, + "learning_rate": 3.4220146222583267e-06, + "loss": 1.7604, + "step": 1067 + }, + { + "epoch": 1.1704109589041096, + "grad_norm": 1.962809443473816, + "learning_rate": 3.4199837530463044e-06, + "loss": 1.8005, + "step": 1068 + }, + { + "epoch": 1.1715068493150684, + "grad_norm": 2.1723763942718506, + "learning_rate": 3.4179528838342813e-06, + "loss": 1.8948, + "step": 1069 + }, + { + "epoch": 1.1726027397260275, + "grad_norm": 2.20245361328125, + "learning_rate": 3.415922014622259e-06, + "loss": 1.8458, + "step": 1070 + }, + { + "epoch": 1.1726027397260275, + "eval_loss": 2.0126123428344727, + "eval_runtime": 37.5473, + "eval_samples_per_second": 26.633, + "eval_steps_per_second": 1.119, + "step": 1070 + }, + { + "epoch": 1.1736986301369863, + "grad_norm": 2.070758104324341, + "learning_rate": 3.413891145410236e-06, + "loss": 1.8374, + "step": 1071 + }, + { + "epoch": 1.174794520547945, + "grad_norm": 2.0977511405944824, + "learning_rate": 3.4118602761982135e-06, + "loss": 1.8631, + "step": 1072 + }, + { + "epoch": 1.1758904109589041, + "grad_norm": 2.0646073818206787, + "learning_rate": 3.4098294069861904e-06, + "loss": 1.8153, + "step": 1073 + }, + { + "epoch": 1.176986301369863, + "grad_norm": 2.1012344360351562, + "learning_rate": 3.4077985377741673e-06, + "loss": 1.7812, + "step": 1074 + }, + { + "epoch": 1.178082191780822, + "grad_norm": 1.999519944190979, + "learning_rate": 3.405767668562145e-06, + "loss": 1.866, + "step": 1075 + }, + { + "epoch": 1.1791780821917808, + "grad_norm": 2.0217132568359375, + "learning_rate": 3.403736799350122e-06, + "loss": 1.847, + "step": 1076 + }, + { + "epoch": 1.1802739726027398, + "grad_norm": 2.030902862548828, + "learning_rate": 3.4017059301380995e-06, + "loss": 1.7724, + "step": 1077 + }, + { + "epoch": 1.1813698630136986, + "grad_norm": 2.22278094291687, + "learning_rate": 3.3996750609260764e-06, + "loss": 1.893, + "step": 1078 + }, + { + "epoch": 1.1824657534246574, + "grad_norm": 2.078679323196411, + "learning_rate": 3.397644191714054e-06, + "loss": 1.7899, + "step": 1079 + }, + { + "epoch": 1.1835616438356165, + "grad_norm": 2.0326967239379883, + "learning_rate": 3.395613322502031e-06, + "loss": 1.7932, + "step": 1080 + }, + { + "epoch": 1.1835616438356165, + "eval_loss": 2.0050764083862305, + "eval_runtime": 37.5549, + "eval_samples_per_second": 26.628, + "eval_steps_per_second": 1.118, + "step": 1080 + }, + { + "epoch": 1.1846575342465753, + "grad_norm": 2.1296298503875732, + "learning_rate": 3.3935824532900086e-06, + "loss": 1.7671, + "step": 1081 + }, + { + "epoch": 1.1857534246575343, + "grad_norm": 2.0533270835876465, + "learning_rate": 3.3915515840779855e-06, + "loss": 1.8058, + "step": 1082 + }, + { + "epoch": 1.1868493150684931, + "grad_norm": 2.0194473266601562, + "learning_rate": 3.389520714865963e-06, + "loss": 1.8066, + "step": 1083 + }, + { + "epoch": 1.1879452054794521, + "grad_norm": 2.052131175994873, + "learning_rate": 3.38748984565394e-06, + "loss": 1.8345, + "step": 1084 + }, + { + "epoch": 1.189041095890411, + "grad_norm": 2.246558666229248, + "learning_rate": 3.3854589764419177e-06, + "loss": 1.7786, + "step": 1085 + }, + { + "epoch": 1.1901369863013698, + "grad_norm": 1.9524049758911133, + "learning_rate": 3.3834281072298946e-06, + "loss": 1.7955, + "step": 1086 + }, + { + "epoch": 1.1912328767123288, + "grad_norm": 2.156313180923462, + "learning_rate": 3.381397238017872e-06, + "loss": 1.7714, + "step": 1087 + }, + { + "epoch": 1.1923287671232876, + "grad_norm": 2.26092267036438, + "learning_rate": 3.379366368805849e-06, + "loss": 1.8689, + "step": 1088 + }, + { + "epoch": 1.1934246575342466, + "grad_norm": 2.0419435501098633, + "learning_rate": 3.3773354995938264e-06, + "loss": 1.8186, + "step": 1089 + }, + { + "epoch": 1.1945205479452055, + "grad_norm": 2.0528366565704346, + "learning_rate": 3.3753046303818037e-06, + "loss": 1.8072, + "step": 1090 + }, + { + "epoch": 1.1945205479452055, + "eval_loss": 2.0051960945129395, + "eval_runtime": 37.5423, + "eval_samples_per_second": 26.637, + "eval_steps_per_second": 1.119, + "step": 1090 + }, + { + "epoch": 1.1956164383561645, + "grad_norm": 2.2391040325164795, + "learning_rate": 3.373273761169781e-06, + "loss": 1.8021, + "step": 1091 + }, + { + "epoch": 1.1967123287671233, + "grad_norm": 2.1061277389526367, + "learning_rate": 3.3712428919577582e-06, + "loss": 1.7879, + "step": 1092 + }, + { + "epoch": 1.1978082191780821, + "grad_norm": 1.9935376644134521, + "learning_rate": 3.3692120227457355e-06, + "loss": 1.8273, + "step": 1093 + }, + { + "epoch": 1.1989041095890411, + "grad_norm": 2.112804889678955, + "learning_rate": 3.3671811535337128e-06, + "loss": 1.7549, + "step": 1094 + }, + { + "epoch": 1.2, + "grad_norm": 2.1300928592681885, + "learning_rate": 3.36515028432169e-06, + "loss": 1.7871, + "step": 1095 + }, + { + "epoch": 1.201095890410959, + "grad_norm": 1.9645880460739136, + "learning_rate": 3.3631194151096673e-06, + "loss": 1.8271, + "step": 1096 + }, + { + "epoch": 1.2021917808219178, + "grad_norm": 2.0097038745880127, + "learning_rate": 3.3610885458976446e-06, + "loss": 1.8528, + "step": 1097 + }, + { + "epoch": 1.2032876712328768, + "grad_norm": 1.903214931488037, + "learning_rate": 3.359057676685622e-06, + "loss": 1.8434, + "step": 1098 + }, + { + "epoch": 1.2043835616438356, + "grad_norm": 1.9981344938278198, + "learning_rate": 3.357026807473599e-06, + "loss": 1.7319, + "step": 1099 + }, + { + "epoch": 1.2054794520547945, + "grad_norm": 2.1177191734313965, + "learning_rate": 3.354995938261576e-06, + "loss": 1.7807, + "step": 1100 + }, + { + "epoch": 1.2054794520547945, + "eval_loss": 2.0032761096954346, + "eval_runtime": 37.5645, + "eval_samples_per_second": 26.621, + "eval_steps_per_second": 1.118, + "step": 1100 + }, + { + "epoch": 1.2065753424657535, + "grad_norm": 2.1452767848968506, + "learning_rate": 3.3529650690495537e-06, + "loss": 1.7437, + "step": 1101 + }, + { + "epoch": 1.2076712328767123, + "grad_norm": 2.1550211906433105, + "learning_rate": 3.3509341998375305e-06, + "loss": 1.7849, + "step": 1102 + }, + { + "epoch": 1.208767123287671, + "grad_norm": 2.1381752490997314, + "learning_rate": 3.3489033306255082e-06, + "loss": 1.8142, + "step": 1103 + }, + { + "epoch": 1.2098630136986301, + "grad_norm": 2.1935431957244873, + "learning_rate": 3.346872461413485e-06, + "loss": 1.78, + "step": 1104 + }, + { + "epoch": 1.210958904109589, + "grad_norm": 1.9361294507980347, + "learning_rate": 3.344841592201463e-06, + "loss": 1.7642, + "step": 1105 + }, + { + "epoch": 1.212054794520548, + "grad_norm": 1.9446313381195068, + "learning_rate": 3.3428107229894397e-06, + "loss": 1.7922, + "step": 1106 + }, + { + "epoch": 1.2131506849315068, + "grad_norm": 2.148961305618286, + "learning_rate": 3.3407798537774174e-06, + "loss": 1.8063, + "step": 1107 + }, + { + "epoch": 1.2142465753424658, + "grad_norm": 2.1928062438964844, + "learning_rate": 3.338748984565394e-06, + "loss": 1.8537, + "step": 1108 + }, + { + "epoch": 1.2153424657534246, + "grad_norm": 2.140801191329956, + "learning_rate": 3.336718115353372e-06, + "loss": 1.8624, + "step": 1109 + }, + { + "epoch": 1.2164383561643834, + "grad_norm": 2.046963691711426, + "learning_rate": 3.3346872461413488e-06, + "loss": 1.7829, + "step": 1110 + }, + { + "epoch": 1.2164383561643834, + "eval_loss": 1.9971742630004883, + "eval_runtime": 37.2848, + "eval_samples_per_second": 26.821, + "eval_steps_per_second": 1.126, + "step": 1110 + }, + { + "epoch": 1.2175342465753425, + "grad_norm": 1.8843704462051392, + "learning_rate": 3.3326563769293256e-06, + "loss": 1.7693, + "step": 1111 + }, + { + "epoch": 1.2186301369863013, + "grad_norm": 2.0071821212768555, + "learning_rate": 3.3306255077173033e-06, + "loss": 1.7995, + "step": 1112 + }, + { + "epoch": 1.2197260273972603, + "grad_norm": 1.887171983718872, + "learning_rate": 3.32859463850528e-06, + "loss": 1.8129, + "step": 1113 + }, + { + "epoch": 1.2208219178082191, + "grad_norm": 2.0528452396392822, + "learning_rate": 3.326563769293258e-06, + "loss": 1.7328, + "step": 1114 + }, + { + "epoch": 1.2219178082191782, + "grad_norm": 2.040501594543457, + "learning_rate": 3.3245329000812347e-06, + "loss": 1.7745, + "step": 1115 + }, + { + "epoch": 1.223013698630137, + "grad_norm": 2.36647367477417, + "learning_rate": 3.3225020308692124e-06, + "loss": 1.7952, + "step": 1116 + }, + { + "epoch": 1.2241095890410958, + "grad_norm": 2.126403331756592, + "learning_rate": 3.3204711616571893e-06, + "loss": 1.7465, + "step": 1117 + }, + { + "epoch": 1.2252054794520548, + "grad_norm": 2.0110483169555664, + "learning_rate": 3.318440292445167e-06, + "loss": 1.803, + "step": 1118 + }, + { + "epoch": 1.2263013698630136, + "grad_norm": 2.1778430938720703, + "learning_rate": 3.316409423233144e-06, + "loss": 1.7654, + "step": 1119 + }, + { + "epoch": 1.2273972602739727, + "grad_norm": 2.0968027114868164, + "learning_rate": 3.3143785540211215e-06, + "loss": 1.8546, + "step": 1120 + }, + { + "epoch": 1.2273972602739727, + "eval_loss": 1.998443365097046, + "eval_runtime": 37.4687, + "eval_samples_per_second": 26.689, + "eval_steps_per_second": 1.121, + "step": 1120 + }, + { + "epoch": 1.2284931506849315, + "grad_norm": 1.8872292041778564, + "learning_rate": 3.3123476848090984e-06, + "loss": 1.8028, + "step": 1121 + }, + { + "epoch": 1.2295890410958905, + "grad_norm": 1.944091558456421, + "learning_rate": 3.310316815597076e-06, + "loss": 1.8075, + "step": 1122 + }, + { + "epoch": 1.2306849315068493, + "grad_norm": 2.0419323444366455, + "learning_rate": 3.308285946385053e-06, + "loss": 1.8905, + "step": 1123 + }, + { + "epoch": 1.2317808219178081, + "grad_norm": 1.9407243728637695, + "learning_rate": 3.30625507717303e-06, + "loss": 1.7385, + "step": 1124 + }, + { + "epoch": 1.2328767123287672, + "grad_norm": 2.062077045440674, + "learning_rate": 3.3042242079610075e-06, + "loss": 1.8278, + "step": 1125 + }, + { + "epoch": 1.233972602739726, + "grad_norm": 1.9379059076309204, + "learning_rate": 3.3021933387489847e-06, + "loss": 1.8506, + "step": 1126 + }, + { + "epoch": 1.235068493150685, + "grad_norm": 1.8854087591171265, + "learning_rate": 3.300162469536962e-06, + "loss": 1.7758, + "step": 1127 + }, + { + "epoch": 1.2361643835616438, + "grad_norm": 1.9997806549072266, + "learning_rate": 3.2981316003249393e-06, + "loss": 1.8534, + "step": 1128 + }, + { + "epoch": 1.2372602739726029, + "grad_norm": 2.0156726837158203, + "learning_rate": 3.2961007311129166e-06, + "loss": 1.8255, + "step": 1129 + }, + { + "epoch": 1.2383561643835617, + "grad_norm": 1.9987738132476807, + "learning_rate": 3.294069861900894e-06, + "loss": 1.8838, + "step": 1130 + }, + { + "epoch": 1.2383561643835617, + "eval_loss": 1.9990321397781372, + "eval_runtime": 37.5765, + "eval_samples_per_second": 26.612, + "eval_steps_per_second": 1.118, + "step": 1130 + }, + { + "epoch": 1.2394520547945205, + "grad_norm": 2.068734645843506, + "learning_rate": 3.292038992688871e-06, + "loss": 1.8151, + "step": 1131 + }, + { + "epoch": 1.2405479452054795, + "grad_norm": 2.1235573291778564, + "learning_rate": 3.2900081234768484e-06, + "loss": 1.8229, + "step": 1132 + }, + { + "epoch": 1.2416438356164383, + "grad_norm": 1.8630824089050293, + "learning_rate": 3.2879772542648257e-06, + "loss": 1.86, + "step": 1133 + }, + { + "epoch": 1.2427397260273974, + "grad_norm": 2.0974395275115967, + "learning_rate": 3.285946385052803e-06, + "loss": 1.7366, + "step": 1134 + }, + { + "epoch": 1.2438356164383562, + "grad_norm": 2.13661789894104, + "learning_rate": 3.2839155158407802e-06, + "loss": 1.8313, + "step": 1135 + }, + { + "epoch": 1.244931506849315, + "grad_norm": 1.9267613887786865, + "learning_rate": 3.2818846466287575e-06, + "loss": 1.7315, + "step": 1136 + }, + { + "epoch": 1.246027397260274, + "grad_norm": 2.1487739086151123, + "learning_rate": 3.2798537774167344e-06, + "loss": 1.8709, + "step": 1137 + }, + { + "epoch": 1.2471232876712328, + "grad_norm": 1.9620766639709473, + "learning_rate": 3.277822908204712e-06, + "loss": 1.8008, + "step": 1138 + }, + { + "epoch": 1.2482191780821918, + "grad_norm": 2.388493299484253, + "learning_rate": 3.275792038992689e-06, + "loss": 1.7527, + "step": 1139 + }, + { + "epoch": 1.2493150684931507, + "grad_norm": 1.9988473653793335, + "learning_rate": 3.2737611697806666e-06, + "loss": 1.8082, + "step": 1140 + }, + { + "epoch": 1.2493150684931507, + "eval_loss": 1.9974714517593384, + "eval_runtime": 37.6611, + "eval_samples_per_second": 26.553, + "eval_steps_per_second": 1.115, + "step": 1140 + }, + { + "epoch": 1.2504109589041095, + "grad_norm": 1.950793743133545, + "learning_rate": 3.2717303005686435e-06, + "loss": 1.7694, + "step": 1141 + }, + { + "epoch": 1.2515068493150685, + "grad_norm": 2.2984821796417236, + "learning_rate": 3.269699431356621e-06, + "loss": 1.8068, + "step": 1142 + }, + { + "epoch": 1.2526027397260273, + "grad_norm": 2.0106565952301025, + "learning_rate": 3.267668562144598e-06, + "loss": 1.7931, + "step": 1143 + }, + { + "epoch": 1.2536986301369863, + "grad_norm": 2.0408499240875244, + "learning_rate": 3.2656376929325757e-06, + "loss": 1.7617, + "step": 1144 + }, + { + "epoch": 1.2547945205479452, + "grad_norm": 2.2353122234344482, + "learning_rate": 3.2636068237205526e-06, + "loss": 1.828, + "step": 1145 + }, + { + "epoch": 1.2558904109589042, + "grad_norm": 1.8609519004821777, + "learning_rate": 3.2615759545085303e-06, + "loss": 1.7909, + "step": 1146 + }, + { + "epoch": 1.256986301369863, + "grad_norm": 2.4187569618225098, + "learning_rate": 3.259545085296507e-06, + "loss": 1.8533, + "step": 1147 + }, + { + "epoch": 1.2580821917808218, + "grad_norm": 2.0719733238220215, + "learning_rate": 3.257514216084484e-06, + "loss": 1.7473, + "step": 1148 + }, + { + "epoch": 1.2591780821917808, + "grad_norm": 1.9666911363601685, + "learning_rate": 3.2554833468724617e-06, + "loss": 1.8545, + "step": 1149 + }, + { + "epoch": 1.2602739726027397, + "grad_norm": 2.485783338546753, + "learning_rate": 3.2534524776604385e-06, + "loss": 1.8176, + "step": 1150 + }, + { + "epoch": 1.2602739726027397, + "eval_loss": 1.997860074043274, + "eval_runtime": 37.6063, + "eval_samples_per_second": 26.591, + "eval_steps_per_second": 1.117, + "step": 1150 + }, + { + "epoch": 1.2613698630136987, + "grad_norm": 1.9877129793167114, + "learning_rate": 3.2514216084484162e-06, + "loss": 1.79, + "step": 1151 + }, + { + "epoch": 1.2624657534246575, + "grad_norm": 1.957518219947815, + "learning_rate": 3.249390739236393e-06, + "loss": 1.8203, + "step": 1152 + }, + { + "epoch": 1.2635616438356165, + "grad_norm": 1.995038390159607, + "learning_rate": 3.2473598700243708e-06, + "loss": 1.7729, + "step": 1153 + }, + { + "epoch": 1.2646575342465753, + "grad_norm": 1.9420701265335083, + "learning_rate": 3.2453290008123476e-06, + "loss": 1.8061, + "step": 1154 + }, + { + "epoch": 1.2657534246575342, + "grad_norm": 2.019300699234009, + "learning_rate": 3.2432981316003253e-06, + "loss": 1.8069, + "step": 1155 + }, + { + "epoch": 1.2668493150684932, + "grad_norm": 2.097673177719116, + "learning_rate": 3.241267262388302e-06, + "loss": 1.7651, + "step": 1156 + }, + { + "epoch": 1.267945205479452, + "grad_norm": 2.0599231719970703, + "learning_rate": 3.23923639317628e-06, + "loss": 1.8027, + "step": 1157 + }, + { + "epoch": 1.269041095890411, + "grad_norm": 2.0775444507598877, + "learning_rate": 3.2372055239642567e-06, + "loss": 1.7515, + "step": 1158 + }, + { + "epoch": 1.2701369863013698, + "grad_norm": 2.104642868041992, + "learning_rate": 3.2351746547522344e-06, + "loss": 1.822, + "step": 1159 + }, + { + "epoch": 1.2712328767123289, + "grad_norm": 2.1397576332092285, + "learning_rate": 3.2331437855402113e-06, + "loss": 1.8646, + "step": 1160 + }, + { + "epoch": 1.2712328767123289, + "eval_loss": 1.9921350479125977, + "eval_runtime": 37.6677, + "eval_samples_per_second": 26.548, + "eval_steps_per_second": 1.115, + "step": 1160 + }, + { + "epoch": 1.2723287671232877, + "grad_norm": 1.9359031915664673, + "learning_rate": 3.2311129163281885e-06, + "loss": 1.7583, + "step": 1161 + }, + { + "epoch": 1.2734246575342465, + "grad_norm": 1.988430142402649, + "learning_rate": 3.229082047116166e-06, + "loss": 1.8365, + "step": 1162 + }, + { + "epoch": 1.2745205479452055, + "grad_norm": 2.086852550506592, + "learning_rate": 3.227051177904143e-06, + "loss": 1.8024, + "step": 1163 + }, + { + "epoch": 1.2756164383561643, + "grad_norm": 1.9968465566635132, + "learning_rate": 3.2250203086921204e-06, + "loss": 1.8064, + "step": 1164 + }, + { + "epoch": 1.2767123287671232, + "grad_norm": 2.245166301727295, + "learning_rate": 3.2229894394800977e-06, + "loss": 1.8338, + "step": 1165 + }, + { + "epoch": 1.2778082191780822, + "grad_norm": 2.014070510864258, + "learning_rate": 3.220958570268075e-06, + "loss": 1.7229, + "step": 1166 + }, + { + "epoch": 1.2789041095890412, + "grad_norm": 2.284050464630127, + "learning_rate": 3.218927701056052e-06, + "loss": 1.7914, + "step": 1167 + }, + { + "epoch": 1.28, + "grad_norm": 2.0785140991210938, + "learning_rate": 3.2168968318440295e-06, + "loss": 1.7083, + "step": 1168 + }, + { + "epoch": 1.2810958904109588, + "grad_norm": 1.989413857460022, + "learning_rate": 3.2148659626320068e-06, + "loss": 1.8597, + "step": 1169 + }, + { + "epoch": 1.2821917808219179, + "grad_norm": 2.0537919998168945, + "learning_rate": 3.212835093419984e-06, + "loss": 1.7563, + "step": 1170 + }, + { + "epoch": 1.2821917808219179, + "eval_loss": 1.9913713932037354, + "eval_runtime": 37.6191, + "eval_samples_per_second": 26.582, + "eval_steps_per_second": 1.116, + "step": 1170 + }, + { + "epoch": 1.2832876712328767, + "grad_norm": 2.045679807662964, + "learning_rate": 3.2108042242079613e-06, + "loss": 1.7789, + "step": 1171 + }, + { + "epoch": 1.2843835616438355, + "grad_norm": 1.9195408821105957, + "learning_rate": 3.2087733549959386e-06, + "loss": 1.8514, + "step": 1172 + }, + { + "epoch": 1.2854794520547945, + "grad_norm": 2.1682701110839844, + "learning_rate": 3.206742485783916e-06, + "loss": 1.738, + "step": 1173 + }, + { + "epoch": 1.2865753424657536, + "grad_norm": 2.0337114334106445, + "learning_rate": 3.2047116165718927e-06, + "loss": 1.7598, + "step": 1174 + }, + { + "epoch": 1.2876712328767124, + "grad_norm": 1.9387762546539307, + "learning_rate": 3.2026807473598704e-06, + "loss": 1.8081, + "step": 1175 + }, + { + "epoch": 1.2887671232876712, + "grad_norm": 2.0097672939300537, + "learning_rate": 3.2006498781478473e-06, + "loss": 1.816, + "step": 1176 + }, + { + "epoch": 1.2898630136986302, + "grad_norm": 2.167109489440918, + "learning_rate": 3.198619008935825e-06, + "loss": 1.7523, + "step": 1177 + }, + { + "epoch": 1.290958904109589, + "grad_norm": 2.0411338806152344, + "learning_rate": 3.196588139723802e-06, + "loss": 1.7932, + "step": 1178 + }, + { + "epoch": 1.2920547945205478, + "grad_norm": 2.0663340091705322, + "learning_rate": 3.1945572705117795e-06, + "loss": 1.8088, + "step": 1179 + }, + { + "epoch": 1.2931506849315069, + "grad_norm": 2.3893980979919434, + "learning_rate": 3.1925264012997564e-06, + "loss": 1.8125, + "step": 1180 + }, + { + "epoch": 1.2931506849315069, + "eval_loss": 1.9927020072937012, + "eval_runtime": 37.6397, + "eval_samples_per_second": 26.568, + "eval_steps_per_second": 1.116, + "step": 1180 + }, + { + "epoch": 1.2942465753424657, + "grad_norm": 2.092419147491455, + "learning_rate": 3.190495532087734e-06, + "loss": 1.8122, + "step": 1181 + }, + { + "epoch": 1.2953424657534247, + "grad_norm": 2.1075375080108643, + "learning_rate": 3.188464662875711e-06, + "loss": 1.8671, + "step": 1182 + }, + { + "epoch": 1.2964383561643835, + "grad_norm": 2.2079601287841797, + "learning_rate": 3.1864337936636886e-06, + "loss": 1.7823, + "step": 1183 + }, + { + "epoch": 1.2975342465753426, + "grad_norm": 2.1104960441589355, + "learning_rate": 3.1844029244516655e-06, + "loss": 1.747, + "step": 1184 + }, + { + "epoch": 1.2986301369863014, + "grad_norm": 2.0351643562316895, + "learning_rate": 3.182372055239643e-06, + "loss": 1.7441, + "step": 1185 + }, + { + "epoch": 1.2997260273972602, + "grad_norm": 2.1289052963256836, + "learning_rate": 3.18034118602762e-06, + "loss": 1.8294, + "step": 1186 + }, + { + "epoch": 1.3008219178082192, + "grad_norm": 2.0787370204925537, + "learning_rate": 3.1783103168155973e-06, + "loss": 1.8035, + "step": 1187 + }, + { + "epoch": 1.301917808219178, + "grad_norm": 1.937454104423523, + "learning_rate": 3.1762794476035746e-06, + "loss": 1.7812, + "step": 1188 + }, + { + "epoch": 1.303013698630137, + "grad_norm": 1.9760960340499878, + "learning_rate": 3.174248578391552e-06, + "loss": 1.791, + "step": 1189 + }, + { + "epoch": 1.3041095890410959, + "grad_norm": 2.175269842147827, + "learning_rate": 3.172217709179529e-06, + "loss": 1.7726, + "step": 1190 + }, + { + "epoch": 1.3041095890410959, + "eval_loss": 1.990390419960022, + "eval_runtime": 37.6381, + "eval_samples_per_second": 26.569, + "eval_steps_per_second": 1.116, + "step": 1190 + }, + { + "epoch": 1.305205479452055, + "grad_norm": 2.113403797149658, + "learning_rate": 3.1701868399675064e-06, + "loss": 1.8015, + "step": 1191 + }, + { + "epoch": 1.3063013698630137, + "grad_norm": 2.0455641746520996, + "learning_rate": 3.1681559707554837e-06, + "loss": 1.7409, + "step": 1192 + }, + { + "epoch": 1.3073972602739725, + "grad_norm": 2.040074348449707, + "learning_rate": 3.166125101543461e-06, + "loss": 1.807, + "step": 1193 + }, + { + "epoch": 1.3084931506849315, + "grad_norm": 1.9834939241409302, + "learning_rate": 3.1640942323314382e-06, + "loss": 1.8028, + "step": 1194 + }, + { + "epoch": 1.3095890410958904, + "grad_norm": 1.9999687671661377, + "learning_rate": 3.1620633631194155e-06, + "loss": 1.7475, + "step": 1195 + }, + { + "epoch": 1.3106849315068494, + "grad_norm": 2.0616705417633057, + "learning_rate": 3.1600324939073928e-06, + "loss": 1.794, + "step": 1196 + }, + { + "epoch": 1.3117808219178082, + "grad_norm": 1.911632776260376, + "learning_rate": 3.15800162469537e-06, + "loss": 1.7867, + "step": 1197 + }, + { + "epoch": 1.3128767123287672, + "grad_norm": 2.0012543201446533, + "learning_rate": 3.155970755483347e-06, + "loss": 1.7852, + "step": 1198 + }, + { + "epoch": 1.313972602739726, + "grad_norm": 1.9527541399002075, + "learning_rate": 3.1539398862713246e-06, + "loss": 1.8033, + "step": 1199 + }, + { + "epoch": 1.3150684931506849, + "grad_norm": 1.9985740184783936, + "learning_rate": 3.1519090170593015e-06, + "loss": 1.7907, + "step": 1200 + }, + { + "epoch": 1.3150684931506849, + "eval_loss": 1.9869638681411743, + "eval_runtime": 37.6494, + "eval_samples_per_second": 26.561, + "eval_steps_per_second": 1.116, + "step": 1200 + }, + { + "epoch": 1.316164383561644, + "grad_norm": 1.9314486980438232, + "learning_rate": 3.149878147847279e-06, + "loss": 1.8566, + "step": 1201 + }, + { + "epoch": 1.3172602739726027, + "grad_norm": 2.019169569015503, + "learning_rate": 3.147847278635256e-06, + "loss": 1.8774, + "step": 1202 + }, + { + "epoch": 1.3183561643835615, + "grad_norm": 2.159221887588501, + "learning_rate": 3.1458164094232337e-06, + "loss": 1.766, + "step": 1203 + }, + { + "epoch": 1.3194520547945205, + "grad_norm": 1.9608495235443115, + "learning_rate": 3.1437855402112106e-06, + "loss": 1.7397, + "step": 1204 + }, + { + "epoch": 1.3205479452054796, + "grad_norm": 2.130441427230835, + "learning_rate": 3.1417546709991883e-06, + "loss": 1.8045, + "step": 1205 + }, + { + "epoch": 1.3216438356164384, + "grad_norm": 2.1584694385528564, + "learning_rate": 3.139723801787165e-06, + "loss": 1.7825, + "step": 1206 + }, + { + "epoch": 1.3227397260273972, + "grad_norm": 2.1429080963134766, + "learning_rate": 3.137692932575143e-06, + "loss": 1.8349, + "step": 1207 + }, + { + "epoch": 1.3238356164383562, + "grad_norm": 2.053940534591675, + "learning_rate": 3.1356620633631197e-06, + "loss": 1.7572, + "step": 1208 + }, + { + "epoch": 1.324931506849315, + "grad_norm": 2.047335624694824, + "learning_rate": 3.1336311941510974e-06, + "loss": 1.7954, + "step": 1209 + }, + { + "epoch": 1.3260273972602739, + "grad_norm": 1.9634249210357666, + "learning_rate": 3.1316003249390742e-06, + "loss": 1.8347, + "step": 1210 + }, + { + "epoch": 1.3260273972602739, + "eval_loss": 1.986480474472046, + "eval_runtime": 37.2913, + "eval_samples_per_second": 26.816, + "eval_steps_per_second": 1.126, + "step": 1210 + }, + { + "epoch": 1.3271232876712329, + "grad_norm": 1.951474666595459, + "learning_rate": 3.129569455727051e-06, + "loss": 1.8349, + "step": 1211 + }, + { + "epoch": 1.3282191780821917, + "grad_norm": 2.0119011402130127, + "learning_rate": 3.1275385865150288e-06, + "loss": 1.8089, + "step": 1212 + }, + { + "epoch": 1.3293150684931507, + "grad_norm": 1.967706561088562, + "learning_rate": 3.1255077173030056e-06, + "loss": 1.8109, + "step": 1213 + }, + { + "epoch": 1.3304109589041095, + "grad_norm": 2.049595594406128, + "learning_rate": 3.1234768480909833e-06, + "loss": 1.7944, + "step": 1214 + }, + { + "epoch": 1.3315068493150686, + "grad_norm": 1.9799634218215942, + "learning_rate": 3.12144597887896e-06, + "loss": 1.8284, + "step": 1215 + }, + { + "epoch": 1.3326027397260274, + "grad_norm": 2.050673246383667, + "learning_rate": 3.119415109666938e-06, + "loss": 1.8522, + "step": 1216 + }, + { + "epoch": 1.3336986301369862, + "grad_norm": 2.0240936279296875, + "learning_rate": 3.1173842404549147e-06, + "loss": 1.8074, + "step": 1217 + }, + { + "epoch": 1.3347945205479452, + "grad_norm": 2.2340989112854004, + "learning_rate": 3.1153533712428924e-06, + "loss": 1.8084, + "step": 1218 + }, + { + "epoch": 1.335890410958904, + "grad_norm": 1.9849423170089722, + "learning_rate": 3.1133225020308693e-06, + "loss": 1.8462, + "step": 1219 + }, + { + "epoch": 1.336986301369863, + "grad_norm": 1.9826864004135132, + "learning_rate": 3.111291632818847e-06, + "loss": 1.849, + "step": 1220 + }, + { + "epoch": 1.336986301369863, + "eval_loss": 1.991707682609558, + "eval_runtime": 37.4253, + "eval_samples_per_second": 26.72, + "eval_steps_per_second": 1.122, + "step": 1220 + }, + { + "epoch": 1.3380821917808219, + "grad_norm": 2.1619536876678467, + "learning_rate": 3.109260763606824e-06, + "loss": 1.8123, + "step": 1221 + }, + { + "epoch": 1.339178082191781, + "grad_norm": 1.9246010780334473, + "learning_rate": 3.1072298943948015e-06, + "loss": 1.823, + "step": 1222 + }, + { + "epoch": 1.3402739726027397, + "grad_norm": 1.9495196342468262, + "learning_rate": 3.1051990251827784e-06, + "loss": 1.7867, + "step": 1223 + }, + { + "epoch": 1.3413698630136985, + "grad_norm": 2.0663001537323, + "learning_rate": 3.1031681559707557e-06, + "loss": 1.842, + "step": 1224 + }, + { + "epoch": 1.3424657534246576, + "grad_norm": 2.03849196434021, + "learning_rate": 3.101137286758733e-06, + "loss": 1.8372, + "step": 1225 + }, + { + "epoch": 1.3435616438356164, + "grad_norm": 2.042194366455078, + "learning_rate": 3.09910641754671e-06, + "loss": 1.8331, + "step": 1226 + }, + { + "epoch": 1.3446575342465754, + "grad_norm": 2.059960126876831, + "learning_rate": 3.0970755483346875e-06, + "loss": 1.821, + "step": 1227 + }, + { + "epoch": 1.3457534246575342, + "grad_norm": 2.00468111038208, + "learning_rate": 3.0950446791226648e-06, + "loss": 1.8766, + "step": 1228 + }, + { + "epoch": 1.3468493150684933, + "grad_norm": 1.9484901428222656, + "learning_rate": 3.093013809910642e-06, + "loss": 1.8149, + "step": 1229 + }, + { + "epoch": 1.347945205479452, + "grad_norm": 1.9615859985351562, + "learning_rate": 3.0909829406986193e-06, + "loss": 1.8588, + "step": 1230 + }, + { + "epoch": 1.347945205479452, + "eval_loss": 1.9857571125030518, + "eval_runtime": 37.535, + "eval_samples_per_second": 26.642, + "eval_steps_per_second": 1.119, + "step": 1230 + }, + { + "epoch": 1.3490410958904109, + "grad_norm": 1.921226978302002, + "learning_rate": 3.0889520714865966e-06, + "loss": 1.8331, + "step": 1231 + }, + { + "epoch": 1.35013698630137, + "grad_norm": 1.9341261386871338, + "learning_rate": 3.086921202274574e-06, + "loss": 1.8011, + "step": 1232 + }, + { + "epoch": 1.3512328767123287, + "grad_norm": 2.078364133834839, + "learning_rate": 3.084890333062551e-06, + "loss": 1.7791, + "step": 1233 + }, + { + "epoch": 1.3523287671232875, + "grad_norm": 2.08891224861145, + "learning_rate": 3.0828594638505284e-06, + "loss": 1.8115, + "step": 1234 + }, + { + "epoch": 1.3534246575342466, + "grad_norm": 1.9678375720977783, + "learning_rate": 3.0808285946385057e-06, + "loss": 1.7734, + "step": 1235 + }, + { + "epoch": 1.3545205479452056, + "grad_norm": 1.9388402700424194, + "learning_rate": 3.078797725426483e-06, + "loss": 1.7146, + "step": 1236 + }, + { + "epoch": 1.3556164383561644, + "grad_norm": 1.8211265802383423, + "learning_rate": 3.07676685621446e-06, + "loss": 1.7996, + "step": 1237 + }, + { + "epoch": 1.3567123287671232, + "grad_norm": 1.9366570711135864, + "learning_rate": 3.0747359870024375e-06, + "loss": 1.7668, + "step": 1238 + }, + { + "epoch": 1.3578082191780823, + "grad_norm": 2.061840534210205, + "learning_rate": 3.0727051177904144e-06, + "loss": 1.7954, + "step": 1239 + }, + { + "epoch": 1.358904109589041, + "grad_norm": 1.989840030670166, + "learning_rate": 3.070674248578392e-06, + "loss": 1.8025, + "step": 1240 + }, + { + "epoch": 1.358904109589041, + "eval_loss": 1.9857583045959473, + "eval_runtime": 37.5608, + "eval_samples_per_second": 26.623, + "eval_steps_per_second": 1.118, + "step": 1240 + }, + { + "epoch": 1.3599999999999999, + "grad_norm": 2.065450429916382, + "learning_rate": 3.068643379366369e-06, + "loss": 1.7966, + "step": 1241 + }, + { + "epoch": 1.361095890410959, + "grad_norm": 2.019740581512451, + "learning_rate": 3.0666125101543466e-06, + "loss": 1.7551, + "step": 1242 + }, + { + "epoch": 1.362191780821918, + "grad_norm": 1.9697777032852173, + "learning_rate": 3.0645816409423235e-06, + "loss": 1.8031, + "step": 1243 + }, + { + "epoch": 1.3632876712328768, + "grad_norm": 2.180959463119507, + "learning_rate": 3.062550771730301e-06, + "loss": 1.778, + "step": 1244 + }, + { + "epoch": 1.3643835616438356, + "grad_norm": 2.149937391281128, + "learning_rate": 3.060519902518278e-06, + "loss": 1.7751, + "step": 1245 + }, + { + "epoch": 1.3654794520547946, + "grad_norm": 2.2257728576660156, + "learning_rate": 3.0584890333062557e-06, + "loss": 1.7237, + "step": 1246 + }, + { + "epoch": 1.3665753424657534, + "grad_norm": 2.0412302017211914, + "learning_rate": 3.0564581640942326e-06, + "loss": 1.8021, + "step": 1247 + }, + { + "epoch": 1.3676712328767122, + "grad_norm": 2.2339093685150146, + "learning_rate": 3.0544272948822094e-06, + "loss": 1.7775, + "step": 1248 + }, + { + "epoch": 1.3687671232876713, + "grad_norm": 2.2473273277282715, + "learning_rate": 3.052396425670187e-06, + "loss": 1.8164, + "step": 1249 + }, + { + "epoch": 1.36986301369863, + "grad_norm": 2.095733165740967, + "learning_rate": 3.050365556458164e-06, + "loss": 1.7727, + "step": 1250 + }, + { + "epoch": 1.36986301369863, + "eval_loss": 1.9857548475265503, + "eval_runtime": 37.5548, + "eval_samples_per_second": 26.628, + "eval_steps_per_second": 1.118, + "step": 1250 + }, + { + "epoch": 1.370958904109589, + "grad_norm": 1.9992356300354004, + "learning_rate": 3.0483346872461417e-06, + "loss": 1.7733, + "step": 1251 + }, + { + "epoch": 1.372054794520548, + "grad_norm": 2.071014404296875, + "learning_rate": 3.0463038180341185e-06, + "loss": 1.8311, + "step": 1252 + }, + { + "epoch": 1.373150684931507, + "grad_norm": 2.258652925491333, + "learning_rate": 3.0442729488220962e-06, + "loss": 1.8352, + "step": 1253 + }, + { + "epoch": 1.3742465753424657, + "grad_norm": 2.1131997108459473, + "learning_rate": 3.042242079610073e-06, + "loss": 1.8184, + "step": 1254 + }, + { + "epoch": 1.3753424657534246, + "grad_norm": 1.9667093753814697, + "learning_rate": 3.0402112103980508e-06, + "loss": 1.7438, + "step": 1255 + }, + { + "epoch": 1.3764383561643836, + "grad_norm": 1.984914779663086, + "learning_rate": 3.0381803411860276e-06, + "loss": 1.7794, + "step": 1256 + }, + { + "epoch": 1.3775342465753424, + "grad_norm": 1.989174246788025, + "learning_rate": 3.0361494719740053e-06, + "loss": 1.7815, + "step": 1257 + }, + { + "epoch": 1.3786301369863014, + "grad_norm": 2.1234655380249023, + "learning_rate": 3.034118602761982e-06, + "loss": 1.7752, + "step": 1258 + }, + { + "epoch": 1.3797260273972602, + "grad_norm": 2.0560953617095947, + "learning_rate": 3.03208773354996e-06, + "loss": 1.7577, + "step": 1259 + }, + { + "epoch": 1.3808219178082193, + "grad_norm": 2.129689931869507, + "learning_rate": 3.0300568643379367e-06, + "loss": 1.824, + "step": 1260 + }, + { + "epoch": 1.3808219178082193, + "eval_loss": 1.9812394380569458, + "eval_runtime": 37.5347, + "eval_samples_per_second": 26.642, + "eval_steps_per_second": 1.119, + "step": 1260 + }, + { + "epoch": 1.381917808219178, + "grad_norm": 2.077955722808838, + "learning_rate": 3.028025995125914e-06, + "loss": 1.8084, + "step": 1261 + }, + { + "epoch": 1.383013698630137, + "grad_norm": 2.0110533237457275, + "learning_rate": 3.0259951259138913e-06, + "loss": 1.739, + "step": 1262 + }, + { + "epoch": 1.384109589041096, + "grad_norm": 2.0531344413757324, + "learning_rate": 3.0239642567018686e-06, + "loss": 1.7879, + "step": 1263 + }, + { + "epoch": 1.3852054794520547, + "grad_norm": 2.1550018787384033, + "learning_rate": 3.021933387489846e-06, + "loss": 1.7465, + "step": 1264 + }, + { + "epoch": 1.3863013698630138, + "grad_norm": 2.087143659591675, + "learning_rate": 3.019902518277823e-06, + "loss": 1.7955, + "step": 1265 + }, + { + "epoch": 1.3873972602739726, + "grad_norm": 1.9999810457229614, + "learning_rate": 3.0178716490658004e-06, + "loss": 1.8334, + "step": 1266 + }, + { + "epoch": 1.3884931506849316, + "grad_norm": 2.239922523498535, + "learning_rate": 3.0158407798537777e-06, + "loss": 1.7348, + "step": 1267 + }, + { + "epoch": 1.3895890410958904, + "grad_norm": 1.9183354377746582, + "learning_rate": 3.013809910641755e-06, + "loss": 1.7459, + "step": 1268 + }, + { + "epoch": 1.3906849315068492, + "grad_norm": 2.0096499919891357, + "learning_rate": 3.0117790414297322e-06, + "loss": 1.7556, + "step": 1269 + }, + { + "epoch": 1.3917808219178083, + "grad_norm": 2.225064516067505, + "learning_rate": 3.0097481722177095e-06, + "loss": 1.7945, + "step": 1270 + }, + { + "epoch": 1.3917808219178083, + "eval_loss": 1.982843279838562, + "eval_runtime": 37.521, + "eval_samples_per_second": 26.652, + "eval_steps_per_second": 1.119, + "step": 1270 + }, + { + "epoch": 1.392876712328767, + "grad_norm": 1.9573938846588135, + "learning_rate": 3.0077173030056868e-06, + "loss": 1.7534, + "step": 1271 + }, + { + "epoch": 1.393972602739726, + "grad_norm": 2.0529563426971436, + "learning_rate": 3.005686433793664e-06, + "loss": 1.8091, + "step": 1272 + }, + { + "epoch": 1.395068493150685, + "grad_norm": 2.3113677501678467, + "learning_rate": 3.0036555645816413e-06, + "loss": 1.7672, + "step": 1273 + }, + { + "epoch": 1.396164383561644, + "grad_norm": 2.013543128967285, + "learning_rate": 3.001624695369618e-06, + "loss": 1.8226, + "step": 1274 + }, + { + "epoch": 1.3972602739726028, + "grad_norm": 2.1328022480010986, + "learning_rate": 2.999593826157596e-06, + "loss": 1.7741, + "step": 1275 + }, + { + "epoch": 1.3983561643835616, + "grad_norm": 2.8097736835479736, + "learning_rate": 2.9975629569455727e-06, + "loss": 1.8155, + "step": 1276 + }, + { + "epoch": 1.3994520547945206, + "grad_norm": 2.2032995223999023, + "learning_rate": 2.9955320877335504e-06, + "loss": 1.7768, + "step": 1277 + }, + { + "epoch": 1.4005479452054794, + "grad_norm": 3.310176134109497, + "learning_rate": 2.9935012185215273e-06, + "loss": 1.8155, + "step": 1278 + }, + { + "epoch": 1.4016438356164382, + "grad_norm": 1.9475737810134888, + "learning_rate": 2.991470349309505e-06, + "loss": 1.774, + "step": 1279 + }, + { + "epoch": 1.4027397260273973, + "grad_norm": 2.5729238986968994, + "learning_rate": 2.989439480097482e-06, + "loss": 1.7907, + "step": 1280 + }, + { + "epoch": 1.4027397260273973, + "eval_loss": 1.9811160564422607, + "eval_runtime": 37.5053, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 1280 + }, + { + "epoch": 1.403835616438356, + "grad_norm": 2.3974802494049072, + "learning_rate": 2.9874086108854595e-06, + "loss": 1.7741, + "step": 1281 + }, + { + "epoch": 1.4049315068493151, + "grad_norm": 2.1172287464141846, + "learning_rate": 2.9853777416734364e-06, + "loss": 1.8272, + "step": 1282 + }, + { + "epoch": 1.406027397260274, + "grad_norm": 3.3079774379730225, + "learning_rate": 2.983346872461414e-06, + "loss": 1.8445, + "step": 1283 + }, + { + "epoch": 1.407123287671233, + "grad_norm": 2.07607102394104, + "learning_rate": 2.981316003249391e-06, + "loss": 1.8064, + "step": 1284 + }, + { + "epoch": 1.4082191780821918, + "grad_norm": 2.2769124507904053, + "learning_rate": 2.9792851340373678e-06, + "loss": 1.7858, + "step": 1285 + }, + { + "epoch": 1.4093150684931506, + "grad_norm": 2.791180372238159, + "learning_rate": 2.9772542648253455e-06, + "loss": 1.8238, + "step": 1286 + }, + { + "epoch": 1.4104109589041096, + "grad_norm": 2.1308441162109375, + "learning_rate": 2.9752233956133223e-06, + "loss": 1.7703, + "step": 1287 + }, + { + "epoch": 1.4115068493150684, + "grad_norm": 2.1865315437316895, + "learning_rate": 2.9731925264013e-06, + "loss": 1.7982, + "step": 1288 + }, + { + "epoch": 1.4126027397260275, + "grad_norm": 3.5701634883880615, + "learning_rate": 2.971161657189277e-06, + "loss": 1.7787, + "step": 1289 + }, + { + "epoch": 1.4136986301369863, + "grad_norm": 1.9576916694641113, + "learning_rate": 2.9691307879772546e-06, + "loss": 1.7711, + "step": 1290 + }, + { + "epoch": 1.4136986301369863, + "eval_loss": 1.9843519926071167, + "eval_runtime": 37.5229, + "eval_samples_per_second": 26.65, + "eval_steps_per_second": 1.119, + "step": 1290 + }, + { + "epoch": 1.4147945205479453, + "grad_norm": 2.6046435832977295, + "learning_rate": 2.9670999187652314e-06, + "loss": 1.7637, + "step": 1291 + }, + { + "epoch": 1.415890410958904, + "grad_norm": 3.658503532409668, + "learning_rate": 2.965069049553209e-06, + "loss": 1.8512, + "step": 1292 + }, + { + "epoch": 1.416986301369863, + "grad_norm": 1.887565016746521, + "learning_rate": 2.963038180341186e-06, + "loss": 1.7421, + "step": 1293 + }, + { + "epoch": 1.418082191780822, + "grad_norm": 2.625554323196411, + "learning_rate": 2.9610073111291637e-06, + "loss": 1.8304, + "step": 1294 + }, + { + "epoch": 1.4191780821917808, + "grad_norm": 2.9848883152008057, + "learning_rate": 2.9589764419171405e-06, + "loss": 1.8451, + "step": 1295 + }, + { + "epoch": 1.4202739726027398, + "grad_norm": 1.9818859100341797, + "learning_rate": 2.9569455727051182e-06, + "loss": 1.7466, + "step": 1296 + }, + { + "epoch": 1.4213698630136986, + "grad_norm": 2.090362310409546, + "learning_rate": 2.954914703493095e-06, + "loss": 1.7852, + "step": 1297 + }, + { + "epoch": 1.4224657534246576, + "grad_norm": 2.8507578372955322, + "learning_rate": 2.9528838342810724e-06, + "loss": 1.7629, + "step": 1298 + }, + { + "epoch": 1.4235616438356165, + "grad_norm": 2.5871756076812744, + "learning_rate": 2.9508529650690496e-06, + "loss": 1.76, + "step": 1299 + }, + { + "epoch": 1.4246575342465753, + "grad_norm": 2.08917236328125, + "learning_rate": 2.948822095857027e-06, + "loss": 1.8103, + "step": 1300 + }, + { + "epoch": 1.4246575342465753, + "eval_loss": 1.9850049018859863, + "eval_runtime": 37.4957, + "eval_samples_per_second": 26.67, + "eval_steps_per_second": 1.12, + "step": 1300 + }, + { + "epoch": 1.4257534246575343, + "grad_norm": 3.8839125633239746, + "learning_rate": 2.946791226645004e-06, + "loss": 1.7672, + "step": 1301 + }, + { + "epoch": 1.426849315068493, + "grad_norm": 1.9710010290145874, + "learning_rate": 2.9447603574329815e-06, + "loss": 1.7348, + "step": 1302 + }, + { + "epoch": 1.427945205479452, + "grad_norm": 2.097695827484131, + "learning_rate": 2.9427294882209587e-06, + "loss": 1.7651, + "step": 1303 + }, + { + "epoch": 1.429041095890411, + "grad_norm": 2.714942216873169, + "learning_rate": 2.940698619008936e-06, + "loss": 1.7819, + "step": 1304 + }, + { + "epoch": 1.43013698630137, + "grad_norm": 2.139592170715332, + "learning_rate": 2.9386677497969133e-06, + "loss": 1.7704, + "step": 1305 + }, + { + "epoch": 1.4312328767123288, + "grad_norm": 2.0190887451171875, + "learning_rate": 2.9366368805848906e-06, + "loss": 1.8119, + "step": 1306 + }, + { + "epoch": 1.4323287671232876, + "grad_norm": 2.0607757568359375, + "learning_rate": 2.934606011372868e-06, + "loss": 1.8133, + "step": 1307 + }, + { + "epoch": 1.4334246575342466, + "grad_norm": 2.0289840698242188, + "learning_rate": 2.932575142160845e-06, + "loss": 1.7949, + "step": 1308 + }, + { + "epoch": 1.4345205479452054, + "grad_norm": 2.0278077125549316, + "learning_rate": 2.9305442729488224e-06, + "loss": 1.798, + "step": 1309 + }, + { + "epoch": 1.4356164383561643, + "grad_norm": 2.10603666305542, + "learning_rate": 2.9285134037367997e-06, + "loss": 1.7895, + "step": 1310 + }, + { + "epoch": 1.4356164383561643, + "eval_loss": 1.9791438579559326, + "eval_runtime": 37.2409, + "eval_samples_per_second": 26.852, + "eval_steps_per_second": 1.128, + "step": 1310 + }, + { + "epoch": 1.4367123287671233, + "grad_norm": 2.009817123413086, + "learning_rate": 2.9264825345247765e-06, + "loss": 1.8163, + "step": 1311 + }, + { + "epoch": 1.4378082191780823, + "grad_norm": 2.185163974761963, + "learning_rate": 2.9244516653127542e-06, + "loss": 1.7905, + "step": 1312 + }, + { + "epoch": 1.4389041095890411, + "grad_norm": 2.074709177017212, + "learning_rate": 2.922420796100731e-06, + "loss": 1.7075, + "step": 1313 + }, + { + "epoch": 1.44, + "grad_norm": 2.172123670578003, + "learning_rate": 2.9203899268887088e-06, + "loss": 1.8299, + "step": 1314 + }, + { + "epoch": 1.441095890410959, + "grad_norm": 2.002028226852417, + "learning_rate": 2.9183590576766856e-06, + "loss": 1.7294, + "step": 1315 + }, + { + "epoch": 1.4421917808219178, + "grad_norm": 1.9138332605361938, + "learning_rate": 2.9163281884646633e-06, + "loss": 1.787, + "step": 1316 + }, + { + "epoch": 1.4432876712328766, + "grad_norm": 2.1227173805236816, + "learning_rate": 2.91429731925264e-06, + "loss": 1.8326, + "step": 1317 + }, + { + "epoch": 1.4443835616438356, + "grad_norm": 2.273021697998047, + "learning_rate": 2.912266450040618e-06, + "loss": 1.7765, + "step": 1318 + }, + { + "epoch": 1.4454794520547944, + "grad_norm": 1.9326497316360474, + "learning_rate": 2.9102355808285947e-06, + "loss": 1.7816, + "step": 1319 + }, + { + "epoch": 1.4465753424657535, + "grad_norm": 2.06606125831604, + "learning_rate": 2.9082047116165724e-06, + "loss": 1.7655, + "step": 1320 + }, + { + "epoch": 1.4465753424657535, + "eval_loss": 1.9804142713546753, + "eval_runtime": 37.4214, + "eval_samples_per_second": 26.723, + "eval_steps_per_second": 1.122, + "step": 1320 + }, + { + "epoch": 1.4476712328767123, + "grad_norm": 2.1864142417907715, + "learning_rate": 2.9061738424045493e-06, + "loss": 1.7437, + "step": 1321 + }, + { + "epoch": 1.4487671232876713, + "grad_norm": 2.2248411178588867, + "learning_rate": 2.904142973192527e-06, + "loss": 1.8464, + "step": 1322 + }, + { + "epoch": 1.4498630136986301, + "grad_norm": 1.9936206340789795, + "learning_rate": 2.902112103980504e-06, + "loss": 1.7913, + "step": 1323 + }, + { + "epoch": 1.450958904109589, + "grad_norm": 2.040400505065918, + "learning_rate": 2.900081234768481e-06, + "loss": 1.7823, + "step": 1324 + }, + { + "epoch": 1.452054794520548, + "grad_norm": 2.291858434677124, + "learning_rate": 2.8980503655564584e-06, + "loss": 1.7748, + "step": 1325 + }, + { + "epoch": 1.4531506849315068, + "grad_norm": 1.9450806379318237, + "learning_rate": 2.8960194963444357e-06, + "loss": 1.7461, + "step": 1326 + }, + { + "epoch": 1.4542465753424658, + "grad_norm": 2.1661314964294434, + "learning_rate": 2.893988627132413e-06, + "loss": 1.7708, + "step": 1327 + }, + { + "epoch": 1.4553424657534246, + "grad_norm": 2.0046839714050293, + "learning_rate": 2.8919577579203902e-06, + "loss": 1.7078, + "step": 1328 + }, + { + "epoch": 1.4564383561643837, + "grad_norm": 2.0632035732269287, + "learning_rate": 2.8899268887083675e-06, + "loss": 1.7695, + "step": 1329 + }, + { + "epoch": 1.4575342465753425, + "grad_norm": 2.090724468231201, + "learning_rate": 2.8878960194963448e-06, + "loss": 1.7795, + "step": 1330 + }, + { + "epoch": 1.4575342465753425, + "eval_loss": 1.9744367599487305, + "eval_runtime": 37.5571, + "eval_samples_per_second": 26.626, + "eval_steps_per_second": 1.118, + "step": 1330 + }, + { + "epoch": 1.4586301369863013, + "grad_norm": 2.0671865940093994, + "learning_rate": 2.885865150284322e-06, + "loss": 1.8078, + "step": 1331 + }, + { + "epoch": 1.4597260273972603, + "grad_norm": 2.011948585510254, + "learning_rate": 2.8838342810722993e-06, + "loss": 1.7611, + "step": 1332 + }, + { + "epoch": 1.4608219178082191, + "grad_norm": 2.0578713417053223, + "learning_rate": 2.8818034118602766e-06, + "loss": 1.7392, + "step": 1333 + }, + { + "epoch": 1.4619178082191782, + "grad_norm": 2.1544125080108643, + "learning_rate": 2.879772542648254e-06, + "loss": 1.8183, + "step": 1334 + }, + { + "epoch": 1.463013698630137, + "grad_norm": 2.026211977005005, + "learning_rate": 2.8777416734362307e-06, + "loss": 1.7275, + "step": 1335 + }, + { + "epoch": 1.464109589041096, + "grad_norm": 2.03863525390625, + "learning_rate": 2.8757108042242084e-06, + "loss": 1.8668, + "step": 1336 + }, + { + "epoch": 1.4652054794520548, + "grad_norm": 1.907163381576538, + "learning_rate": 2.8736799350121853e-06, + "loss": 1.8183, + "step": 1337 + }, + { + "epoch": 1.4663013698630136, + "grad_norm": 2.027798891067505, + "learning_rate": 2.871649065800163e-06, + "loss": 1.8373, + "step": 1338 + }, + { + "epoch": 1.4673972602739727, + "grad_norm": 2.0047388076782227, + "learning_rate": 2.86961819658814e-06, + "loss": 1.7657, + "step": 1339 + }, + { + "epoch": 1.4684931506849315, + "grad_norm": 2.065209150314331, + "learning_rate": 2.8675873273761175e-06, + "loss": 1.7852, + "step": 1340 + }, + { + "epoch": 1.4684931506849315, + "eval_loss": 1.9726630449295044, + "eval_runtime": 37.5967, + "eval_samples_per_second": 26.598, + "eval_steps_per_second": 1.117, + "step": 1340 + }, + { + "epoch": 1.4695890410958903, + "grad_norm": 2.0226666927337646, + "learning_rate": 2.8655564581640944e-06, + "loss": 1.7713, + "step": 1341 + }, + { + "epoch": 1.4706849315068493, + "grad_norm": 1.981685996055603, + "learning_rate": 2.863525588952072e-06, + "loss": 1.8389, + "step": 1342 + }, + { + "epoch": 1.4717808219178083, + "grad_norm": 1.9060076475143433, + "learning_rate": 2.861494719740049e-06, + "loss": 1.7214, + "step": 1343 + }, + { + "epoch": 1.4728767123287672, + "grad_norm": 2.032881498336792, + "learning_rate": 2.8594638505280266e-06, + "loss": 1.7371, + "step": 1344 + }, + { + "epoch": 1.473972602739726, + "grad_norm": 1.9562633037567139, + "learning_rate": 2.8574329813160035e-06, + "loss": 1.7757, + "step": 1345 + }, + { + "epoch": 1.475068493150685, + "grad_norm": 1.9932552576065063, + "learning_rate": 2.855402112103981e-06, + "loss": 1.7607, + "step": 1346 + }, + { + "epoch": 1.4761643835616438, + "grad_norm": 2.0549392700195312, + "learning_rate": 2.853371242891958e-06, + "loss": 1.7832, + "step": 1347 + }, + { + "epoch": 1.4772602739726026, + "grad_norm": 2.2332963943481445, + "learning_rate": 2.851340373679935e-06, + "loss": 1.794, + "step": 1348 + }, + { + "epoch": 1.4783561643835617, + "grad_norm": 2.140798807144165, + "learning_rate": 2.8493095044679126e-06, + "loss": 1.8529, + "step": 1349 + }, + { + "epoch": 1.4794520547945205, + "grad_norm": 2.0761284828186035, + "learning_rate": 2.8472786352558894e-06, + "loss": 1.8016, + "step": 1350 + }, + { + "epoch": 1.4794520547945205, + "eval_loss": 1.9772450923919678, + "eval_runtime": 37.6307, + "eval_samples_per_second": 26.574, + "eval_steps_per_second": 1.116, + "step": 1350 + }, + { + "epoch": 1.4805479452054795, + "grad_norm": 2.108281373977661, + "learning_rate": 2.845247766043867e-06, + "loss": 1.7419, + "step": 1351 + }, + { + "epoch": 1.4816438356164383, + "grad_norm": 2.0750250816345215, + "learning_rate": 2.843216896831844e-06, + "loss": 1.735, + "step": 1352 + }, + { + "epoch": 1.4827397260273973, + "grad_norm": 2.021198272705078, + "learning_rate": 2.8411860276198217e-06, + "loss": 1.7526, + "step": 1353 + }, + { + "epoch": 1.4838356164383562, + "grad_norm": 2.0334317684173584, + "learning_rate": 2.8391551584077985e-06, + "loss": 1.8128, + "step": 1354 + }, + { + "epoch": 1.484931506849315, + "grad_norm": 2.0824408531188965, + "learning_rate": 2.8371242891957762e-06, + "loss": 1.8399, + "step": 1355 + }, + { + "epoch": 1.486027397260274, + "grad_norm": 1.9490233659744263, + "learning_rate": 2.835093419983753e-06, + "loss": 1.767, + "step": 1356 + }, + { + "epoch": 1.4871232876712328, + "grad_norm": 2.1065993309020996, + "learning_rate": 2.833062550771731e-06, + "loss": 1.7679, + "step": 1357 + }, + { + "epoch": 1.4882191780821918, + "grad_norm": 2.0148558616638184, + "learning_rate": 2.8310316815597076e-06, + "loss": 1.7386, + "step": 1358 + }, + { + "epoch": 1.4893150684931507, + "grad_norm": 2.188098669052124, + "learning_rate": 2.8290008123476853e-06, + "loss": 1.7778, + "step": 1359 + }, + { + "epoch": 1.4904109589041097, + "grad_norm": 2.2202253341674805, + "learning_rate": 2.826969943135662e-06, + "loss": 1.7739, + "step": 1360 + }, + { + "epoch": 1.4904109589041097, + "eval_loss": 1.9797139167785645, + "eval_runtime": 37.6095, + "eval_samples_per_second": 26.589, + "eval_steps_per_second": 1.117, + "step": 1360 + }, + { + "epoch": 1.4915068493150685, + "grad_norm": 2.0920464992523193, + "learning_rate": 2.8249390739236395e-06, + "loss": 1.7811, + "step": 1361 + }, + { + "epoch": 1.4926027397260273, + "grad_norm": 1.9884718656539917, + "learning_rate": 2.8229082047116167e-06, + "loss": 1.7444, + "step": 1362 + }, + { + "epoch": 1.4936986301369863, + "grad_norm": 2.1209335327148438, + "learning_rate": 2.820877335499594e-06, + "loss": 1.7638, + "step": 1363 + }, + { + "epoch": 1.4947945205479451, + "grad_norm": 1.9715428352355957, + "learning_rate": 2.8188464662875713e-06, + "loss": 1.7163, + "step": 1364 + }, + { + "epoch": 1.4958904109589042, + "grad_norm": 1.9295417070388794, + "learning_rate": 2.8168155970755486e-06, + "loss": 1.7399, + "step": 1365 + }, + { + "epoch": 1.496986301369863, + "grad_norm": 1.9574483633041382, + "learning_rate": 2.814784727863526e-06, + "loss": 1.759, + "step": 1366 + }, + { + "epoch": 1.498082191780822, + "grad_norm": 2.0707592964172363, + "learning_rate": 2.812753858651503e-06, + "loss": 1.791, + "step": 1367 + }, + { + "epoch": 1.4991780821917808, + "grad_norm": 2.24857497215271, + "learning_rate": 2.8107229894394804e-06, + "loss": 1.7692, + "step": 1368 + }, + { + "epoch": 1.5002739726027396, + "grad_norm": 2.0012736320495605, + "learning_rate": 2.8086921202274577e-06, + "loss": 1.7251, + "step": 1369 + }, + { + "epoch": 1.5013698630136987, + "grad_norm": 2.120032787322998, + "learning_rate": 2.806661251015435e-06, + "loss": 1.8065, + "step": 1370 + }, + { + "epoch": 1.5013698630136987, + "eval_loss": 1.9802494049072266, + "eval_runtime": 37.6191, + "eval_samples_per_second": 26.582, + "eval_steps_per_second": 1.116, + "step": 1370 + }, + { + "epoch": 1.5024657534246575, + "grad_norm": 2.059269905090332, + "learning_rate": 2.8046303818034122e-06, + "loss": 1.849, + "step": 1371 + }, + { + "epoch": 1.5035616438356163, + "grad_norm": 2.005443811416626, + "learning_rate": 2.802599512591389e-06, + "loss": 1.8966, + "step": 1372 + }, + { + "epoch": 1.5046575342465753, + "grad_norm": 1.9510835409164429, + "learning_rate": 2.8005686433793668e-06, + "loss": 1.7712, + "step": 1373 + }, + { + "epoch": 1.5057534246575344, + "grad_norm": 2.0802578926086426, + "learning_rate": 2.7985377741673436e-06, + "loss": 1.724, + "step": 1374 + }, + { + "epoch": 1.5068493150684932, + "grad_norm": 2.250236988067627, + "learning_rate": 2.7965069049553213e-06, + "loss": 1.7818, + "step": 1375 + }, + { + "epoch": 1.507945205479452, + "grad_norm": 2.288304567337036, + "learning_rate": 2.794476035743298e-06, + "loss": 1.8265, + "step": 1376 + }, + { + "epoch": 1.509041095890411, + "grad_norm": 2.1732380390167236, + "learning_rate": 2.792445166531276e-06, + "loss": 1.8009, + "step": 1377 + }, + { + "epoch": 1.5101369863013698, + "grad_norm": 2.1012215614318848, + "learning_rate": 2.7904142973192527e-06, + "loss": 1.8032, + "step": 1378 + }, + { + "epoch": 1.5112328767123286, + "grad_norm": 2.122478723526001, + "learning_rate": 2.7883834281072304e-06, + "loss": 1.741, + "step": 1379 + }, + { + "epoch": 1.5123287671232877, + "grad_norm": 2.114091396331787, + "learning_rate": 2.7863525588952073e-06, + "loss": 1.7347, + "step": 1380 + }, + { + "epoch": 1.5123287671232877, + "eval_loss": 1.9719252586364746, + "eval_runtime": 37.6053, + "eval_samples_per_second": 26.592, + "eval_steps_per_second": 1.117, + "step": 1380 + }, + { + "epoch": 1.5134246575342467, + "grad_norm": 2.059864044189453, + "learning_rate": 2.784321689683185e-06, + "loss": 1.7383, + "step": 1381 + }, + { + "epoch": 1.5145205479452055, + "grad_norm": 2.051453113555908, + "learning_rate": 2.782290820471162e-06, + "loss": 1.7363, + "step": 1382 + }, + { + "epoch": 1.5156164383561643, + "grad_norm": 2.1645872592926025, + "learning_rate": 2.7802599512591395e-06, + "loss": 1.8204, + "step": 1383 + }, + { + "epoch": 1.5167123287671234, + "grad_norm": 2.1003007888793945, + "learning_rate": 2.7782290820471164e-06, + "loss": 1.7997, + "step": 1384 + }, + { + "epoch": 1.5178082191780822, + "grad_norm": 1.9776486158370972, + "learning_rate": 2.7761982128350932e-06, + "loss": 1.8082, + "step": 1385 + }, + { + "epoch": 1.518904109589041, + "grad_norm": 1.9704468250274658, + "learning_rate": 2.774167343623071e-06, + "loss": 1.8202, + "step": 1386 + }, + { + "epoch": 1.52, + "grad_norm": 2.102381944656372, + "learning_rate": 2.772136474411048e-06, + "loss": 1.7491, + "step": 1387 + }, + { + "epoch": 1.521095890410959, + "grad_norm": 2.0659008026123047, + "learning_rate": 2.7701056051990255e-06, + "loss": 1.6852, + "step": 1388 + }, + { + "epoch": 1.5221917808219179, + "grad_norm": 2.0750720500946045, + "learning_rate": 2.7680747359870023e-06, + "loss": 1.7838, + "step": 1389 + }, + { + "epoch": 1.5232876712328767, + "grad_norm": 2.0162465572357178, + "learning_rate": 2.76604386677498e-06, + "loss": 1.7667, + "step": 1390 + }, + { + "epoch": 1.5232876712328767, + "eval_loss": 1.9698952436447144, + "eval_runtime": 37.6248, + "eval_samples_per_second": 26.578, + "eval_steps_per_second": 1.116, + "step": 1390 + }, + { + "epoch": 1.5243835616438357, + "grad_norm": 2.224431037902832, + "learning_rate": 2.764012997562957e-06, + "loss": 1.8349, + "step": 1391 + }, + { + "epoch": 1.5254794520547945, + "grad_norm": 2.0104691982269287, + "learning_rate": 2.7619821283509346e-06, + "loss": 1.7213, + "step": 1392 + }, + { + "epoch": 1.5265753424657533, + "grad_norm": 1.9719268083572388, + "learning_rate": 2.7599512591389114e-06, + "loss": 1.807, + "step": 1393 + }, + { + "epoch": 1.5276712328767124, + "grad_norm": 1.996199607849121, + "learning_rate": 2.757920389926889e-06, + "loss": 1.7653, + "step": 1394 + }, + { + "epoch": 1.5287671232876714, + "grad_norm": 2.1448185443878174, + "learning_rate": 2.755889520714866e-06, + "loss": 1.7484, + "step": 1395 + }, + { + "epoch": 1.52986301369863, + "grad_norm": 2.108182907104492, + "learning_rate": 2.7538586515028437e-06, + "loss": 1.7219, + "step": 1396 + }, + { + "epoch": 1.530958904109589, + "grad_norm": 2.0446760654449463, + "learning_rate": 2.7518277822908205e-06, + "loss": 1.7776, + "step": 1397 + }, + { + "epoch": 1.532054794520548, + "grad_norm": 2.0954434871673584, + "learning_rate": 2.749796913078798e-06, + "loss": 1.8224, + "step": 1398 + }, + { + "epoch": 1.5331506849315069, + "grad_norm": 1.979532241821289, + "learning_rate": 2.747766043866775e-06, + "loss": 1.7704, + "step": 1399 + }, + { + "epoch": 1.5342465753424657, + "grad_norm": 1.8814504146575928, + "learning_rate": 2.7457351746547524e-06, + "loss": 1.8426, + "step": 1400 + }, + { + "epoch": 1.5342465753424657, + "eval_loss": 1.9686505794525146, + "eval_runtime": 37.6118, + "eval_samples_per_second": 26.587, + "eval_steps_per_second": 1.117, + "step": 1400 + }, + { + "epoch": 1.5353424657534247, + "grad_norm": 1.9090824127197266, + "learning_rate": 2.7437043054427297e-06, + "loss": 1.8815, + "step": 1401 + }, + { + "epoch": 1.5364383561643835, + "grad_norm": 2.0893030166625977, + "learning_rate": 2.741673436230707e-06, + "loss": 1.7769, + "step": 1402 + }, + { + "epoch": 1.5375342465753423, + "grad_norm": 2.0485997200012207, + "learning_rate": 2.739642567018684e-06, + "loss": 1.7477, + "step": 1403 + }, + { + "epoch": 1.5386301369863014, + "grad_norm": 2.1117589473724365, + "learning_rate": 2.7376116978066615e-06, + "loss": 1.7929, + "step": 1404 + }, + { + "epoch": 1.5397260273972604, + "grad_norm": 2.239506721496582, + "learning_rate": 2.7355808285946388e-06, + "loss": 1.787, + "step": 1405 + }, + { + "epoch": 1.5408219178082192, + "grad_norm": 2.030825614929199, + "learning_rate": 2.733549959382616e-06, + "loss": 1.7477, + "step": 1406 + }, + { + "epoch": 1.541917808219178, + "grad_norm": 2.074589967727661, + "learning_rate": 2.7315190901705933e-06, + "loss": 1.8296, + "step": 1407 + }, + { + "epoch": 1.543013698630137, + "grad_norm": 1.9294577836990356, + "learning_rate": 2.7294882209585706e-06, + "loss": 1.7911, + "step": 1408 + }, + { + "epoch": 1.5441095890410959, + "grad_norm": 1.9394869804382324, + "learning_rate": 2.727457351746548e-06, + "loss": 1.752, + "step": 1409 + }, + { + "epoch": 1.5452054794520547, + "grad_norm": 2.060983657836914, + "learning_rate": 2.725426482534525e-06, + "loss": 1.7456, + "step": 1410 + }, + { + "epoch": 1.5452054794520547, + "eval_loss": 1.9721039533615112, + "eval_runtime": 37.3127, + "eval_samples_per_second": 26.801, + "eval_steps_per_second": 1.126, + "step": 1410 + }, + { + "epoch": 1.5463013698630137, + "grad_norm": 1.973060131072998, + "learning_rate": 2.723395613322502e-06, + "loss": 1.787, + "step": 1411 + }, + { + "epoch": 1.5473972602739727, + "grad_norm": 2.001887559890747, + "learning_rate": 2.7213647441104797e-06, + "loss": 1.7681, + "step": 1412 + }, + { + "epoch": 1.5484931506849315, + "grad_norm": 2.193004608154297, + "learning_rate": 2.7193338748984565e-06, + "loss": 1.7744, + "step": 1413 + }, + { + "epoch": 1.5495890410958904, + "grad_norm": 2.15061092376709, + "learning_rate": 2.7173030056864342e-06, + "loss": 1.9328, + "step": 1414 + }, + { + "epoch": 1.5506849315068494, + "grad_norm": 2.1413729190826416, + "learning_rate": 2.715272136474411e-06, + "loss": 1.7168, + "step": 1415 + }, + { + "epoch": 1.5517808219178082, + "grad_norm": 2.1227574348449707, + "learning_rate": 2.713241267262389e-06, + "loss": 1.7845, + "step": 1416 + }, + { + "epoch": 1.552876712328767, + "grad_norm": 2.091916799545288, + "learning_rate": 2.7112103980503656e-06, + "loss": 1.7459, + "step": 1417 + }, + { + "epoch": 1.553972602739726, + "grad_norm": 2.022514820098877, + "learning_rate": 2.7091795288383433e-06, + "loss": 1.8067, + "step": 1418 + }, + { + "epoch": 1.555068493150685, + "grad_norm": 2.039914131164551, + "learning_rate": 2.70714865962632e-06, + "loss": 1.7561, + "step": 1419 + }, + { + "epoch": 1.5561643835616439, + "grad_norm": 2.186922073364258, + "learning_rate": 2.705117790414298e-06, + "loss": 1.7604, + "step": 1420 + }, + { + "epoch": 1.5561643835616439, + "eval_loss": 1.9736120700836182, + "eval_runtime": 37.4922, + "eval_samples_per_second": 26.672, + "eval_steps_per_second": 1.12, + "step": 1420 + }, + { + "epoch": 1.5572602739726027, + "grad_norm": 2.0078210830688477, + "learning_rate": 2.7030869212022747e-06, + "loss": 1.8195, + "step": 1421 + }, + { + "epoch": 1.5583561643835617, + "grad_norm": 2.03193998336792, + "learning_rate": 2.7010560519902516e-06, + "loss": 1.7429, + "step": 1422 + }, + { + "epoch": 1.5594520547945205, + "grad_norm": 2.0053257942199707, + "learning_rate": 2.6990251827782293e-06, + "loss": 1.7643, + "step": 1423 + }, + { + "epoch": 1.5605479452054793, + "grad_norm": 1.9069421291351318, + "learning_rate": 2.696994313566206e-06, + "loss": 1.7462, + "step": 1424 + }, + { + "epoch": 1.5616438356164384, + "grad_norm": 1.9090805053710938, + "learning_rate": 2.694963444354184e-06, + "loss": 1.8425, + "step": 1425 + }, + { + "epoch": 1.5627397260273974, + "grad_norm": 1.9518672227859497, + "learning_rate": 2.6929325751421607e-06, + "loss": 1.7524, + "step": 1426 + }, + { + "epoch": 1.563835616438356, + "grad_norm": 1.9681649208068848, + "learning_rate": 2.6909017059301384e-06, + "loss": 1.8132, + "step": 1427 + }, + { + "epoch": 1.564931506849315, + "grad_norm": 2.0240578651428223, + "learning_rate": 2.6888708367181153e-06, + "loss": 1.7179, + "step": 1428 + }, + { + "epoch": 1.566027397260274, + "grad_norm": 2.120497703552246, + "learning_rate": 2.686839967506093e-06, + "loss": 1.7982, + "step": 1429 + }, + { + "epoch": 1.5671232876712329, + "grad_norm": 2.030435800552368, + "learning_rate": 2.68480909829407e-06, + "loss": 1.6864, + "step": 1430 + }, + { + "epoch": 1.5671232876712329, + "eval_loss": 1.9688317775726318, + "eval_runtime": 37.6029, + "eval_samples_per_second": 26.594, + "eval_steps_per_second": 1.117, + "step": 1430 + }, + { + "epoch": 1.5682191780821917, + "grad_norm": 2.083026170730591, + "learning_rate": 2.6827782290820475e-06, + "loss": 1.6925, + "step": 1431 + }, + { + "epoch": 1.5693150684931507, + "grad_norm": 2.0663950443267822, + "learning_rate": 2.6807473598700244e-06, + "loss": 1.8128, + "step": 1432 + }, + { + "epoch": 1.5704109589041098, + "grad_norm": 2.019975185394287, + "learning_rate": 2.678716490658002e-06, + "loss": 1.7813, + "step": 1433 + }, + { + "epoch": 1.5715068493150683, + "grad_norm": 2.0030932426452637, + "learning_rate": 2.676685621445979e-06, + "loss": 1.7864, + "step": 1434 + }, + { + "epoch": 1.5726027397260274, + "grad_norm": 2.0622000694274902, + "learning_rate": 2.674654752233956e-06, + "loss": 1.7946, + "step": 1435 + }, + { + "epoch": 1.5736986301369864, + "grad_norm": 2.003072738647461, + "learning_rate": 2.6726238830219335e-06, + "loss": 1.785, + "step": 1436 + }, + { + "epoch": 1.5747945205479452, + "grad_norm": 2.0511863231658936, + "learning_rate": 2.6705930138099107e-06, + "loss": 1.8211, + "step": 1437 + }, + { + "epoch": 1.575890410958904, + "grad_norm": 1.9955956935882568, + "learning_rate": 2.668562144597888e-06, + "loss": 1.6871, + "step": 1438 + }, + { + "epoch": 1.576986301369863, + "grad_norm": 2.0979461669921875, + "learning_rate": 2.6665312753858653e-06, + "loss": 1.7806, + "step": 1439 + }, + { + "epoch": 1.5780821917808219, + "grad_norm": 1.9976561069488525, + "learning_rate": 2.6645004061738426e-06, + "loss": 1.7798, + "step": 1440 + }, + { + "epoch": 1.5780821917808219, + "eval_loss": 1.9716663360595703, + "eval_runtime": 37.6084, + "eval_samples_per_second": 26.59, + "eval_steps_per_second": 1.117, + "step": 1440 + }, + { + "epoch": 1.5791780821917807, + "grad_norm": 2.1817586421966553, + "learning_rate": 2.66246953696182e-06, + "loss": 1.7292, + "step": 1441 + }, + { + "epoch": 1.5802739726027397, + "grad_norm": 2.0021443367004395, + "learning_rate": 2.660438667749797e-06, + "loss": 1.7075, + "step": 1442 + }, + { + "epoch": 1.5813698630136988, + "grad_norm": 1.8750648498535156, + "learning_rate": 2.6584077985377744e-06, + "loss": 1.7591, + "step": 1443 + }, + { + "epoch": 1.5824657534246576, + "grad_norm": 1.9670345783233643, + "learning_rate": 2.6563769293257517e-06, + "loss": 1.77, + "step": 1444 + }, + { + "epoch": 1.5835616438356164, + "grad_norm": 1.9381967782974243, + "learning_rate": 2.654346060113729e-06, + "loss": 1.7303, + "step": 1445 + }, + { + "epoch": 1.5846575342465754, + "grad_norm": 2.0975873470306396, + "learning_rate": 2.6523151909017062e-06, + "loss": 1.7475, + "step": 1446 + }, + { + "epoch": 1.5857534246575342, + "grad_norm": 1.9539753198623657, + "learning_rate": 2.6502843216896835e-06, + "loss": 1.7323, + "step": 1447 + }, + { + "epoch": 1.586849315068493, + "grad_norm": 1.962587833404541, + "learning_rate": 2.6482534524776603e-06, + "loss": 1.8429, + "step": 1448 + }, + { + "epoch": 1.587945205479452, + "grad_norm": 1.9619494676589966, + "learning_rate": 2.646222583265638e-06, + "loss": 1.8039, + "step": 1449 + }, + { + "epoch": 1.589041095890411, + "grad_norm": 2.091921091079712, + "learning_rate": 2.644191714053615e-06, + "loss": 1.781, + "step": 1450 + }, + { + "epoch": 1.589041095890411, + "eval_loss": 1.9683992862701416, + "eval_runtime": 37.624, + "eval_samples_per_second": 26.579, + "eval_steps_per_second": 1.116, + "step": 1450 + }, + { + "epoch": 1.59013698630137, + "grad_norm": 2.1267876625061035, + "learning_rate": 2.6421608448415926e-06, + "loss": 1.7357, + "step": 1451 + }, + { + "epoch": 1.5912328767123287, + "grad_norm": 2.2417800426483154, + "learning_rate": 2.6401299756295694e-06, + "loss": 1.7245, + "step": 1452 + }, + { + "epoch": 1.5923287671232877, + "grad_norm": 2.1259522438049316, + "learning_rate": 2.638099106417547e-06, + "loss": 1.7858, + "step": 1453 + }, + { + "epoch": 1.5934246575342466, + "grad_norm": 2.128701686859131, + "learning_rate": 2.636068237205524e-06, + "loss": 1.7375, + "step": 1454 + }, + { + "epoch": 1.5945205479452054, + "grad_norm": 2.028517246246338, + "learning_rate": 2.6340373679935017e-06, + "loss": 1.8399, + "step": 1455 + }, + { + "epoch": 1.5956164383561644, + "grad_norm": 2.1483123302459717, + "learning_rate": 2.6320064987814785e-06, + "loss": 1.6903, + "step": 1456 + }, + { + "epoch": 1.5967123287671234, + "grad_norm": 2.1509265899658203, + "learning_rate": 2.6299756295694562e-06, + "loss": 1.7911, + "step": 1457 + }, + { + "epoch": 1.5978082191780822, + "grad_norm": 2.1323459148406982, + "learning_rate": 2.627944760357433e-06, + "loss": 1.8291, + "step": 1458 + }, + { + "epoch": 1.598904109589041, + "grad_norm": 2.0712826251983643, + "learning_rate": 2.6259138911454104e-06, + "loss": 1.8398, + "step": 1459 + }, + { + "epoch": 1.6, + "grad_norm": 2.0771944522857666, + "learning_rate": 2.6238830219333877e-06, + "loss": 1.8176, + "step": 1460 + }, + { + "epoch": 1.6, + "eval_loss": 1.962112545967102, + "eval_runtime": 37.5906, + "eval_samples_per_second": 26.602, + "eval_steps_per_second": 1.117, + "step": 1460 + }, + { + "epoch": 1.601095890410959, + "grad_norm": 2.064944267272949, + "learning_rate": 2.621852152721365e-06, + "loss": 1.6995, + "step": 1461 + }, + { + "epoch": 1.6021917808219177, + "grad_norm": 2.0380799770355225, + "learning_rate": 2.619821283509342e-06, + "loss": 1.8017, + "step": 1462 + }, + { + "epoch": 1.6032876712328767, + "grad_norm": 2.0653076171875, + "learning_rate": 2.6177904142973195e-06, + "loss": 1.7324, + "step": 1463 + }, + { + "epoch": 1.6043835616438358, + "grad_norm": 2.042045831680298, + "learning_rate": 2.6157595450852968e-06, + "loss": 1.7875, + "step": 1464 + }, + { + "epoch": 1.6054794520547944, + "grad_norm": 1.9798356294631958, + "learning_rate": 2.613728675873274e-06, + "loss": 1.8024, + "step": 1465 + }, + { + "epoch": 1.6065753424657534, + "grad_norm": 1.9917240142822266, + "learning_rate": 2.6116978066612513e-06, + "loss": 1.8298, + "step": 1466 + }, + { + "epoch": 1.6076712328767124, + "grad_norm": 1.9990789890289307, + "learning_rate": 2.6096669374492286e-06, + "loss": 1.7757, + "step": 1467 + }, + { + "epoch": 1.6087671232876712, + "grad_norm": 1.9390219449996948, + "learning_rate": 2.607636068237206e-06, + "loss": 1.732, + "step": 1468 + }, + { + "epoch": 1.60986301369863, + "grad_norm": 2.0136828422546387, + "learning_rate": 2.605605199025183e-06, + "loss": 1.7942, + "step": 1469 + }, + { + "epoch": 1.610958904109589, + "grad_norm": 1.9484944343566895, + "learning_rate": 2.6035743298131604e-06, + "loss": 1.7656, + "step": 1470 + }, + { + "epoch": 1.610958904109589, + "eval_loss": 1.9646728038787842, + "eval_runtime": 37.5868, + "eval_samples_per_second": 26.605, + "eval_steps_per_second": 1.117, + "step": 1470 + }, + { + "epoch": 1.612054794520548, + "grad_norm": 2.0478456020355225, + "learning_rate": 2.6015434606011377e-06, + "loss": 1.8037, + "step": 1471 + }, + { + "epoch": 1.6131506849315067, + "grad_norm": 2.0135908126831055, + "learning_rate": 2.5995125913891145e-06, + "loss": 1.7815, + "step": 1472 + }, + { + "epoch": 1.6142465753424657, + "grad_norm": 1.994864583015442, + "learning_rate": 2.5974817221770922e-06, + "loss": 1.7643, + "step": 1473 + }, + { + "epoch": 1.6153424657534248, + "grad_norm": 2.0730109214782715, + "learning_rate": 2.595450852965069e-06, + "loss": 1.8038, + "step": 1474 + }, + { + "epoch": 1.6164383561643836, + "grad_norm": 2.012881278991699, + "learning_rate": 2.593419983753047e-06, + "loss": 1.8071, + "step": 1475 + }, + { + "epoch": 1.6175342465753424, + "grad_norm": 2.20212459564209, + "learning_rate": 2.5913891145410236e-06, + "loss": 1.7495, + "step": 1476 + }, + { + "epoch": 1.6186301369863014, + "grad_norm": 1.971611499786377, + "learning_rate": 2.5893582453290013e-06, + "loss": 1.8128, + "step": 1477 + }, + { + "epoch": 1.6197260273972602, + "grad_norm": 2.0107650756835938, + "learning_rate": 2.587327376116978e-06, + "loss": 1.7531, + "step": 1478 + }, + { + "epoch": 1.620821917808219, + "grad_norm": 1.9367616176605225, + "learning_rate": 2.585296506904956e-06, + "loss": 1.7576, + "step": 1479 + }, + { + "epoch": 1.621917808219178, + "grad_norm": 1.8660348653793335, + "learning_rate": 2.5832656376929327e-06, + "loss": 1.7428, + "step": 1480 + }, + { + "epoch": 1.621917808219178, + "eval_loss": 1.9651310443878174, + "eval_runtime": 37.5912, + "eval_samples_per_second": 26.602, + "eval_steps_per_second": 1.117, + "step": 1480 + }, + { + "epoch": 1.6230136986301371, + "grad_norm": 1.980948567390442, + "learning_rate": 2.5812347684809104e-06, + "loss": 1.6846, + "step": 1481 + }, + { + "epoch": 1.624109589041096, + "grad_norm": 1.9366756677627563, + "learning_rate": 2.5792038992688873e-06, + "loss": 1.8548, + "step": 1482 + }, + { + "epoch": 1.6252054794520547, + "grad_norm": 1.9535239934921265, + "learning_rate": 2.577173030056865e-06, + "loss": 1.6751, + "step": 1483 + }, + { + "epoch": 1.6263013698630138, + "grad_norm": 1.9503936767578125, + "learning_rate": 2.575142160844842e-06, + "loss": 1.7936, + "step": 1484 + }, + { + "epoch": 1.6273972602739726, + "grad_norm": 1.9823237657546997, + "learning_rate": 2.5731112916328187e-06, + "loss": 1.7992, + "step": 1485 + }, + { + "epoch": 1.6284931506849314, + "grad_norm": 1.936607003211975, + "learning_rate": 2.5710804224207964e-06, + "loss": 1.7297, + "step": 1486 + }, + { + "epoch": 1.6295890410958904, + "grad_norm": 1.9715105295181274, + "learning_rate": 2.5690495532087733e-06, + "loss": 1.6904, + "step": 1487 + }, + { + "epoch": 1.6306849315068495, + "grad_norm": 1.954744815826416, + "learning_rate": 2.567018683996751e-06, + "loss": 1.8027, + "step": 1488 + }, + { + "epoch": 1.6317808219178083, + "grad_norm": 1.910043716430664, + "learning_rate": 2.564987814784728e-06, + "loss": 1.7263, + "step": 1489 + }, + { + "epoch": 1.632876712328767, + "grad_norm": 1.9333269596099854, + "learning_rate": 2.5629569455727055e-06, + "loss": 1.7592, + "step": 1490 + }, + { + "epoch": 1.632876712328767, + "eval_loss": 1.9617011547088623, + "eval_runtime": 37.5866, + "eval_samples_per_second": 26.605, + "eval_steps_per_second": 1.117, + "step": 1490 + }, + { + "epoch": 1.633972602739726, + "grad_norm": 1.937027931213379, + "learning_rate": 2.5609260763606824e-06, + "loss": 1.7499, + "step": 1491 + }, + { + "epoch": 1.635068493150685, + "grad_norm": 1.9691696166992188, + "learning_rate": 2.55889520714866e-06, + "loss": 1.8642, + "step": 1492 + }, + { + "epoch": 1.6361643835616437, + "grad_norm": 1.9791373014450073, + "learning_rate": 2.556864337936637e-06, + "loss": 1.7512, + "step": 1493 + }, + { + "epoch": 1.6372602739726028, + "grad_norm": 2.035804510116577, + "learning_rate": 2.5548334687246146e-06, + "loss": 1.7625, + "step": 1494 + }, + { + "epoch": 1.6383561643835618, + "grad_norm": 1.9504729509353638, + "learning_rate": 2.5528025995125915e-06, + "loss": 1.8128, + "step": 1495 + }, + { + "epoch": 1.6394520547945204, + "grad_norm": 2.119971513748169, + "learning_rate": 2.550771730300569e-06, + "loss": 1.7902, + "step": 1496 + }, + { + "epoch": 1.6405479452054794, + "grad_norm": 1.9479756355285645, + "learning_rate": 2.548740861088546e-06, + "loss": 1.8019, + "step": 1497 + }, + { + "epoch": 1.6416438356164385, + "grad_norm": 2.0506067276000977, + "learning_rate": 2.5467099918765233e-06, + "loss": 1.6934, + "step": 1498 + }, + { + "epoch": 1.6427397260273973, + "grad_norm": 2.146449089050293, + "learning_rate": 2.5446791226645006e-06, + "loss": 1.6957, + "step": 1499 + }, + { + "epoch": 1.643835616438356, + "grad_norm": 1.9801750183105469, + "learning_rate": 2.542648253452478e-06, + "loss": 1.7984, + "step": 1500 + }, + { + "epoch": 1.643835616438356, + "eval_loss": 1.9647246599197388, + "eval_runtime": 37.5676, + "eval_samples_per_second": 26.619, + "eval_steps_per_second": 1.118, + "step": 1500 + }, + { + "epoch": 1.644931506849315, + "grad_norm": 2.122166872024536, + "learning_rate": 2.540617384240455e-06, + "loss": 1.7736, + "step": 1501 + }, + { + "epoch": 1.6460273972602741, + "grad_norm": 1.9945837259292603, + "learning_rate": 2.5385865150284324e-06, + "loss": 1.7362, + "step": 1502 + }, + { + "epoch": 1.6471232876712327, + "grad_norm": 1.977905511856079, + "learning_rate": 2.5365556458164097e-06, + "loss": 1.7556, + "step": 1503 + }, + { + "epoch": 1.6482191780821918, + "grad_norm": 2.03883957862854, + "learning_rate": 2.534524776604387e-06, + "loss": 1.7662, + "step": 1504 + }, + { + "epoch": 1.6493150684931508, + "grad_norm": 1.985999584197998, + "learning_rate": 2.5324939073923642e-06, + "loss": 1.6532, + "step": 1505 + }, + { + "epoch": 1.6504109589041096, + "grad_norm": 1.9366891384124756, + "learning_rate": 2.5304630381803415e-06, + "loss": 1.7577, + "step": 1506 + }, + { + "epoch": 1.6515068493150684, + "grad_norm": 2.0977749824523926, + "learning_rate": 2.5284321689683188e-06, + "loss": 1.7595, + "step": 1507 + }, + { + "epoch": 1.6526027397260274, + "grad_norm": 1.9759522676467896, + "learning_rate": 2.526401299756296e-06, + "loss": 1.742, + "step": 1508 + }, + { + "epoch": 1.6536986301369863, + "grad_norm": 1.9824578762054443, + "learning_rate": 2.524370430544273e-06, + "loss": 1.7691, + "step": 1509 + }, + { + "epoch": 1.654794520547945, + "grad_norm": 2.033693790435791, + "learning_rate": 2.5223395613322506e-06, + "loss": 1.7244, + "step": 1510 + }, + { + "epoch": 1.654794520547945, + "eval_loss": 1.9656455516815186, + "eval_runtime": 37.226, + "eval_samples_per_second": 26.863, + "eval_steps_per_second": 1.128, + "step": 1510 + }, + { + "epoch": 1.655890410958904, + "grad_norm": 2.015699625015259, + "learning_rate": 2.5203086921202274e-06, + "loss": 1.7977, + "step": 1511 + }, + { + "epoch": 1.6569863013698631, + "grad_norm": 2.0714852809906006, + "learning_rate": 2.518277822908205e-06, + "loss": 1.7488, + "step": 1512 + }, + { + "epoch": 1.658082191780822, + "grad_norm": 1.9523553848266602, + "learning_rate": 2.516246953696182e-06, + "loss": 1.7877, + "step": 1513 + }, + { + "epoch": 1.6591780821917808, + "grad_norm": 2.1122264862060547, + "learning_rate": 2.5142160844841597e-06, + "loss": 1.7888, + "step": 1514 + }, + { + "epoch": 1.6602739726027398, + "grad_norm": 1.9710755348205566, + "learning_rate": 2.5121852152721365e-06, + "loss": 1.7781, + "step": 1515 + }, + { + "epoch": 1.6613698630136986, + "grad_norm": 1.9889023303985596, + "learning_rate": 2.5101543460601142e-06, + "loss": 1.7769, + "step": 1516 + }, + { + "epoch": 1.6624657534246574, + "grad_norm": 2.0876429080963135, + "learning_rate": 2.508123476848091e-06, + "loss": 1.7106, + "step": 1517 + }, + { + "epoch": 1.6635616438356164, + "grad_norm": 2.0601608753204346, + "learning_rate": 2.506092607636069e-06, + "loss": 1.7985, + "step": 1518 + }, + { + "epoch": 1.6646575342465755, + "grad_norm": 1.9425318241119385, + "learning_rate": 2.5040617384240457e-06, + "loss": 1.7326, + "step": 1519 + }, + { + "epoch": 1.6657534246575343, + "grad_norm": 2.1803112030029297, + "learning_rate": 2.5020308692120234e-06, + "loss": 1.7241, + "step": 1520 + }, + { + "epoch": 1.6657534246575343, + "eval_loss": 1.9666804075241089, + "eval_runtime": 37.3882, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 1520 + }, + { + "epoch": 1.666849315068493, + "grad_norm": 2.4585888385772705, + "learning_rate": 2.5e-06, + "loss": 1.8266, + "step": 1521 + }, + { + "epoch": 1.6679452054794521, + "grad_norm": 2.0859873294830322, + "learning_rate": 2.4979691307879775e-06, + "loss": 1.7332, + "step": 1522 + }, + { + "epoch": 1.669041095890411, + "grad_norm": 1.9705928564071655, + "learning_rate": 2.4959382615759548e-06, + "loss": 1.7264, + "step": 1523 + }, + { + "epoch": 1.6701369863013698, + "grad_norm": 2.631394147872925, + "learning_rate": 2.493907392363932e-06, + "loss": 1.8336, + "step": 1524 + }, + { + "epoch": 1.6712328767123288, + "grad_norm": 2.0018208026885986, + "learning_rate": 2.4918765231519093e-06, + "loss": 1.7892, + "step": 1525 + }, + { + "epoch": 1.6723287671232878, + "grad_norm": 2.0302157402038574, + "learning_rate": 2.4898456539398866e-06, + "loss": 1.7424, + "step": 1526 + }, + { + "epoch": 1.6734246575342466, + "grad_norm": 2.2957849502563477, + "learning_rate": 2.487814784727864e-06, + "loss": 1.8317, + "step": 1527 + }, + { + "epoch": 1.6745205479452054, + "grad_norm": 2.0042552947998047, + "learning_rate": 2.4857839155158407e-06, + "loss": 1.7548, + "step": 1528 + }, + { + "epoch": 1.6756164383561645, + "grad_norm": 2.0509517192840576, + "learning_rate": 2.483753046303818e-06, + "loss": 1.754, + "step": 1529 + }, + { + "epoch": 1.6767123287671233, + "grad_norm": 2.132749080657959, + "learning_rate": 2.4817221770917953e-06, + "loss": 1.7559, + "step": 1530 + }, + { + "epoch": 1.6767123287671233, + "eval_loss": 1.961317777633667, + "eval_runtime": 37.4576, + "eval_samples_per_second": 26.697, + "eval_steps_per_second": 1.121, + "step": 1530 + }, + { + "epoch": 1.677808219178082, + "grad_norm": 2.0966835021972656, + "learning_rate": 2.4796913078797725e-06, + "loss": 1.7919, + "step": 1531 + }, + { + "epoch": 1.6789041095890411, + "grad_norm": 1.9800058603286743, + "learning_rate": 2.47766043866775e-06, + "loss": 1.7834, + "step": 1532 + }, + { + "epoch": 1.6800000000000002, + "grad_norm": 1.9882386922836304, + "learning_rate": 2.475629569455727e-06, + "loss": 1.7594, + "step": 1533 + }, + { + "epoch": 1.6810958904109587, + "grad_norm": 1.9218485355377197, + "learning_rate": 2.4735987002437044e-06, + "loss": 1.7347, + "step": 1534 + }, + { + "epoch": 1.6821917808219178, + "grad_norm": 1.996363878250122, + "learning_rate": 2.4715678310316816e-06, + "loss": 1.7529, + "step": 1535 + }, + { + "epoch": 1.6832876712328768, + "grad_norm": 1.961236596107483, + "learning_rate": 2.469536961819659e-06, + "loss": 1.7561, + "step": 1536 + }, + { + "epoch": 1.6843835616438356, + "grad_norm": 2.024998664855957, + "learning_rate": 2.467506092607636e-06, + "loss": 1.7408, + "step": 1537 + }, + { + "epoch": 1.6854794520547944, + "grad_norm": 2.069365978240967, + "learning_rate": 2.4654752233956135e-06, + "loss": 1.7252, + "step": 1538 + }, + { + "epoch": 1.6865753424657535, + "grad_norm": 2.128957509994507, + "learning_rate": 2.4634443541835907e-06, + "loss": 1.8226, + "step": 1539 + }, + { + "epoch": 1.6876712328767123, + "grad_norm": 2.031522274017334, + "learning_rate": 2.461413484971568e-06, + "loss": 1.7584, + "step": 1540 + }, + { + "epoch": 1.6876712328767123, + "eval_loss": 1.959444522857666, + "eval_runtime": 37.5055, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 1540 + }, + { + "epoch": 1.688767123287671, + "grad_norm": 1.9841612577438354, + "learning_rate": 2.4593826157595453e-06, + "loss": 1.8442, + "step": 1541 + }, + { + "epoch": 1.6898630136986301, + "grad_norm": 2.141650915145874, + "learning_rate": 2.4573517465475226e-06, + "loss": 1.7004, + "step": 1542 + }, + { + "epoch": 1.6909589041095892, + "grad_norm": 1.9188008308410645, + "learning_rate": 2.4553208773355e-06, + "loss": 1.7469, + "step": 1543 + }, + { + "epoch": 1.692054794520548, + "grad_norm": 1.8655012845993042, + "learning_rate": 2.453290008123477e-06, + "loss": 1.7632, + "step": 1544 + }, + { + "epoch": 1.6931506849315068, + "grad_norm": 2.1063475608825684, + "learning_rate": 2.4512591389114544e-06, + "loss": 1.761, + "step": 1545 + }, + { + "epoch": 1.6942465753424658, + "grad_norm": 2.0043179988861084, + "learning_rate": 2.4492282696994317e-06, + "loss": 1.7716, + "step": 1546 + }, + { + "epoch": 1.6953424657534246, + "grad_norm": 1.9580390453338623, + "learning_rate": 2.447197400487409e-06, + "loss": 1.772, + "step": 1547 + }, + { + "epoch": 1.6964383561643834, + "grad_norm": 2.3084945678710938, + "learning_rate": 2.4451665312753862e-06, + "loss": 1.7995, + "step": 1548 + }, + { + "epoch": 1.6975342465753425, + "grad_norm": 2.001495599746704, + "learning_rate": 2.4431356620633635e-06, + "loss": 1.7981, + "step": 1549 + }, + { + "epoch": 1.6986301369863015, + "grad_norm": 1.936398983001709, + "learning_rate": 2.4411047928513408e-06, + "loss": 1.7586, + "step": 1550 + }, + { + "epoch": 1.6986301369863015, + "eval_loss": 1.9544179439544678, + "eval_runtime": 37.5022, + "eval_samples_per_second": 26.665, + "eval_steps_per_second": 1.12, + "step": 1550 + }, + { + "epoch": 1.6997260273972603, + "grad_norm": 2.043818712234497, + "learning_rate": 2.439073923639318e-06, + "loss": 1.7741, + "step": 1551 + }, + { + "epoch": 1.7008219178082191, + "grad_norm": 2.249664068222046, + "learning_rate": 2.4370430544272953e-06, + "loss": 1.7257, + "step": 1552 + }, + { + "epoch": 1.7019178082191782, + "grad_norm": 2.014836072921753, + "learning_rate": 2.435012185215272e-06, + "loss": 1.7108, + "step": 1553 + }, + { + "epoch": 1.703013698630137, + "grad_norm": 2.298334836959839, + "learning_rate": 2.4329813160032495e-06, + "loss": 1.7339, + "step": 1554 + }, + { + "epoch": 1.7041095890410958, + "grad_norm": 2.2216415405273438, + "learning_rate": 2.4309504467912267e-06, + "loss": 1.7671, + "step": 1555 + }, + { + "epoch": 1.7052054794520548, + "grad_norm": 2.092634677886963, + "learning_rate": 2.428919577579204e-06, + "loss": 1.7422, + "step": 1556 + }, + { + "epoch": 1.7063013698630138, + "grad_norm": 2.7248661518096924, + "learning_rate": 2.4268887083671813e-06, + "loss": 1.8349, + "step": 1557 + }, + { + "epoch": 1.7073972602739727, + "grad_norm": 2.0364444255828857, + "learning_rate": 2.4248578391551586e-06, + "loss": 1.717, + "step": 1558 + }, + { + "epoch": 1.7084931506849315, + "grad_norm": 1.9016317129135132, + "learning_rate": 2.422826969943136e-06, + "loss": 1.7417, + "step": 1559 + }, + { + "epoch": 1.7095890410958905, + "grad_norm": 2.5477652549743652, + "learning_rate": 2.420796100731113e-06, + "loss": 1.7043, + "step": 1560 + }, + { + "epoch": 1.7095890410958905, + "eval_loss": 1.9544798135757446, + "eval_runtime": 37.5221, + "eval_samples_per_second": 26.651, + "eval_steps_per_second": 1.119, + "step": 1560 + }, + { + "epoch": 1.7106849315068493, + "grad_norm": 2.017261028289795, + "learning_rate": 2.4187652315190904e-06, + "loss": 1.7602, + "step": 1561 + }, + { + "epoch": 1.7117808219178081, + "grad_norm": 2.2273268699645996, + "learning_rate": 2.4167343623070677e-06, + "loss": 1.7726, + "step": 1562 + }, + { + "epoch": 1.7128767123287671, + "grad_norm": 2.1747512817382812, + "learning_rate": 2.414703493095045e-06, + "loss": 1.7603, + "step": 1563 + }, + { + "epoch": 1.7139726027397262, + "grad_norm": 2.0712475776672363, + "learning_rate": 2.4126726238830222e-06, + "loss": 1.787, + "step": 1564 + }, + { + "epoch": 1.7150684931506848, + "grad_norm": 2.02577805519104, + "learning_rate": 2.410641754670999e-06, + "loss": 1.7366, + "step": 1565 + }, + { + "epoch": 1.7161643835616438, + "grad_norm": 2.1076927185058594, + "learning_rate": 2.4086108854589763e-06, + "loss": 1.7998, + "step": 1566 + }, + { + "epoch": 1.7172602739726028, + "grad_norm": 2.062802314758301, + "learning_rate": 2.4065800162469536e-06, + "loss": 1.7321, + "step": 1567 + }, + { + "epoch": 1.7183561643835616, + "grad_norm": 2.1243865489959717, + "learning_rate": 2.404549147034931e-06, + "loss": 1.7654, + "step": 1568 + }, + { + "epoch": 1.7194520547945205, + "grad_norm": 2.2422707080841064, + "learning_rate": 2.402518277822908e-06, + "loss": 1.8068, + "step": 1569 + }, + { + "epoch": 1.7205479452054795, + "grad_norm": 1.927902340888977, + "learning_rate": 2.4004874086108854e-06, + "loss": 1.7288, + "step": 1570 + }, + { + "epoch": 1.7205479452054795, + "eval_loss": 1.9516545534133911, + "eval_runtime": 37.531, + "eval_samples_per_second": 26.645, + "eval_steps_per_second": 1.119, + "step": 1570 + }, + { + "epoch": 1.7216438356164385, + "grad_norm": 2.273752450942993, + "learning_rate": 2.3984565393988627e-06, + "loss": 1.7971, + "step": 1571 + }, + { + "epoch": 1.7227397260273971, + "grad_norm": 2.0363776683807373, + "learning_rate": 2.39642567018684e-06, + "loss": 1.7286, + "step": 1572 + }, + { + "epoch": 1.7238356164383561, + "grad_norm": 1.9637500047683716, + "learning_rate": 2.3943948009748173e-06, + "loss": 1.7435, + "step": 1573 + }, + { + "epoch": 1.7249315068493152, + "grad_norm": 1.9725103378295898, + "learning_rate": 2.3923639317627945e-06, + "loss": 1.7712, + "step": 1574 + }, + { + "epoch": 1.726027397260274, + "grad_norm": 2.019195795059204, + "learning_rate": 2.390333062550772e-06, + "loss": 1.7193, + "step": 1575 + }, + { + "epoch": 1.7271232876712328, + "grad_norm": 1.9915928840637207, + "learning_rate": 2.388302193338749e-06, + "loss": 1.6404, + "step": 1576 + }, + { + "epoch": 1.7282191780821918, + "grad_norm": 2.2263803482055664, + "learning_rate": 2.3862713241267264e-06, + "loss": 1.8404, + "step": 1577 + }, + { + "epoch": 1.7293150684931506, + "grad_norm": 1.9913668632507324, + "learning_rate": 2.3842404549147037e-06, + "loss": 1.7559, + "step": 1578 + }, + { + "epoch": 1.7304109589041095, + "grad_norm": 1.9765959978103638, + "learning_rate": 2.382209585702681e-06, + "loss": 1.7134, + "step": 1579 + }, + { + "epoch": 1.7315068493150685, + "grad_norm": 1.941770076751709, + "learning_rate": 2.380178716490658e-06, + "loss": 1.8003, + "step": 1580 + }, + { + "epoch": 1.7315068493150685, + "eval_loss": 1.9468334913253784, + "eval_runtime": 37.5401, + "eval_samples_per_second": 26.638, + "eval_steps_per_second": 1.119, + "step": 1580 + }, + { + "epoch": 1.7326027397260275, + "grad_norm": 2.0806727409362793, + "learning_rate": 2.3781478472786355e-06, + "loss": 1.7534, + "step": 1581 + }, + { + "epoch": 1.7336986301369863, + "grad_norm": 2.050828695297241, + "learning_rate": 2.3761169780666128e-06, + "loss": 1.7915, + "step": 1582 + }, + { + "epoch": 1.7347945205479451, + "grad_norm": 1.9756336212158203, + "learning_rate": 2.37408610885459e-06, + "loss": 1.8137, + "step": 1583 + }, + { + "epoch": 1.7358904109589042, + "grad_norm": 2.020108938217163, + "learning_rate": 2.3720552396425673e-06, + "loss": 1.7455, + "step": 1584 + }, + { + "epoch": 1.736986301369863, + "grad_norm": 2.0776844024658203, + "learning_rate": 2.3700243704305446e-06, + "loss": 1.8213, + "step": 1585 + }, + { + "epoch": 1.7380821917808218, + "grad_norm": 2.089085340499878, + "learning_rate": 2.367993501218522e-06, + "loss": 1.7593, + "step": 1586 + }, + { + "epoch": 1.7391780821917808, + "grad_norm": 2.106271982192993, + "learning_rate": 2.365962632006499e-06, + "loss": 1.7649, + "step": 1587 + }, + { + "epoch": 1.7402739726027399, + "grad_norm": 2.0556890964508057, + "learning_rate": 2.3639317627944764e-06, + "loss": 1.7973, + "step": 1588 + }, + { + "epoch": 1.7413698630136987, + "grad_norm": 1.9738290309906006, + "learning_rate": 2.3619008935824537e-06, + "loss": 1.7834, + "step": 1589 + }, + { + "epoch": 1.7424657534246575, + "grad_norm": 1.9077492952346802, + "learning_rate": 2.3598700243704305e-06, + "loss": 1.7862, + "step": 1590 + }, + { + "epoch": 1.7424657534246575, + "eval_loss": 1.9491550922393799, + "eval_runtime": 37.5401, + "eval_samples_per_second": 26.638, + "eval_steps_per_second": 1.119, + "step": 1590 + }, + { + "epoch": 1.7435616438356165, + "grad_norm": 1.895346999168396, + "learning_rate": 2.357839155158408e-06, + "loss": 1.7813, + "step": 1591 + }, + { + "epoch": 1.7446575342465753, + "grad_norm": 1.9841216802597046, + "learning_rate": 2.355808285946385e-06, + "loss": 1.7871, + "step": 1592 + }, + { + "epoch": 1.7457534246575341, + "grad_norm": 2.0224013328552246, + "learning_rate": 2.3537774167343624e-06, + "loss": 1.8079, + "step": 1593 + }, + { + "epoch": 1.7468493150684932, + "grad_norm": 1.92066490650177, + "learning_rate": 2.3517465475223396e-06, + "loss": 1.7708, + "step": 1594 + }, + { + "epoch": 1.7479452054794522, + "grad_norm": 2.0495505332946777, + "learning_rate": 2.349715678310317e-06, + "loss": 1.77, + "step": 1595 + }, + { + "epoch": 1.749041095890411, + "grad_norm": 2.0403316020965576, + "learning_rate": 2.347684809098294e-06, + "loss": 1.7339, + "step": 1596 + }, + { + "epoch": 1.7501369863013698, + "grad_norm": 1.9952152967453003, + "learning_rate": 2.3456539398862715e-06, + "loss": 1.7859, + "step": 1597 + }, + { + "epoch": 1.7512328767123289, + "grad_norm": 1.980825662612915, + "learning_rate": 2.3436230706742487e-06, + "loss": 1.7315, + "step": 1598 + }, + { + "epoch": 1.7523287671232877, + "grad_norm": 2.0775835514068604, + "learning_rate": 2.341592201462226e-06, + "loss": 1.7615, + "step": 1599 + }, + { + "epoch": 1.7534246575342465, + "grad_norm": 2.0406181812286377, + "learning_rate": 2.3395613322502033e-06, + "loss": 1.7475, + "step": 1600 + }, + { + "epoch": 1.7534246575342465, + "eval_loss": 1.9491811990737915, + "eval_runtime": 37.558, + "eval_samples_per_second": 26.626, + "eval_steps_per_second": 1.118, + "step": 1600 + }, + { + "epoch": 1.7545205479452055, + "grad_norm": 1.983383059501648, + "learning_rate": 2.3375304630381806e-06, + "loss": 1.7135, + "step": 1601 + }, + { + "epoch": 1.7556164383561645, + "grad_norm": 2.065302610397339, + "learning_rate": 2.335499593826158e-06, + "loss": 1.7148, + "step": 1602 + }, + { + "epoch": 1.7567123287671231, + "grad_norm": 2.0635147094726562, + "learning_rate": 2.333468724614135e-06, + "loss": 1.7718, + "step": 1603 + }, + { + "epoch": 1.7578082191780822, + "grad_norm": 2.1237823963165283, + "learning_rate": 2.3314378554021124e-06, + "loss": 1.7893, + "step": 1604 + }, + { + "epoch": 1.7589041095890412, + "grad_norm": 2.0154919624328613, + "learning_rate": 2.3294069861900897e-06, + "loss": 1.7171, + "step": 1605 + }, + { + "epoch": 1.76, + "grad_norm": 1.9559892416000366, + "learning_rate": 2.327376116978067e-06, + "loss": 1.7144, + "step": 1606 + }, + { + "epoch": 1.7610958904109588, + "grad_norm": 2.076488733291626, + "learning_rate": 2.3253452477660442e-06, + "loss": 1.8004, + "step": 1607 + }, + { + "epoch": 1.7621917808219179, + "grad_norm": 2.0016939640045166, + "learning_rate": 2.3233143785540215e-06, + "loss": 1.7352, + "step": 1608 + }, + { + "epoch": 1.7632876712328767, + "grad_norm": 2.081190586090088, + "learning_rate": 2.3212835093419988e-06, + "loss": 1.7833, + "step": 1609 + }, + { + "epoch": 1.7643835616438355, + "grad_norm": 2.0911076068878174, + "learning_rate": 2.319252640129976e-06, + "loss": 1.7238, + "step": 1610 + }, + { + "epoch": 1.7643835616438355, + "eval_loss": 1.9489142894744873, + "eval_runtime": 37.2922, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 1610 + }, + { + "epoch": 1.7654794520547945, + "grad_norm": 2.1103882789611816, + "learning_rate": 2.3172217709179533e-06, + "loss": 1.8207, + "step": 1611 + }, + { + "epoch": 1.7665753424657535, + "grad_norm": 1.964347243309021, + "learning_rate": 2.3151909017059306e-06, + "loss": 1.7946, + "step": 1612 + }, + { + "epoch": 1.7676712328767124, + "grad_norm": 2.2181637287139893, + "learning_rate": 2.313160032493908e-06, + "loss": 1.7078, + "step": 1613 + }, + { + "epoch": 1.7687671232876712, + "grad_norm": 2.0376179218292236, + "learning_rate": 2.311129163281885e-06, + "loss": 1.7869, + "step": 1614 + }, + { + "epoch": 1.7698630136986302, + "grad_norm": 1.9327906370162964, + "learning_rate": 2.309098294069862e-06, + "loss": 1.7155, + "step": 1615 + }, + { + "epoch": 1.770958904109589, + "grad_norm": 2.0721137523651123, + "learning_rate": 2.3070674248578393e-06, + "loss": 1.7097, + "step": 1616 + }, + { + "epoch": 1.7720547945205478, + "grad_norm": 2.1026227474212646, + "learning_rate": 2.3050365556458166e-06, + "loss": 1.7692, + "step": 1617 + }, + { + "epoch": 1.7731506849315068, + "grad_norm": 2.130305528640747, + "learning_rate": 2.303005686433794e-06, + "loss": 1.7808, + "step": 1618 + }, + { + "epoch": 1.7742465753424659, + "grad_norm": 2.0941500663757324, + "learning_rate": 2.300974817221771e-06, + "loss": 1.7389, + "step": 1619 + }, + { + "epoch": 1.7753424657534247, + "grad_norm": 2.0324289798736572, + "learning_rate": 2.2989439480097484e-06, + "loss": 1.7024, + "step": 1620 + }, + { + "epoch": 1.7753424657534247, + "eval_loss": 1.9513062238693237, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 1620 + }, + { + "epoch": 1.7764383561643835, + "grad_norm": 2.1128101348876953, + "learning_rate": 2.2969130787977257e-06, + "loss": 1.7577, + "step": 1621 + }, + { + "epoch": 1.7775342465753425, + "grad_norm": 1.93445885181427, + "learning_rate": 2.294882209585703e-06, + "loss": 1.7829, + "step": 1622 + }, + { + "epoch": 1.7786301369863013, + "grad_norm": 2.091400146484375, + "learning_rate": 2.2928513403736802e-06, + "loss": 1.7095, + "step": 1623 + }, + { + "epoch": 1.7797260273972602, + "grad_norm": 2.117192029953003, + "learning_rate": 2.2908204711616575e-06, + "loss": 1.753, + "step": 1624 + }, + { + "epoch": 1.7808219178082192, + "grad_norm": 2.0938236713409424, + "learning_rate": 2.2887896019496348e-06, + "loss": 1.8042, + "step": 1625 + }, + { + "epoch": 1.7819178082191782, + "grad_norm": 2.0304651260375977, + "learning_rate": 2.286758732737612e-06, + "loss": 1.7359, + "step": 1626 + }, + { + "epoch": 1.783013698630137, + "grad_norm": 2.0851471424102783, + "learning_rate": 2.284727863525589e-06, + "loss": 1.7473, + "step": 1627 + }, + { + "epoch": 1.7841095890410958, + "grad_norm": 2.043586254119873, + "learning_rate": 2.282696994313566e-06, + "loss": 1.799, + "step": 1628 + }, + { + "epoch": 1.7852054794520549, + "grad_norm": 2.0063164234161377, + "learning_rate": 2.2806661251015434e-06, + "loss": 1.7716, + "step": 1629 + }, + { + "epoch": 1.7863013698630137, + "grad_norm": 2.0192148685455322, + "learning_rate": 2.2786352558895207e-06, + "loss": 1.7392, + "step": 1630 + }, + { + "epoch": 1.7863013698630137, + "eval_loss": 1.9452368021011353, + "eval_runtime": 37.578, + "eval_samples_per_second": 26.611, + "eval_steps_per_second": 1.118, + "step": 1630 + }, + { + "epoch": 1.7873972602739725, + "grad_norm": 1.9149479866027832, + "learning_rate": 2.276604386677498e-06, + "loss": 1.699, + "step": 1631 + }, + { + "epoch": 1.7884931506849315, + "grad_norm": 2.102924346923828, + "learning_rate": 2.2745735174654753e-06, + "loss": 1.7245, + "step": 1632 + }, + { + "epoch": 1.7895890410958906, + "grad_norm": 2.1121819019317627, + "learning_rate": 2.2725426482534526e-06, + "loss": 1.8049, + "step": 1633 + }, + { + "epoch": 1.7906849315068492, + "grad_norm": 2.0112698078155518, + "learning_rate": 2.27051177904143e-06, + "loss": 1.7468, + "step": 1634 + }, + { + "epoch": 1.7917808219178082, + "grad_norm": 1.9730697870254517, + "learning_rate": 2.268480909829407e-06, + "loss": 1.7492, + "step": 1635 + }, + { + "epoch": 1.7928767123287672, + "grad_norm": 1.9572551250457764, + "learning_rate": 2.2664500406173844e-06, + "loss": 1.7533, + "step": 1636 + }, + { + "epoch": 1.793972602739726, + "grad_norm": 2.0028905868530273, + "learning_rate": 2.2644191714053617e-06, + "loss": 1.7939, + "step": 1637 + }, + { + "epoch": 1.7950684931506848, + "grad_norm": 2.0761966705322266, + "learning_rate": 2.262388302193339e-06, + "loss": 1.767, + "step": 1638 + }, + { + "epoch": 1.7961643835616439, + "grad_norm": 1.9057316780090332, + "learning_rate": 2.260357432981316e-06, + "loss": 1.7469, + "step": 1639 + }, + { + "epoch": 1.7972602739726027, + "grad_norm": 2.0405161380767822, + "learning_rate": 2.2583265637692935e-06, + "loss": 1.7658, + "step": 1640 + }, + { + "epoch": 1.7972602739726027, + "eval_loss": 1.9442158937454224, + "eval_runtime": 37.6343, + "eval_samples_per_second": 26.571, + "eval_steps_per_second": 1.116, + "step": 1640 + }, + { + "epoch": 1.7983561643835615, + "grad_norm": 1.9433516263961792, + "learning_rate": 2.2562956945572708e-06, + "loss": 1.7424, + "step": 1641 + }, + { + "epoch": 1.7994520547945205, + "grad_norm": 1.9645577669143677, + "learning_rate": 2.254264825345248e-06, + "loss": 1.7521, + "step": 1642 + }, + { + "epoch": 1.8005479452054796, + "grad_norm": 1.9828311204910278, + "learning_rate": 2.2522339561332253e-06, + "loss": 1.7529, + "step": 1643 + }, + { + "epoch": 1.8016438356164384, + "grad_norm": 1.9738712310791016, + "learning_rate": 2.2502030869212026e-06, + "loss": 1.7616, + "step": 1644 + }, + { + "epoch": 1.8027397260273972, + "grad_norm": 1.8852397203445435, + "learning_rate": 2.24817221770918e-06, + "loss": 1.8189, + "step": 1645 + }, + { + "epoch": 1.8038356164383562, + "grad_norm": 1.974819302558899, + "learning_rate": 2.246141348497157e-06, + "loss": 1.7895, + "step": 1646 + }, + { + "epoch": 1.804931506849315, + "grad_norm": 2.036257266998291, + "learning_rate": 2.2441104792851344e-06, + "loss": 1.738, + "step": 1647 + }, + { + "epoch": 1.8060273972602738, + "grad_norm": 1.9910297393798828, + "learning_rate": 2.2420796100731117e-06, + "loss": 1.7585, + "step": 1648 + }, + { + "epoch": 1.8071232876712329, + "grad_norm": 2.1427969932556152, + "learning_rate": 2.240048740861089e-06, + "loss": 1.7866, + "step": 1649 + }, + { + "epoch": 1.808219178082192, + "grad_norm": 2.018359661102295, + "learning_rate": 2.2380178716490662e-06, + "loss": 1.639, + "step": 1650 + }, + { + "epoch": 1.808219178082192, + "eval_loss": 1.9436448812484741, + "eval_runtime": 37.6307, + "eval_samples_per_second": 26.574, + "eval_steps_per_second": 1.116, + "step": 1650 + }, + { + "epoch": 1.8093150684931507, + "grad_norm": 2.058678388595581, + "learning_rate": 2.2359870024370435e-06, + "loss": 1.7789, + "step": 1651 + }, + { + "epoch": 1.8104109589041095, + "grad_norm": 2.056478977203369, + "learning_rate": 2.2339561332250204e-06, + "loss": 1.7878, + "step": 1652 + }, + { + "epoch": 1.8115068493150686, + "grad_norm": 2.010570526123047, + "learning_rate": 2.2319252640129976e-06, + "loss": 1.7651, + "step": 1653 + }, + { + "epoch": 1.8126027397260274, + "grad_norm": 1.9039926528930664, + "learning_rate": 2.229894394800975e-06, + "loss": 1.7009, + "step": 1654 + }, + { + "epoch": 1.8136986301369862, + "grad_norm": 2.100445032119751, + "learning_rate": 2.227863525588952e-06, + "loss": 1.6879, + "step": 1655 + }, + { + "epoch": 1.8147945205479452, + "grad_norm": 2.057743787765503, + "learning_rate": 2.2258326563769295e-06, + "loss": 1.6992, + "step": 1656 + }, + { + "epoch": 1.8158904109589042, + "grad_norm": 1.9845513105392456, + "learning_rate": 2.2238017871649067e-06, + "loss": 1.6893, + "step": 1657 + }, + { + "epoch": 1.816986301369863, + "grad_norm": 2.033980369567871, + "learning_rate": 2.221770917952884e-06, + "loss": 1.7497, + "step": 1658 + }, + { + "epoch": 1.8180821917808219, + "grad_norm": 2.1005442142486572, + "learning_rate": 2.2197400487408613e-06, + "loss": 1.7936, + "step": 1659 + }, + { + "epoch": 1.819178082191781, + "grad_norm": 2.017122268676758, + "learning_rate": 2.2177091795288386e-06, + "loss": 1.6999, + "step": 1660 + }, + { + "epoch": 1.819178082191781, + "eval_loss": 1.9451484680175781, + "eval_runtime": 37.6149, + "eval_samples_per_second": 26.585, + "eval_steps_per_second": 1.117, + "step": 1660 + }, + { + "epoch": 1.8202739726027397, + "grad_norm": 2.0698463916778564, + "learning_rate": 2.215678310316816e-06, + "loss": 1.7521, + "step": 1661 + }, + { + "epoch": 1.8213698630136985, + "grad_norm": 1.897350788116455, + "learning_rate": 2.213647441104793e-06, + "loss": 1.7611, + "step": 1662 + }, + { + "epoch": 1.8224657534246576, + "grad_norm": 2.1905267238616943, + "learning_rate": 2.2116165718927704e-06, + "loss": 1.7269, + "step": 1663 + }, + { + "epoch": 1.8235616438356166, + "grad_norm": 2.0472371578216553, + "learning_rate": 2.2095857026807473e-06, + "loss": 1.7581, + "step": 1664 + }, + { + "epoch": 1.8246575342465754, + "grad_norm": 1.9712332487106323, + "learning_rate": 2.2075548334687245e-06, + "loss": 1.7023, + "step": 1665 + }, + { + "epoch": 1.8257534246575342, + "grad_norm": 2.0489535331726074, + "learning_rate": 2.205523964256702e-06, + "loss": 1.7291, + "step": 1666 + }, + { + "epoch": 1.8268493150684932, + "grad_norm": 2.5053207874298096, + "learning_rate": 2.203493095044679e-06, + "loss": 1.7523, + "step": 1667 + }, + { + "epoch": 1.827945205479452, + "grad_norm": 2.1530115604400635, + "learning_rate": 2.2014622258326564e-06, + "loss": 1.7854, + "step": 1668 + }, + { + "epoch": 1.8290410958904109, + "grad_norm": 2.003650188446045, + "learning_rate": 2.1994313566206336e-06, + "loss": 1.7953, + "step": 1669 + }, + { + "epoch": 1.83013698630137, + "grad_norm": 2.125225305557251, + "learning_rate": 2.197400487408611e-06, + "loss": 1.7215, + "step": 1670 + }, + { + "epoch": 1.83013698630137, + "eval_loss": 1.9451876878738403, + "eval_runtime": 37.6594, + "eval_samples_per_second": 26.554, + "eval_steps_per_second": 1.115, + "step": 1670 + }, + { + "epoch": 1.831232876712329, + "grad_norm": 2.058640956878662, + "learning_rate": 2.195369618196588e-06, + "loss": 1.7039, + "step": 1671 + }, + { + "epoch": 1.8323287671232875, + "grad_norm": 2.1126749515533447, + "learning_rate": 2.1933387489845655e-06, + "loss": 1.7385, + "step": 1672 + }, + { + "epoch": 1.8334246575342465, + "grad_norm": 2.0567469596862793, + "learning_rate": 2.1913078797725427e-06, + "loss": 1.7499, + "step": 1673 + }, + { + "epoch": 1.8345205479452056, + "grad_norm": 2.3803439140319824, + "learning_rate": 2.18927701056052e-06, + "loss": 1.8254, + "step": 1674 + }, + { + "epoch": 1.8356164383561644, + "grad_norm": 2.8527748584747314, + "learning_rate": 2.1872461413484973e-06, + "loss": 1.7606, + "step": 1675 + }, + { + "epoch": 1.8367123287671232, + "grad_norm": 2.1428439617156982, + "learning_rate": 2.1852152721364746e-06, + "loss": 1.7619, + "step": 1676 + }, + { + "epoch": 1.8378082191780822, + "grad_norm": 2.0377211570739746, + "learning_rate": 2.183184402924452e-06, + "loss": 1.7093, + "step": 1677 + }, + { + "epoch": 1.838904109589041, + "grad_norm": 1.9712910652160645, + "learning_rate": 2.181153533712429e-06, + "loss": 1.7788, + "step": 1678 + }, + { + "epoch": 1.8399999999999999, + "grad_norm": 2.0599145889282227, + "learning_rate": 2.1791226645004064e-06, + "loss": 1.7474, + "step": 1679 + }, + { + "epoch": 1.841095890410959, + "grad_norm": 1.9770036935806274, + "learning_rate": 2.1770917952883837e-06, + "loss": 1.7317, + "step": 1680 + }, + { + "epoch": 1.841095890410959, + "eval_loss": 1.9447766542434692, + "eval_runtime": 37.654, + "eval_samples_per_second": 26.558, + "eval_steps_per_second": 1.115, + "step": 1680 + }, + { + "epoch": 1.842191780821918, + "grad_norm": 2.1299326419830322, + "learning_rate": 2.175060926076361e-06, + "loss": 1.8011, + "step": 1681 + }, + { + "epoch": 1.8432876712328767, + "grad_norm": 2.134857416152954, + "learning_rate": 2.1730300568643382e-06, + "loss": 1.83, + "step": 1682 + }, + { + "epoch": 1.8443835616438355, + "grad_norm": 2.0131702423095703, + "learning_rate": 2.1709991876523155e-06, + "loss": 1.732, + "step": 1683 + }, + { + "epoch": 1.8454794520547946, + "grad_norm": 2.1677494049072266, + "learning_rate": 2.1689683184402928e-06, + "loss": 1.7118, + "step": 1684 + }, + { + "epoch": 1.8465753424657534, + "grad_norm": 2.325538158416748, + "learning_rate": 2.16693744922827e-06, + "loss": 1.7821, + "step": 1685 + }, + { + "epoch": 1.8476712328767122, + "grad_norm": 2.033168077468872, + "learning_rate": 2.1649065800162473e-06, + "loss": 1.7539, + "step": 1686 + }, + { + "epoch": 1.8487671232876712, + "grad_norm": 2.085318088531494, + "learning_rate": 2.1628757108042246e-06, + "loss": 1.725, + "step": 1687 + }, + { + "epoch": 1.8498630136986303, + "grad_norm": 1.9267048835754395, + "learning_rate": 2.160844841592202e-06, + "loss": 1.7066, + "step": 1688 + }, + { + "epoch": 1.850958904109589, + "grad_norm": 1.999786376953125, + "learning_rate": 2.1588139723801787e-06, + "loss": 1.7066, + "step": 1689 + }, + { + "epoch": 1.8520547945205479, + "grad_norm": 1.986025094985962, + "learning_rate": 2.156783103168156e-06, + "loss": 1.7326, + "step": 1690 + }, + { + "epoch": 1.8520547945205479, + "eval_loss": 1.9433118104934692, + "eval_runtime": 37.6538, + "eval_samples_per_second": 26.558, + "eval_steps_per_second": 1.115, + "step": 1690 + }, + { + "epoch": 1.853150684931507, + "grad_norm": 2.0098400115966797, + "learning_rate": 2.1547522339561333e-06, + "loss": 1.7078, + "step": 1691 + }, + { + "epoch": 1.8542465753424657, + "grad_norm": 1.9441081285476685, + "learning_rate": 2.1527213647441106e-06, + "loss": 1.7653, + "step": 1692 + }, + { + "epoch": 1.8553424657534245, + "grad_norm": 2.151911497116089, + "learning_rate": 2.150690495532088e-06, + "loss": 1.8066, + "step": 1693 + }, + { + "epoch": 1.8564383561643836, + "grad_norm": 2.04116153717041, + "learning_rate": 2.148659626320065e-06, + "loss": 1.7735, + "step": 1694 + }, + { + "epoch": 1.8575342465753426, + "grad_norm": 1.9916558265686035, + "learning_rate": 2.1466287571080424e-06, + "loss": 1.7543, + "step": 1695 + }, + { + "epoch": 1.8586301369863014, + "grad_norm": 2.1071197986602783, + "learning_rate": 2.1445978878960197e-06, + "loss": 1.7333, + "step": 1696 + }, + { + "epoch": 1.8597260273972602, + "grad_norm": 2.0217881202697754, + "learning_rate": 2.142567018683997e-06, + "loss": 1.7443, + "step": 1697 + }, + { + "epoch": 1.8608219178082193, + "grad_norm": 1.9675788879394531, + "learning_rate": 2.140536149471974e-06, + "loss": 1.7969, + "step": 1698 + }, + { + "epoch": 1.861917808219178, + "grad_norm": 2.0711920261383057, + "learning_rate": 2.1385052802599515e-06, + "loss": 1.7168, + "step": 1699 + }, + { + "epoch": 1.8630136986301369, + "grad_norm": 1.9705638885498047, + "learning_rate": 2.1364744110479288e-06, + "loss": 1.7211, + "step": 1700 + }, + { + "epoch": 1.8630136986301369, + "eval_loss": 1.941272497177124, + "eval_runtime": 37.6762, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 1700 + }, + { + "epoch": 1.864109589041096, + "grad_norm": 1.979668140411377, + "learning_rate": 2.134443541835906e-06, + "loss": 1.7783, + "step": 1701 + }, + { + "epoch": 1.865205479452055, + "grad_norm": 1.9930697679519653, + "learning_rate": 2.132412672623883e-06, + "loss": 1.7303, + "step": 1702 + }, + { + "epoch": 1.8663013698630135, + "grad_norm": 2.0750374794006348, + "learning_rate": 2.13038180341186e-06, + "loss": 1.7056, + "step": 1703 + }, + { + "epoch": 1.8673972602739726, + "grad_norm": 2.368161201477051, + "learning_rate": 2.1283509341998374e-06, + "loss": 1.7582, + "step": 1704 + }, + { + "epoch": 1.8684931506849316, + "grad_norm": 2.0597310066223145, + "learning_rate": 2.1263200649878147e-06, + "loss": 1.8028, + "step": 1705 + }, + { + "epoch": 1.8695890410958904, + "grad_norm": 1.947609543800354, + "learning_rate": 2.124289195775792e-06, + "loss": 1.7256, + "step": 1706 + }, + { + "epoch": 1.8706849315068492, + "grad_norm": 2.348658800125122, + "learning_rate": 2.1222583265637693e-06, + "loss": 1.748, + "step": 1707 + }, + { + "epoch": 1.8717808219178083, + "grad_norm": 1.988966464996338, + "learning_rate": 2.1202274573517465e-06, + "loss": 1.6602, + "step": 1708 + }, + { + "epoch": 1.872876712328767, + "grad_norm": 2.143030881881714, + "learning_rate": 2.118196588139724e-06, + "loss": 1.7074, + "step": 1709 + }, + { + "epoch": 1.8739726027397259, + "grad_norm": 2.046783685684204, + "learning_rate": 2.116165718927701e-06, + "loss": 1.7106, + "step": 1710 + }, + { + "epoch": 1.8739726027397259, + "eval_loss": 1.940579891204834, + "eval_runtime": 37.2642, + "eval_samples_per_second": 26.835, + "eval_steps_per_second": 1.127, + "step": 1710 + }, + { + "epoch": 1.875068493150685, + "grad_norm": 2.121497869491577, + "learning_rate": 2.1141348497156784e-06, + "loss": 1.7409, + "step": 1711 + }, + { + "epoch": 1.876164383561644, + "grad_norm": 2.0360069274902344, + "learning_rate": 2.1121039805036556e-06, + "loss": 1.7169, + "step": 1712 + }, + { + "epoch": 1.8772602739726028, + "grad_norm": 2.0190343856811523, + "learning_rate": 2.110073111291633e-06, + "loss": 1.7046, + "step": 1713 + }, + { + "epoch": 1.8783561643835616, + "grad_norm": 2.238844156265259, + "learning_rate": 2.10804224207961e-06, + "loss": 1.7483, + "step": 1714 + }, + { + "epoch": 1.8794520547945206, + "grad_norm": 1.9576469659805298, + "learning_rate": 2.1060113728675875e-06, + "loss": 1.7805, + "step": 1715 + }, + { + "epoch": 1.8805479452054794, + "grad_norm": 2.168437957763672, + "learning_rate": 2.1039805036555647e-06, + "loss": 1.7652, + "step": 1716 + }, + { + "epoch": 1.8816438356164382, + "grad_norm": 2.2372653484344482, + "learning_rate": 2.101949634443542e-06, + "loss": 1.6498, + "step": 1717 + }, + { + "epoch": 1.8827397260273973, + "grad_norm": 2.1206254959106445, + "learning_rate": 2.0999187652315193e-06, + "loss": 1.7426, + "step": 1718 + }, + { + "epoch": 1.8838356164383563, + "grad_norm": 1.9150874614715576, + "learning_rate": 2.0978878960194966e-06, + "loss": 1.6928, + "step": 1719 + }, + { + "epoch": 1.884931506849315, + "grad_norm": 2.3700966835021973, + "learning_rate": 2.095857026807474e-06, + "loss": 1.7584, + "step": 1720 + }, + { + "epoch": 1.884931506849315, + "eval_loss": 1.939314603805542, + "eval_runtime": 37.4061, + "eval_samples_per_second": 26.734, + "eval_steps_per_second": 1.123, + "step": 1720 + }, + { + "epoch": 1.886027397260274, + "grad_norm": 2.0733463764190674, + "learning_rate": 2.093826157595451e-06, + "loss": 1.7501, + "step": 1721 + }, + { + "epoch": 1.887123287671233, + "grad_norm": 1.9543569087982178, + "learning_rate": 2.0917952883834284e-06, + "loss": 1.7111, + "step": 1722 + }, + { + "epoch": 1.8882191780821918, + "grad_norm": 2.035921573638916, + "learning_rate": 2.0897644191714057e-06, + "loss": 1.7679, + "step": 1723 + }, + { + "epoch": 1.8893150684931506, + "grad_norm": 2.073953628540039, + "learning_rate": 2.087733549959383e-06, + "loss": 1.7863, + "step": 1724 + }, + { + "epoch": 1.8904109589041096, + "grad_norm": 2.1159274578094482, + "learning_rate": 2.0857026807473602e-06, + "loss": 1.6988, + "step": 1725 + }, + { + "epoch": 1.8915068493150686, + "grad_norm": 1.907313585281372, + "learning_rate": 2.0836718115353375e-06, + "loss": 1.6797, + "step": 1726 + }, + { + "epoch": 1.8926027397260274, + "grad_norm": 2.2749409675598145, + "learning_rate": 2.0816409423233144e-06, + "loss": 1.801, + "step": 1727 + }, + { + "epoch": 1.8936986301369862, + "grad_norm": 2.110377073287964, + "learning_rate": 2.0796100731112916e-06, + "loss": 1.7025, + "step": 1728 + }, + { + "epoch": 1.8947945205479453, + "grad_norm": 1.9137041568756104, + "learning_rate": 2.077579203899269e-06, + "loss": 1.7262, + "step": 1729 + }, + { + "epoch": 1.895890410958904, + "grad_norm": 2.1976664066314697, + "learning_rate": 2.075548334687246e-06, + "loss": 1.6892, + "step": 1730 + }, + { + "epoch": 1.895890410958904, + "eval_loss": 1.9381723403930664, + "eval_runtime": 37.4975, + "eval_samples_per_second": 26.668, + "eval_steps_per_second": 1.12, + "step": 1730 + }, + { + "epoch": 1.896986301369863, + "grad_norm": 2.191030502319336, + "learning_rate": 2.0735174654752235e-06, + "loss": 1.7521, + "step": 1731 + }, + { + "epoch": 1.898082191780822, + "grad_norm": 2.0503571033477783, + "learning_rate": 2.0714865962632007e-06, + "loss": 1.6853, + "step": 1732 + }, + { + "epoch": 1.899178082191781, + "grad_norm": 2.152529716491699, + "learning_rate": 2.069455727051178e-06, + "loss": 1.738, + "step": 1733 + }, + { + "epoch": 1.9002739726027398, + "grad_norm": 2.284041166305542, + "learning_rate": 2.0674248578391553e-06, + "loss": 1.7166, + "step": 1734 + }, + { + "epoch": 1.9013698630136986, + "grad_norm": 2.0965638160705566, + "learning_rate": 2.0653939886271326e-06, + "loss": 1.735, + "step": 1735 + }, + { + "epoch": 1.9024657534246576, + "grad_norm": 2.2820167541503906, + "learning_rate": 2.06336311941511e-06, + "loss": 1.7737, + "step": 1736 + }, + { + "epoch": 1.9035616438356164, + "grad_norm": 1.982020616531372, + "learning_rate": 2.061332250203087e-06, + "loss": 1.7198, + "step": 1737 + }, + { + "epoch": 1.9046575342465752, + "grad_norm": 1.9861208200454712, + "learning_rate": 2.0593013809910644e-06, + "loss": 1.6935, + "step": 1738 + }, + { + "epoch": 1.9057534246575343, + "grad_norm": 1.9662915468215942, + "learning_rate": 2.0572705117790417e-06, + "loss": 1.7297, + "step": 1739 + }, + { + "epoch": 1.9068493150684933, + "grad_norm": 2.0439398288726807, + "learning_rate": 2.055239642567019e-06, + "loss": 1.7678, + "step": 1740 + }, + { + "epoch": 1.9068493150684933, + "eval_loss": 1.9400287866592407, + "eval_runtime": 37.5492, + "eval_samples_per_second": 26.632, + "eval_steps_per_second": 1.119, + "step": 1740 + }, + { + "epoch": 1.907945205479452, + "grad_norm": 2.0104117393493652, + "learning_rate": 2.0532087733549962e-06, + "loss": 1.747, + "step": 1741 + }, + { + "epoch": 1.909041095890411, + "grad_norm": 2.0162367820739746, + "learning_rate": 2.0511779041429735e-06, + "loss": 1.7472, + "step": 1742 + }, + { + "epoch": 1.91013698630137, + "grad_norm": 2.0113325119018555, + "learning_rate": 2.0491470349309508e-06, + "loss": 1.7131, + "step": 1743 + }, + { + "epoch": 1.9112328767123288, + "grad_norm": 2.220263957977295, + "learning_rate": 2.047116165718928e-06, + "loss": 1.7515, + "step": 1744 + }, + { + "epoch": 1.9123287671232876, + "grad_norm": 2.0188612937927246, + "learning_rate": 2.0450852965069053e-06, + "loss": 1.7453, + "step": 1745 + }, + { + "epoch": 1.9134246575342466, + "grad_norm": 1.929147481918335, + "learning_rate": 2.0430544272948826e-06, + "loss": 1.7991, + "step": 1746 + }, + { + "epoch": 1.9145205479452054, + "grad_norm": 2.0564963817596436, + "learning_rate": 2.04102355808286e-06, + "loss": 1.734, + "step": 1747 + }, + { + "epoch": 1.9156164383561642, + "grad_norm": 2.012883186340332, + "learning_rate": 2.038992688870837e-06, + "loss": 1.6779, + "step": 1748 + }, + { + "epoch": 1.9167123287671233, + "grad_norm": 2.0084073543548584, + "learning_rate": 2.0369618196588144e-06, + "loss": 1.7054, + "step": 1749 + }, + { + "epoch": 1.9178082191780823, + "grad_norm": 1.972630262374878, + "learning_rate": 2.0349309504467917e-06, + "loss": 1.717, + "step": 1750 + }, + { + "epoch": 1.9178082191780823, + "eval_loss": 1.9397809505462646, + "eval_runtime": 37.5324, + "eval_samples_per_second": 26.644, + "eval_steps_per_second": 1.119, + "step": 1750 + }, + { + "epoch": 1.9189041095890411, + "grad_norm": 2.020827293395996, + "learning_rate": 2.032900081234769e-06, + "loss": 1.7623, + "step": 1751 + }, + { + "epoch": 1.92, + "grad_norm": 1.8884387016296387, + "learning_rate": 2.030869212022746e-06, + "loss": 1.7724, + "step": 1752 + }, + { + "epoch": 1.921095890410959, + "grad_norm": 1.9301886558532715, + "learning_rate": 2.028838342810723e-06, + "loss": 1.7308, + "step": 1753 + }, + { + "epoch": 1.9221917808219178, + "grad_norm": 1.9553035497665405, + "learning_rate": 2.0268074735987004e-06, + "loss": 1.7143, + "step": 1754 + }, + { + "epoch": 1.9232876712328766, + "grad_norm": 2.130143165588379, + "learning_rate": 2.0247766043866777e-06, + "loss": 1.7022, + "step": 1755 + }, + { + "epoch": 1.9243835616438356, + "grad_norm": 1.955505609512329, + "learning_rate": 2.022745735174655e-06, + "loss": 1.7804, + "step": 1756 + }, + { + "epoch": 1.9254794520547946, + "grad_norm": 2.070336103439331, + "learning_rate": 2.020714865962632e-06, + "loss": 1.7459, + "step": 1757 + }, + { + "epoch": 1.9265753424657535, + "grad_norm": 1.9322236776351929, + "learning_rate": 2.0186839967506095e-06, + "loss": 1.7644, + "step": 1758 + }, + { + "epoch": 1.9276712328767123, + "grad_norm": 2.033029556274414, + "learning_rate": 2.0166531275385868e-06, + "loss": 1.7389, + "step": 1759 + }, + { + "epoch": 1.9287671232876713, + "grad_norm": 1.879835605621338, + "learning_rate": 2.014622258326564e-06, + "loss": 1.7037, + "step": 1760 + }, + { + "epoch": 1.9287671232876713, + "eval_loss": 1.9346613883972168, + "eval_runtime": 37.53, + "eval_samples_per_second": 26.645, + "eval_steps_per_second": 1.119, + "step": 1760 + }, + { + "epoch": 1.9298630136986301, + "grad_norm": 1.9985371828079224, + "learning_rate": 2.0125913891145413e-06, + "loss": 1.7497, + "step": 1761 + }, + { + "epoch": 1.930958904109589, + "grad_norm": 2.2449045181274414, + "learning_rate": 2.0105605199025186e-06, + "loss": 1.7356, + "step": 1762 + }, + { + "epoch": 1.932054794520548, + "grad_norm": 1.9392081499099731, + "learning_rate": 2.008529650690496e-06, + "loss": 1.701, + "step": 1763 + }, + { + "epoch": 1.933150684931507, + "grad_norm": 2.14847469329834, + "learning_rate": 2.0064987814784727e-06, + "loss": 1.7607, + "step": 1764 + }, + { + "epoch": 1.9342465753424658, + "grad_norm": 2.1175179481506348, + "learning_rate": 2.00446791226645e-06, + "loss": 1.7172, + "step": 1765 + }, + { + "epoch": 1.9353424657534246, + "grad_norm": 1.955820918083191, + "learning_rate": 2.0024370430544273e-06, + "loss": 1.7179, + "step": 1766 + }, + { + "epoch": 1.9364383561643836, + "grad_norm": 2.132200241088867, + "learning_rate": 2.0004061738424045e-06, + "loss": 1.7273, + "step": 1767 + }, + { + "epoch": 1.9375342465753425, + "grad_norm": 2.010092258453369, + "learning_rate": 1.998375304630382e-06, + "loss": 1.7799, + "step": 1768 + }, + { + "epoch": 1.9386301369863013, + "grad_norm": 2.1509854793548584, + "learning_rate": 1.996344435418359e-06, + "loss": 1.7387, + "step": 1769 + }, + { + "epoch": 1.9397260273972603, + "grad_norm": 2.038088083267212, + "learning_rate": 1.9943135662063364e-06, + "loss": 1.759, + "step": 1770 + }, + { + "epoch": 1.9397260273972603, + "eval_loss": 1.9372096061706543, + "eval_runtime": 37.5206, + "eval_samples_per_second": 26.652, + "eval_steps_per_second": 1.119, + "step": 1770 + }, + { + "epoch": 1.9408219178082193, + "grad_norm": 2.068157434463501, + "learning_rate": 1.9922826969943136e-06, + "loss": 1.742, + "step": 1771 + }, + { + "epoch": 1.941917808219178, + "grad_norm": 2.1352615356445312, + "learning_rate": 1.990251827782291e-06, + "loss": 1.7749, + "step": 1772 + }, + { + "epoch": 1.943013698630137, + "grad_norm": 1.9305949211120605, + "learning_rate": 1.988220958570268e-06, + "loss": 1.7521, + "step": 1773 + }, + { + "epoch": 1.944109589041096, + "grad_norm": 2.052032232284546, + "learning_rate": 1.9861900893582455e-06, + "loss": 1.7603, + "step": 1774 + }, + { + "epoch": 1.9452054794520548, + "grad_norm": 2.0203394889831543, + "learning_rate": 1.9841592201462227e-06, + "loss": 1.7699, + "step": 1775 + }, + { + "epoch": 1.9463013698630136, + "grad_norm": 2.469848871231079, + "learning_rate": 1.9821283509342e-06, + "loss": 1.693, + "step": 1776 + }, + { + "epoch": 1.9473972602739726, + "grad_norm": 1.9579914808273315, + "learning_rate": 1.9800974817221773e-06, + "loss": 1.7018, + "step": 1777 + }, + { + "epoch": 1.9484931506849315, + "grad_norm": 2.114840269088745, + "learning_rate": 1.9780666125101546e-06, + "loss": 1.7514, + "step": 1778 + }, + { + "epoch": 1.9495890410958903, + "grad_norm": 2.061809539794922, + "learning_rate": 1.976035743298132e-06, + "loss": 1.7095, + "step": 1779 + }, + { + "epoch": 1.9506849315068493, + "grad_norm": 1.9839788675308228, + "learning_rate": 1.974004874086109e-06, + "loss": 1.7928, + "step": 1780 + }, + { + "epoch": 1.9506849315068493, + "eval_loss": 1.941945195198059, + "eval_runtime": 37.518, + "eval_samples_per_second": 26.654, + "eval_steps_per_second": 1.119, + "step": 1780 + }, + { + "epoch": 1.9517808219178083, + "grad_norm": 1.9607301950454712, + "learning_rate": 1.9719740048740864e-06, + "loss": 1.6679, + "step": 1781 + }, + { + "epoch": 1.9528767123287671, + "grad_norm": 2.0004162788391113, + "learning_rate": 1.9699431356620637e-06, + "loss": 1.716, + "step": 1782 + }, + { + "epoch": 1.953972602739726, + "grad_norm": 2.09340763092041, + "learning_rate": 1.967912266450041e-06, + "loss": 1.7144, + "step": 1783 + }, + { + "epoch": 1.955068493150685, + "grad_norm": 1.997462511062622, + "learning_rate": 1.9658813972380182e-06, + "loss": 1.7049, + "step": 1784 + }, + { + "epoch": 1.9561643835616438, + "grad_norm": 1.9848195314407349, + "learning_rate": 1.9638505280259955e-06, + "loss": 1.8262, + "step": 1785 + }, + { + "epoch": 1.9572602739726026, + "grad_norm": 2.0498056411743164, + "learning_rate": 1.9618196588139728e-06, + "loss": 1.6989, + "step": 1786 + }, + { + "epoch": 1.9583561643835616, + "grad_norm": 2.038849115371704, + "learning_rate": 1.95978878960195e-06, + "loss": 1.74, + "step": 1787 + }, + { + "epoch": 1.9594520547945207, + "grad_norm": 1.9121922254562378, + "learning_rate": 1.9577579203899273e-06, + "loss": 1.7999, + "step": 1788 + }, + { + "epoch": 1.9605479452054795, + "grad_norm": 1.8990696668624878, + "learning_rate": 1.955727051177904e-06, + "loss": 1.7795, + "step": 1789 + }, + { + "epoch": 1.9616438356164383, + "grad_norm": 1.983232021331787, + "learning_rate": 1.9536961819658815e-06, + "loss": 1.7099, + "step": 1790 + }, + { + "epoch": 1.9616438356164383, + "eval_loss": 1.9464988708496094, + "eval_runtime": 37.5024, + "eval_samples_per_second": 26.665, + "eval_steps_per_second": 1.12, + "step": 1790 + }, + { + "epoch": 1.9627397260273973, + "grad_norm": 1.9355418682098389, + "learning_rate": 1.9516653127538587e-06, + "loss": 1.6934, + "step": 1791 + }, + { + "epoch": 1.9638356164383561, + "grad_norm": 2.057600736618042, + "learning_rate": 1.949634443541836e-06, + "loss": 1.6919, + "step": 1792 + }, + { + "epoch": 1.964931506849315, + "grad_norm": 2.046416759490967, + "learning_rate": 1.9476035743298133e-06, + "loss": 1.6956, + "step": 1793 + }, + { + "epoch": 1.966027397260274, + "grad_norm": 2.022989273071289, + "learning_rate": 1.9455727051177906e-06, + "loss": 1.7653, + "step": 1794 + }, + { + "epoch": 1.967123287671233, + "grad_norm": 2.277711868286133, + "learning_rate": 1.943541835905768e-06, + "loss": 1.7288, + "step": 1795 + }, + { + "epoch": 1.9682191780821918, + "grad_norm": 1.9851937294006348, + "learning_rate": 1.941510966693745e-06, + "loss": 1.779, + "step": 1796 + }, + { + "epoch": 1.9693150684931506, + "grad_norm": 2.059169054031372, + "learning_rate": 1.9394800974817224e-06, + "loss": 1.6873, + "step": 1797 + }, + { + "epoch": 1.9704109589041097, + "grad_norm": 1.9894514083862305, + "learning_rate": 1.9374492282696997e-06, + "loss": 1.7645, + "step": 1798 + }, + { + "epoch": 1.9715068493150685, + "grad_norm": 2.059819459915161, + "learning_rate": 1.935418359057677e-06, + "loss": 1.7939, + "step": 1799 + }, + { + "epoch": 1.9726027397260273, + "grad_norm": 1.8777337074279785, + "learning_rate": 1.9333874898456542e-06, + "loss": 1.7561, + "step": 1800 + }, + { + "epoch": 1.9726027397260273, + "eval_loss": 1.9419773817062378, + "eval_runtime": 37.5263, + "eval_samples_per_second": 26.648, + "eval_steps_per_second": 1.119, + "step": 1800 + }, + { + "epoch": 1.9736986301369863, + "grad_norm": 2.049135446548462, + "learning_rate": 1.931356620633631e-06, + "loss": 1.7762, + "step": 1801 + }, + { + "epoch": 1.9747945205479454, + "grad_norm": 2.0334010124206543, + "learning_rate": 1.9293257514216083e-06, + "loss": 1.678, + "step": 1802 + }, + { + "epoch": 1.9758904109589042, + "grad_norm": 2.0800070762634277, + "learning_rate": 1.9272948822095856e-06, + "loss": 1.7344, + "step": 1803 + }, + { + "epoch": 1.976986301369863, + "grad_norm": 2.1043014526367188, + "learning_rate": 1.925264012997563e-06, + "loss": 1.6738, + "step": 1804 + }, + { + "epoch": 1.978082191780822, + "grad_norm": 1.9936349391937256, + "learning_rate": 1.92323314378554e-06, + "loss": 1.7321, + "step": 1805 + }, + { + "epoch": 1.9791780821917808, + "grad_norm": 1.977447271347046, + "learning_rate": 1.9212022745735174e-06, + "loss": 1.7599, + "step": 1806 + }, + { + "epoch": 1.9802739726027396, + "grad_norm": 2.0987250804901123, + "learning_rate": 1.9191714053614947e-06, + "loss": 1.6954, + "step": 1807 + }, + { + "epoch": 1.9813698630136987, + "grad_norm": 2.0406241416931152, + "learning_rate": 1.917140536149472e-06, + "loss": 1.6982, + "step": 1808 + }, + { + "epoch": 1.9824657534246577, + "grad_norm": 1.9119296073913574, + "learning_rate": 1.9151096669374493e-06, + "loss": 1.6717, + "step": 1809 + }, + { + "epoch": 1.9835616438356163, + "grad_norm": 2.2667031288146973, + "learning_rate": 1.9130787977254266e-06, + "loss": 1.798, + "step": 1810 + }, + { + "epoch": 1.9835616438356163, + "eval_loss": 1.9404252767562866, + "eval_runtime": 37.2788, + "eval_samples_per_second": 26.825, + "eval_steps_per_second": 1.127, + "step": 1810 + }, + { + "epoch": 1.9846575342465753, + "grad_norm": 2.0808026790618896, + "learning_rate": 1.911047928513404e-06, + "loss": 1.7559, + "step": 1811 + }, + { + "epoch": 1.9857534246575343, + "grad_norm": 2.097626209259033, + "learning_rate": 1.909017059301381e-06, + "loss": 1.6999, + "step": 1812 + }, + { + "epoch": 1.9868493150684932, + "grad_norm": 2.0065059661865234, + "learning_rate": 1.9069861900893586e-06, + "loss": 1.7037, + "step": 1813 + }, + { + "epoch": 1.987945205479452, + "grad_norm": 1.8728374242782593, + "learning_rate": 1.9049553208773354e-06, + "loss": 1.759, + "step": 1814 + }, + { + "epoch": 1.989041095890411, + "grad_norm": 2.369331121444702, + "learning_rate": 1.9029244516653127e-06, + "loss": 1.7124, + "step": 1815 + }, + { + "epoch": 1.9901369863013698, + "grad_norm": 2.0041840076446533, + "learning_rate": 1.90089358245329e-06, + "loss": 1.661, + "step": 1816 + }, + { + "epoch": 1.9912328767123286, + "grad_norm": 2.1467182636260986, + "learning_rate": 1.8988627132412673e-06, + "loss": 1.7145, + "step": 1817 + }, + { + "epoch": 1.9923287671232877, + "grad_norm": 2.373904228210449, + "learning_rate": 1.8968318440292445e-06, + "loss": 1.7563, + "step": 1818 + }, + { + "epoch": 1.9934246575342467, + "grad_norm": 2.093635082244873, + "learning_rate": 1.8948009748172218e-06, + "loss": 1.666, + "step": 1819 + }, + { + "epoch": 1.9945205479452055, + "grad_norm": 2.120913505554199, + "learning_rate": 1.892770105605199e-06, + "loss": 1.7626, + "step": 1820 + }, + { + "epoch": 1.9945205479452055, + "eval_loss": 1.9345192909240723, + "eval_runtime": 37.475, + "eval_samples_per_second": 26.684, + "eval_steps_per_second": 1.121, + "step": 1820 + }, + { + "epoch": 1.9956164383561643, + "grad_norm": 1.9989229440689087, + "learning_rate": 1.8907392363931764e-06, + "loss": 1.7584, + "step": 1821 + }, + { + "epoch": 1.9967123287671233, + "grad_norm": 1.961505651473999, + "learning_rate": 1.8887083671811536e-06, + "loss": 1.7206, + "step": 1822 + }, + { + "epoch": 1.9978082191780822, + "grad_norm": 1.9231300354003906, + "learning_rate": 1.886677497969131e-06, + "loss": 1.7375, + "step": 1823 + }, + { + "epoch": 1.998904109589041, + "grad_norm": 1.9450154304504395, + "learning_rate": 1.8846466287571082e-06, + "loss": 1.6322, + "step": 1824 + }, + { + "epoch": 2.0, + "grad_norm": 2.037381649017334, + "learning_rate": 1.8826157595450855e-06, + "loss": 1.7536, + "step": 1825 + }, + { + "epoch": 2.001095890410959, + "grad_norm": 2.1451215744018555, + "learning_rate": 1.8805848903330625e-06, + "loss": 1.6514, + "step": 1826 + }, + { + "epoch": 2.0021917808219176, + "grad_norm": 2.216177225112915, + "learning_rate": 1.8785540211210398e-06, + "loss": 1.8134, + "step": 1827 + }, + { + "epoch": 2.0032876712328767, + "grad_norm": 2.0705394744873047, + "learning_rate": 1.876523151909017e-06, + "loss": 1.7279, + "step": 1828 + }, + { + "epoch": 2.0043835616438357, + "grad_norm": 2.261056661605835, + "learning_rate": 1.8744922826969944e-06, + "loss": 1.6768, + "step": 1829 + }, + { + "epoch": 2.0054794520547947, + "grad_norm": 2.4416584968566895, + "learning_rate": 1.8724614134849716e-06, + "loss": 1.6638, + "step": 1830 + }, + { + "epoch": 2.0054794520547947, + "eval_loss": 1.9376901388168335, + "eval_runtime": 37.5572, + "eval_samples_per_second": 26.626, + "eval_steps_per_second": 1.118, + "step": 1830 + }, + { + "epoch": 2.0065753424657533, + "grad_norm": 2.193220615386963, + "learning_rate": 1.870430544272949e-06, + "loss": 1.6861, + "step": 1831 + }, + { + "epoch": 2.0076712328767123, + "grad_norm": 2.0660276412963867, + "learning_rate": 1.8683996750609262e-06, + "loss": 1.793, + "step": 1832 + }, + { + "epoch": 2.0087671232876714, + "grad_norm": 2.226266622543335, + "learning_rate": 1.8663688058489035e-06, + "loss": 1.68, + "step": 1833 + }, + { + "epoch": 2.00986301369863, + "grad_norm": 2.3646411895751953, + "learning_rate": 1.8643379366368807e-06, + "loss": 1.6938, + "step": 1834 + }, + { + "epoch": 2.010958904109589, + "grad_norm": 2.390726327896118, + "learning_rate": 1.862307067424858e-06, + "loss": 1.7118, + "step": 1835 + }, + { + "epoch": 2.012054794520548, + "grad_norm": 2.257854461669922, + "learning_rate": 1.8602761982128353e-06, + "loss": 1.7064, + "step": 1836 + }, + { + "epoch": 2.013150684931507, + "grad_norm": 2.3134212493896484, + "learning_rate": 1.8582453290008126e-06, + "loss": 1.7517, + "step": 1837 + }, + { + "epoch": 2.0142465753424657, + "grad_norm": 2.399315357208252, + "learning_rate": 1.8562144597887898e-06, + "loss": 1.6798, + "step": 1838 + }, + { + "epoch": 2.0153424657534247, + "grad_norm": 2.262643814086914, + "learning_rate": 1.854183590576767e-06, + "loss": 1.6974, + "step": 1839 + }, + { + "epoch": 2.0164383561643837, + "grad_norm": 2.1609957218170166, + "learning_rate": 1.8521527213647442e-06, + "loss": 1.7383, + "step": 1840 + }, + { + "epoch": 2.0164383561643837, + "eval_loss": 1.9400956630706787, + "eval_runtime": 37.5932, + "eval_samples_per_second": 26.601, + "eval_steps_per_second": 1.117, + "step": 1840 + }, + { + "epoch": 2.0175342465753423, + "grad_norm": 2.2183380126953125, + "learning_rate": 1.8501218521527215e-06, + "loss": 1.6514, + "step": 1841 + }, + { + "epoch": 2.0186301369863013, + "grad_norm": 2.4489705562591553, + "learning_rate": 1.8480909829406987e-06, + "loss": 1.624, + "step": 1842 + }, + { + "epoch": 2.0197260273972604, + "grad_norm": 2.357877016067505, + "learning_rate": 1.846060113728676e-06, + "loss": 1.6499, + "step": 1843 + }, + { + "epoch": 2.020821917808219, + "grad_norm": 2.161865472793579, + "learning_rate": 1.8440292445166533e-06, + "loss": 1.704, + "step": 1844 + }, + { + "epoch": 2.021917808219178, + "grad_norm": 2.388451337814331, + "learning_rate": 1.8419983753046306e-06, + "loss": 1.6626, + "step": 1845 + }, + { + "epoch": 2.023013698630137, + "grad_norm": 2.3482980728149414, + "learning_rate": 1.8399675060926078e-06, + "loss": 1.7347, + "step": 1846 + }, + { + "epoch": 2.024109589041096, + "grad_norm": 2.3431859016418457, + "learning_rate": 1.8379366368805851e-06, + "loss": 1.71, + "step": 1847 + }, + { + "epoch": 2.0252054794520546, + "grad_norm": 2.30924129486084, + "learning_rate": 1.8359057676685624e-06, + "loss": 1.6768, + "step": 1848 + }, + { + "epoch": 2.0263013698630137, + "grad_norm": 2.2544329166412354, + "learning_rate": 1.8338748984565397e-06, + "loss": 1.6454, + "step": 1849 + }, + { + "epoch": 2.0273972602739727, + "grad_norm": 2.282646656036377, + "learning_rate": 1.831844029244517e-06, + "loss": 1.6812, + "step": 1850 + }, + { + "epoch": 2.0273972602739727, + "eval_loss": 1.9371191263198853, + "eval_runtime": 37.5968, + "eval_samples_per_second": 26.598, + "eval_steps_per_second": 1.117, + "step": 1850 + }, + { + "epoch": 2.0284931506849313, + "grad_norm": 2.1591529846191406, + "learning_rate": 1.829813160032494e-06, + "loss": 1.6457, + "step": 1851 + }, + { + "epoch": 2.0295890410958903, + "grad_norm": 2.1543376445770264, + "learning_rate": 1.8277822908204713e-06, + "loss": 1.6693, + "step": 1852 + }, + { + "epoch": 2.0306849315068494, + "grad_norm": 2.351182699203491, + "learning_rate": 1.8257514216084486e-06, + "loss": 1.7067, + "step": 1853 + }, + { + "epoch": 2.0317808219178084, + "grad_norm": 2.354083299636841, + "learning_rate": 1.8237205523964258e-06, + "loss": 1.6863, + "step": 1854 + }, + { + "epoch": 2.032876712328767, + "grad_norm": 2.2278695106506348, + "learning_rate": 1.8216896831844031e-06, + "loss": 1.7006, + "step": 1855 + }, + { + "epoch": 2.033972602739726, + "grad_norm": 2.2142486572265625, + "learning_rate": 1.8196588139723804e-06, + "loss": 1.7451, + "step": 1856 + }, + { + "epoch": 2.035068493150685, + "grad_norm": 2.348388433456421, + "learning_rate": 1.8176279447603577e-06, + "loss": 1.6306, + "step": 1857 + }, + { + "epoch": 2.0361643835616436, + "grad_norm": 2.6112802028656006, + "learning_rate": 1.815597075548335e-06, + "loss": 1.6695, + "step": 1858 + }, + { + "epoch": 2.0372602739726027, + "grad_norm": 2.3125290870666504, + "learning_rate": 1.8135662063363122e-06, + "loss": 1.697, + "step": 1859 + }, + { + "epoch": 2.0383561643835617, + "grad_norm": 2.2591099739074707, + "learning_rate": 1.8115353371242895e-06, + "loss": 1.7316, + "step": 1860 + }, + { + "epoch": 2.0383561643835617, + "eval_loss": 1.944149374961853, + "eval_runtime": 37.5996, + "eval_samples_per_second": 26.596, + "eval_steps_per_second": 1.117, + "step": 1860 + }, + { + "epoch": 2.0394520547945207, + "grad_norm": 2.5762269496917725, + "learning_rate": 1.8095044679122668e-06, + "loss": 1.6358, + "step": 1861 + }, + { + "epoch": 2.0405479452054793, + "grad_norm": 2.259580373764038, + "learning_rate": 1.807473598700244e-06, + "loss": 1.6748, + "step": 1862 + }, + { + "epoch": 2.0416438356164384, + "grad_norm": 2.3359739780426025, + "learning_rate": 1.805442729488221e-06, + "loss": 1.6656, + "step": 1863 + }, + { + "epoch": 2.0427397260273974, + "grad_norm": 2.4480202198028564, + "learning_rate": 1.8034118602761982e-06, + "loss": 1.6973, + "step": 1864 + }, + { + "epoch": 2.043835616438356, + "grad_norm": 2.1894607543945312, + "learning_rate": 1.8013809910641754e-06, + "loss": 1.6933, + "step": 1865 + }, + { + "epoch": 2.044931506849315, + "grad_norm": 2.2062199115753174, + "learning_rate": 1.7993501218521527e-06, + "loss": 1.6931, + "step": 1866 + }, + { + "epoch": 2.046027397260274, + "grad_norm": 2.568087339401245, + "learning_rate": 1.79731925264013e-06, + "loss": 1.6388, + "step": 1867 + }, + { + "epoch": 2.047123287671233, + "grad_norm": 2.3496286869049072, + "learning_rate": 1.7952883834281073e-06, + "loss": 1.7154, + "step": 1868 + }, + { + "epoch": 2.0482191780821917, + "grad_norm": 2.1612911224365234, + "learning_rate": 1.7932575142160846e-06, + "loss": 1.6577, + "step": 1869 + }, + { + "epoch": 2.0493150684931507, + "grad_norm": 2.4662342071533203, + "learning_rate": 1.7912266450040618e-06, + "loss": 1.7104, + "step": 1870 + }, + { + "epoch": 2.0493150684931507, + "eval_loss": 1.9412809610366821, + "eval_runtime": 37.5984, + "eval_samples_per_second": 26.597, + "eval_steps_per_second": 1.117, + "step": 1870 + }, + { + "epoch": 2.0504109589041097, + "grad_norm": 2.6574866771698, + "learning_rate": 1.789195775792039e-06, + "loss": 1.7146, + "step": 1871 + }, + { + "epoch": 2.0515068493150683, + "grad_norm": 2.208397388458252, + "learning_rate": 1.7871649065800164e-06, + "loss": 1.6098, + "step": 1872 + }, + { + "epoch": 2.0526027397260274, + "grad_norm": 2.366791009902954, + "learning_rate": 1.7851340373679937e-06, + "loss": 1.7543, + "step": 1873 + }, + { + "epoch": 2.0536986301369864, + "grad_norm": 2.2281816005706787, + "learning_rate": 1.783103168155971e-06, + "loss": 1.6461, + "step": 1874 + }, + { + "epoch": 2.0547945205479454, + "grad_norm": 2.212224245071411, + "learning_rate": 1.7810722989439482e-06, + "loss": 1.6629, + "step": 1875 + }, + { + "epoch": 2.055890410958904, + "grad_norm": 2.206946611404419, + "learning_rate": 1.7790414297319253e-06, + "loss": 1.6917, + "step": 1876 + }, + { + "epoch": 2.056986301369863, + "grad_norm": 2.1313860416412354, + "learning_rate": 1.7770105605199025e-06, + "loss": 1.6515, + "step": 1877 + }, + { + "epoch": 2.058082191780822, + "grad_norm": 2.247718572616577, + "learning_rate": 1.7749796913078798e-06, + "loss": 1.6462, + "step": 1878 + }, + { + "epoch": 2.0591780821917807, + "grad_norm": 2.2024714946746826, + "learning_rate": 1.772948822095857e-06, + "loss": 1.7021, + "step": 1879 + }, + { + "epoch": 2.0602739726027397, + "grad_norm": 2.155611515045166, + "learning_rate": 1.7709179528838344e-06, + "loss": 1.7425, + "step": 1880 + }, + { + "epoch": 2.0602739726027397, + "eval_loss": 1.9357531070709229, + "eval_runtime": 37.6069, + "eval_samples_per_second": 26.591, + "eval_steps_per_second": 1.117, + "step": 1880 + }, + { + "epoch": 2.0613698630136987, + "grad_norm": 2.1168251037597656, + "learning_rate": 1.7688870836718116e-06, + "loss": 1.7279, + "step": 1881 + }, + { + "epoch": 2.0624657534246573, + "grad_norm": 2.420991897583008, + "learning_rate": 1.766856214459789e-06, + "loss": 1.6794, + "step": 1882 + }, + { + "epoch": 2.0635616438356164, + "grad_norm": 2.290802001953125, + "learning_rate": 1.7648253452477662e-06, + "loss": 1.6749, + "step": 1883 + }, + { + "epoch": 2.0646575342465754, + "grad_norm": 2.2239034175872803, + "learning_rate": 1.7627944760357435e-06, + "loss": 1.6624, + "step": 1884 + }, + { + "epoch": 2.0657534246575344, + "grad_norm": 2.248377799987793, + "learning_rate": 1.7607636068237208e-06, + "loss": 1.7431, + "step": 1885 + }, + { + "epoch": 2.066849315068493, + "grad_norm": 2.428326368331909, + "learning_rate": 1.758732737611698e-06, + "loss": 1.7145, + "step": 1886 + }, + { + "epoch": 2.067945205479452, + "grad_norm": 2.3253471851348877, + "learning_rate": 1.7567018683996753e-06, + "loss": 1.6947, + "step": 1887 + }, + { + "epoch": 2.069041095890411, + "grad_norm": 2.386906623840332, + "learning_rate": 1.7546709991876524e-06, + "loss": 1.6723, + "step": 1888 + }, + { + "epoch": 2.0701369863013697, + "grad_norm": 2.076267957687378, + "learning_rate": 1.7526401299756296e-06, + "loss": 1.6877, + "step": 1889 + }, + { + "epoch": 2.0712328767123287, + "grad_norm": 2.1380043029785156, + "learning_rate": 1.750609260763607e-06, + "loss": 1.6879, + "step": 1890 + }, + { + "epoch": 2.0712328767123287, + "eval_loss": 1.9421926736831665, + "eval_runtime": 37.6339, + "eval_samples_per_second": 26.572, + "eval_steps_per_second": 1.116, + "step": 1890 + }, + { + "epoch": 2.0723287671232877, + "grad_norm": 2.2633447647094727, + "learning_rate": 1.7485783915515842e-06, + "loss": 1.6757, + "step": 1891 + }, + { + "epoch": 2.0734246575342468, + "grad_norm": 2.253002882003784, + "learning_rate": 1.7465475223395615e-06, + "loss": 1.6757, + "step": 1892 + }, + { + "epoch": 2.0745205479452054, + "grad_norm": 2.2368297576904297, + "learning_rate": 1.7445166531275387e-06, + "loss": 1.6518, + "step": 1893 + }, + { + "epoch": 2.0756164383561644, + "grad_norm": 2.3998613357543945, + "learning_rate": 1.742485783915516e-06, + "loss": 1.7935, + "step": 1894 + }, + { + "epoch": 2.0767123287671234, + "grad_norm": 2.323230743408203, + "learning_rate": 1.7404549147034933e-06, + "loss": 1.7621, + "step": 1895 + }, + { + "epoch": 2.077808219178082, + "grad_norm": 2.468967914581299, + "learning_rate": 1.7384240454914706e-06, + "loss": 1.6097, + "step": 1896 + }, + { + "epoch": 2.078904109589041, + "grad_norm": 3.209561586380005, + "learning_rate": 1.7363931762794478e-06, + "loss": 1.6866, + "step": 1897 + }, + { + "epoch": 2.08, + "grad_norm": 2.26240611076355, + "learning_rate": 1.7343623070674251e-06, + "loss": 1.7269, + "step": 1898 + }, + { + "epoch": 2.081095890410959, + "grad_norm": 2.625722885131836, + "learning_rate": 1.7323314378554024e-06, + "loss": 1.6778, + "step": 1899 + }, + { + "epoch": 2.0821917808219177, + "grad_norm": 2.4800055027008057, + "learning_rate": 1.7303005686433797e-06, + "loss": 1.6867, + "step": 1900 + }, + { + "epoch": 2.0821917808219177, + "eval_loss": 1.9384645223617554, + "eval_runtime": 37.6202, + "eval_samples_per_second": 26.581, + "eval_steps_per_second": 1.116, + "step": 1900 + }, + { + "epoch": 2.0832876712328767, + "grad_norm": 2.2881827354431152, + "learning_rate": 1.7282696994313567e-06, + "loss": 1.674, + "step": 1901 + }, + { + "epoch": 2.0843835616438358, + "grad_norm": 2.640097141265869, + "learning_rate": 1.726238830219334e-06, + "loss": 1.6879, + "step": 1902 + }, + { + "epoch": 2.0854794520547943, + "grad_norm": 2.3600308895111084, + "learning_rate": 1.7242079610073113e-06, + "loss": 1.6793, + "step": 1903 + }, + { + "epoch": 2.0865753424657534, + "grad_norm": 2.4701266288757324, + "learning_rate": 1.7221770917952886e-06, + "loss": 1.7504, + "step": 1904 + }, + { + "epoch": 2.0876712328767124, + "grad_norm": 2.3872451782226562, + "learning_rate": 1.7201462225832658e-06, + "loss": 1.7028, + "step": 1905 + }, + { + "epoch": 2.0887671232876714, + "grad_norm": 2.306307315826416, + "learning_rate": 1.7181153533712431e-06, + "loss": 1.7063, + "step": 1906 + }, + { + "epoch": 2.08986301369863, + "grad_norm": 2.2499122619628906, + "learning_rate": 1.7160844841592204e-06, + "loss": 1.6069, + "step": 1907 + }, + { + "epoch": 2.090958904109589, + "grad_norm": 2.276212215423584, + "learning_rate": 1.7140536149471977e-06, + "loss": 1.6452, + "step": 1908 + }, + { + "epoch": 2.092054794520548, + "grad_norm": 2.1463613510131836, + "learning_rate": 1.712022745735175e-06, + "loss": 1.684, + "step": 1909 + }, + { + "epoch": 2.0931506849315067, + "grad_norm": 2.1183440685272217, + "learning_rate": 1.7099918765231522e-06, + "loss": 1.676, + "step": 1910 + }, + { + "epoch": 2.0931506849315067, + "eval_loss": 1.9333924055099487, + "eval_runtime": 37.3166, + "eval_samples_per_second": 26.798, + "eval_steps_per_second": 1.126, + "step": 1910 + }, + { + "epoch": 2.0942465753424657, + "grad_norm": 2.09332013130188, + "learning_rate": 1.7079610073111295e-06, + "loss": 1.6946, + "step": 1911 + }, + { + "epoch": 2.0953424657534248, + "grad_norm": 2.136695384979248, + "learning_rate": 1.7059301380991068e-06, + "loss": 1.6224, + "step": 1912 + }, + { + "epoch": 2.0964383561643833, + "grad_norm": 2.3098766803741455, + "learning_rate": 1.7038992688870836e-06, + "loss": 1.6837, + "step": 1913 + }, + { + "epoch": 2.0975342465753424, + "grad_norm": 2.1490230560302734, + "learning_rate": 1.701868399675061e-06, + "loss": 1.6797, + "step": 1914 + }, + { + "epoch": 2.0986301369863014, + "grad_norm": 2.2379367351531982, + "learning_rate": 1.6998375304630382e-06, + "loss": 1.6976, + "step": 1915 + }, + { + "epoch": 2.0997260273972604, + "grad_norm": 2.1653738021850586, + "learning_rate": 1.6978066612510155e-06, + "loss": 1.7651, + "step": 1916 + }, + { + "epoch": 2.100821917808219, + "grad_norm": 2.195258378982544, + "learning_rate": 1.6957757920389927e-06, + "loss": 1.7226, + "step": 1917 + }, + { + "epoch": 2.101917808219178, + "grad_norm": 2.226472854614258, + "learning_rate": 1.69374492282697e-06, + "loss": 1.6828, + "step": 1918 + }, + { + "epoch": 2.103013698630137, + "grad_norm": 2.2270874977111816, + "learning_rate": 1.6917140536149473e-06, + "loss": 1.6502, + "step": 1919 + }, + { + "epoch": 2.1041095890410957, + "grad_norm": 2.2078800201416016, + "learning_rate": 1.6896831844029246e-06, + "loss": 1.7411, + "step": 1920 + }, + { + "epoch": 2.1041095890410957, + "eval_loss": 1.9392459392547607, + "eval_runtime": 37.484, + "eval_samples_per_second": 26.678, + "eval_steps_per_second": 1.12, + "step": 1920 + }, + { + "epoch": 2.1052054794520547, + "grad_norm": 2.1973047256469727, + "learning_rate": 1.6876523151909018e-06, + "loss": 1.6904, + "step": 1921 + }, + { + "epoch": 2.1063013698630138, + "grad_norm": 2.2516322135925293, + "learning_rate": 1.6856214459788791e-06, + "loss": 1.7042, + "step": 1922 + }, + { + "epoch": 2.107397260273973, + "grad_norm": 2.2418224811553955, + "learning_rate": 1.6835905767668564e-06, + "loss": 1.6569, + "step": 1923 + }, + { + "epoch": 2.1084931506849314, + "grad_norm": 2.303800582885742, + "learning_rate": 1.6815597075548337e-06, + "loss": 1.6577, + "step": 1924 + }, + { + "epoch": 2.1095890410958904, + "grad_norm": 2.098238229751587, + "learning_rate": 1.679528838342811e-06, + "loss": 1.7339, + "step": 1925 + }, + { + "epoch": 2.1106849315068494, + "grad_norm": 2.1578710079193115, + "learning_rate": 1.677497969130788e-06, + "loss": 1.6533, + "step": 1926 + }, + { + "epoch": 2.111780821917808, + "grad_norm": 2.2567641735076904, + "learning_rate": 1.6754670999187653e-06, + "loss": 1.6664, + "step": 1927 + }, + { + "epoch": 2.112876712328767, + "grad_norm": 2.1851046085357666, + "learning_rate": 1.6734362307067426e-06, + "loss": 1.686, + "step": 1928 + }, + { + "epoch": 2.113972602739726, + "grad_norm": 2.2019577026367188, + "learning_rate": 1.6714053614947198e-06, + "loss": 1.6363, + "step": 1929 + }, + { + "epoch": 2.115068493150685, + "grad_norm": 2.3287160396575928, + "learning_rate": 1.669374492282697e-06, + "loss": 1.6491, + "step": 1930 + }, + { + "epoch": 2.115068493150685, + "eval_loss": 1.936156153678894, + "eval_runtime": 37.5733, + "eval_samples_per_second": 26.615, + "eval_steps_per_second": 1.118, + "step": 1930 + }, + { + "epoch": 2.1161643835616437, + "grad_norm": 2.391085386276245, + "learning_rate": 1.6673436230706744e-06, + "loss": 1.7278, + "step": 1931 + }, + { + "epoch": 2.1172602739726027, + "grad_norm": 2.2187516689300537, + "learning_rate": 1.6653127538586517e-06, + "loss": 1.7211, + "step": 1932 + }, + { + "epoch": 2.118356164383562, + "grad_norm": 2.4330122470855713, + "learning_rate": 1.663281884646629e-06, + "loss": 1.6563, + "step": 1933 + }, + { + "epoch": 2.1194520547945204, + "grad_norm": 2.3581886291503906, + "learning_rate": 1.6612510154346062e-06, + "loss": 1.685, + "step": 1934 + }, + { + "epoch": 2.1205479452054794, + "grad_norm": 2.2456486225128174, + "learning_rate": 1.6592201462225835e-06, + "loss": 1.7389, + "step": 1935 + }, + { + "epoch": 2.1216438356164384, + "grad_norm": 2.313720703125, + "learning_rate": 1.6571892770105608e-06, + "loss": 1.7131, + "step": 1936 + }, + { + "epoch": 2.1227397260273975, + "grad_norm": 2.68670916557312, + "learning_rate": 1.655158407798538e-06, + "loss": 1.7427, + "step": 1937 + }, + { + "epoch": 2.123835616438356, + "grad_norm": 2.226874351501465, + "learning_rate": 1.653127538586515e-06, + "loss": 1.6905, + "step": 1938 + }, + { + "epoch": 2.124931506849315, + "grad_norm": 2.5537056922912598, + "learning_rate": 1.6510966693744924e-06, + "loss": 1.7219, + "step": 1939 + }, + { + "epoch": 2.126027397260274, + "grad_norm": 2.402783155441284, + "learning_rate": 1.6490658001624696e-06, + "loss": 1.7356, + "step": 1940 + }, + { + "epoch": 2.126027397260274, + "eval_loss": 1.9355976581573486, + "eval_runtime": 37.619, + "eval_samples_per_second": 26.582, + "eval_steps_per_second": 1.116, + "step": 1940 + }, + { + "epoch": 2.1271232876712327, + "grad_norm": 2.2399065494537354, + "learning_rate": 1.647034930950447e-06, + "loss": 1.7298, + "step": 1941 + }, + { + "epoch": 2.1282191780821917, + "grad_norm": 2.268878936767578, + "learning_rate": 1.6450040617384242e-06, + "loss": 1.6646, + "step": 1942 + }, + { + "epoch": 2.1293150684931508, + "grad_norm": 2.5293900966644287, + "learning_rate": 1.6429731925264015e-06, + "loss": 1.6783, + "step": 1943 + }, + { + "epoch": 2.1304109589041094, + "grad_norm": 2.386493682861328, + "learning_rate": 1.6409423233143788e-06, + "loss": 1.7065, + "step": 1944 + }, + { + "epoch": 2.1315068493150684, + "grad_norm": 2.3141331672668457, + "learning_rate": 1.638911454102356e-06, + "loss": 1.6734, + "step": 1945 + }, + { + "epoch": 2.1326027397260274, + "grad_norm": 2.199573040008545, + "learning_rate": 1.6368805848903333e-06, + "loss": 1.6422, + "step": 1946 + }, + { + "epoch": 2.1336986301369865, + "grad_norm": 2.346104145050049, + "learning_rate": 1.6348497156783106e-06, + "loss": 1.7251, + "step": 1947 + }, + { + "epoch": 2.134794520547945, + "grad_norm": 2.2806875705718994, + "learning_rate": 1.6328188464662879e-06, + "loss": 1.6595, + "step": 1948 + }, + { + "epoch": 2.135890410958904, + "grad_norm": 2.187328815460205, + "learning_rate": 1.6307879772542651e-06, + "loss": 1.6762, + "step": 1949 + }, + { + "epoch": 2.136986301369863, + "grad_norm": 2.2514591217041016, + "learning_rate": 1.628757108042242e-06, + "loss": 1.7047, + "step": 1950 + }, + { + "epoch": 2.136986301369863, + "eval_loss": 1.9337656497955322, + "eval_runtime": 37.5966, + "eval_samples_per_second": 26.598, + "eval_steps_per_second": 1.117, + "step": 1950 + }, + { + "epoch": 2.138082191780822, + "grad_norm": 2.318274736404419, + "learning_rate": 1.6267262388302193e-06, + "loss": 1.7352, + "step": 1951 + }, + { + "epoch": 2.1391780821917807, + "grad_norm": 2.274162769317627, + "learning_rate": 1.6246953696181965e-06, + "loss": 1.6815, + "step": 1952 + }, + { + "epoch": 2.1402739726027398, + "grad_norm": 2.3822526931762695, + "learning_rate": 1.6226645004061738e-06, + "loss": 1.705, + "step": 1953 + }, + { + "epoch": 2.141369863013699, + "grad_norm": 2.217129945755005, + "learning_rate": 1.620633631194151e-06, + "loss": 1.7824, + "step": 1954 + }, + { + "epoch": 2.1424657534246574, + "grad_norm": 2.3034861087799072, + "learning_rate": 1.6186027619821284e-06, + "loss": 1.713, + "step": 1955 + }, + { + "epoch": 2.1435616438356164, + "grad_norm": 2.218284845352173, + "learning_rate": 1.6165718927701056e-06, + "loss": 1.6663, + "step": 1956 + }, + { + "epoch": 2.1446575342465755, + "grad_norm": 2.3420121669769287, + "learning_rate": 1.614541023558083e-06, + "loss": 1.7072, + "step": 1957 + }, + { + "epoch": 2.145753424657534, + "grad_norm": 2.182140350341797, + "learning_rate": 1.6125101543460602e-06, + "loss": 1.7134, + "step": 1958 + }, + { + "epoch": 2.146849315068493, + "grad_norm": 2.2779390811920166, + "learning_rate": 1.6104792851340375e-06, + "loss": 1.6748, + "step": 1959 + }, + { + "epoch": 2.147945205479452, + "grad_norm": 2.247450113296509, + "learning_rate": 1.6084484159220147e-06, + "loss": 1.696, + "step": 1960 + }, + { + "epoch": 2.147945205479452, + "eval_loss": 1.9329941272735596, + "eval_runtime": 37.6058, + "eval_samples_per_second": 26.592, + "eval_steps_per_second": 1.117, + "step": 1960 + }, + { + "epoch": 2.149041095890411, + "grad_norm": 2.3096683025360107, + "learning_rate": 1.606417546709992e-06, + "loss": 1.6729, + "step": 1961 + }, + { + "epoch": 2.1501369863013697, + "grad_norm": 2.14719557762146, + "learning_rate": 1.6043866774979693e-06, + "loss": 1.6615, + "step": 1962 + }, + { + "epoch": 2.1512328767123288, + "grad_norm": 2.2620556354522705, + "learning_rate": 1.6023558082859464e-06, + "loss": 1.6833, + "step": 1963 + }, + { + "epoch": 2.152328767123288, + "grad_norm": 2.373412847518921, + "learning_rate": 1.6003249390739236e-06, + "loss": 1.6486, + "step": 1964 + }, + { + "epoch": 2.1534246575342464, + "grad_norm": 2.223722219467163, + "learning_rate": 1.598294069861901e-06, + "loss": 1.6888, + "step": 1965 + }, + { + "epoch": 2.1545205479452054, + "grad_norm": 2.1869406700134277, + "learning_rate": 1.5962632006498782e-06, + "loss": 1.6619, + "step": 1966 + }, + { + "epoch": 2.1556164383561645, + "grad_norm": 2.241142988204956, + "learning_rate": 1.5942323314378555e-06, + "loss": 1.731, + "step": 1967 + }, + { + "epoch": 2.1567123287671235, + "grad_norm": 2.145240306854248, + "learning_rate": 1.5922014622258327e-06, + "loss": 1.6332, + "step": 1968 + }, + { + "epoch": 2.157808219178082, + "grad_norm": 2.206467390060425, + "learning_rate": 1.59017059301381e-06, + "loss": 1.6529, + "step": 1969 + }, + { + "epoch": 2.158904109589041, + "grad_norm": 2.1281487941741943, + "learning_rate": 1.5881397238017873e-06, + "loss": 1.7022, + "step": 1970 + }, + { + "epoch": 2.158904109589041, + "eval_loss": 1.9336646795272827, + "eval_runtime": 37.6226, + "eval_samples_per_second": 26.58, + "eval_steps_per_second": 1.116, + "step": 1970 + }, + { + "epoch": 2.16, + "grad_norm": 2.8022873401641846, + "learning_rate": 1.5861088545897646e-06, + "loss": 1.7133, + "step": 1971 + }, + { + "epoch": 2.1610958904109587, + "grad_norm": 2.151744842529297, + "learning_rate": 1.5840779853777418e-06, + "loss": 1.6314, + "step": 1972 + }, + { + "epoch": 2.1621917808219178, + "grad_norm": 2.284320831298828, + "learning_rate": 1.5820471161657191e-06, + "loss": 1.6845, + "step": 1973 + }, + { + "epoch": 2.163287671232877, + "grad_norm": 2.205044746398926, + "learning_rate": 1.5800162469536964e-06, + "loss": 1.6653, + "step": 1974 + }, + { + "epoch": 2.1643835616438354, + "grad_norm": 2.794811725616455, + "learning_rate": 1.5779853777416735e-06, + "loss": 1.7294, + "step": 1975 + }, + { + "epoch": 2.1654794520547944, + "grad_norm": 2.311840295791626, + "learning_rate": 1.5759545085296507e-06, + "loss": 1.7145, + "step": 1976 + }, + { + "epoch": 2.1665753424657535, + "grad_norm": 2.2681734561920166, + "learning_rate": 1.573923639317628e-06, + "loss": 1.6922, + "step": 1977 + }, + { + "epoch": 2.1676712328767125, + "grad_norm": 2.1687207221984863, + "learning_rate": 1.5718927701056053e-06, + "loss": 1.704, + "step": 1978 + }, + { + "epoch": 2.168767123287671, + "grad_norm": 2.2467267513275146, + "learning_rate": 1.5698619008935826e-06, + "loss": 1.7009, + "step": 1979 + }, + { + "epoch": 2.16986301369863, + "grad_norm": 2.178468942642212, + "learning_rate": 1.5678310316815598e-06, + "loss": 1.7103, + "step": 1980 + }, + { + "epoch": 2.16986301369863, + "eval_loss": 1.930389404296875, + "eval_runtime": 37.6086, + "eval_samples_per_second": 26.59, + "eval_steps_per_second": 1.117, + "step": 1980 + }, + { + "epoch": 2.170958904109589, + "grad_norm": 2.1918745040893555, + "learning_rate": 1.5658001624695371e-06, + "loss": 1.6473, + "step": 1981 + }, + { + "epoch": 2.172054794520548, + "grad_norm": 2.3625173568725586, + "learning_rate": 1.5637692932575144e-06, + "loss": 1.7239, + "step": 1982 + }, + { + "epoch": 2.1731506849315068, + "grad_norm": 2.2162346839904785, + "learning_rate": 1.5617384240454917e-06, + "loss": 1.6735, + "step": 1983 + }, + { + "epoch": 2.174246575342466, + "grad_norm": 2.3919036388397217, + "learning_rate": 1.559707554833469e-06, + "loss": 1.6845, + "step": 1984 + }, + { + "epoch": 2.175342465753425, + "grad_norm": 2.2709097862243652, + "learning_rate": 1.5576766856214462e-06, + "loss": 1.7364, + "step": 1985 + }, + { + "epoch": 2.1764383561643834, + "grad_norm": 2.1636414527893066, + "learning_rate": 1.5556458164094235e-06, + "loss": 1.6056, + "step": 1986 + }, + { + "epoch": 2.1775342465753424, + "grad_norm": 2.14188289642334, + "learning_rate": 1.5536149471974008e-06, + "loss": 1.6773, + "step": 1987 + }, + { + "epoch": 2.1786301369863015, + "grad_norm": 2.2092809677124023, + "learning_rate": 1.5515840779853778e-06, + "loss": 1.6333, + "step": 1988 + }, + { + "epoch": 2.17972602739726, + "grad_norm": 2.1823160648345947, + "learning_rate": 1.549553208773355e-06, + "loss": 1.6927, + "step": 1989 + }, + { + "epoch": 2.180821917808219, + "grad_norm": 2.3488991260528564, + "learning_rate": 1.5475223395613324e-06, + "loss": 1.6764, + "step": 1990 + }, + { + "epoch": 2.180821917808219, + "eval_loss": 1.934288740158081, + "eval_runtime": 37.6139, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 1990 + }, + { + "epoch": 2.181917808219178, + "grad_norm": 2.336214780807495, + "learning_rate": 1.5454914703493097e-06, + "loss": 1.6683, + "step": 1991 + }, + { + "epoch": 2.183013698630137, + "grad_norm": 2.243014097213745, + "learning_rate": 1.543460601137287e-06, + "loss": 1.7561, + "step": 1992 + }, + { + "epoch": 2.1841095890410958, + "grad_norm": 2.1899123191833496, + "learning_rate": 1.5414297319252642e-06, + "loss": 1.7209, + "step": 1993 + }, + { + "epoch": 2.185205479452055, + "grad_norm": 2.4623351097106934, + "learning_rate": 1.5393988627132415e-06, + "loss": 1.7158, + "step": 1994 + }, + { + "epoch": 2.186301369863014, + "grad_norm": 2.2736823558807373, + "learning_rate": 1.5373679935012188e-06, + "loss": 1.6794, + "step": 1995 + }, + { + "epoch": 2.1873972602739724, + "grad_norm": 2.2148497104644775, + "learning_rate": 1.535337124289196e-06, + "loss": 1.6247, + "step": 1996 + }, + { + "epoch": 2.1884931506849314, + "grad_norm": 2.3276689052581787, + "learning_rate": 1.5333062550771733e-06, + "loss": 1.6835, + "step": 1997 + }, + { + "epoch": 2.1895890410958905, + "grad_norm": 2.3846237659454346, + "learning_rate": 1.5312753858651506e-06, + "loss": 1.771, + "step": 1998 + }, + { + "epoch": 2.1906849315068495, + "grad_norm": 2.2715535163879395, + "learning_rate": 1.5292445166531279e-06, + "loss": 1.6797, + "step": 1999 + }, + { + "epoch": 2.191780821917808, + "grad_norm": 2.1311490535736084, + "learning_rate": 1.5272136474411047e-06, + "loss": 1.6968, + "step": 2000 + }, + { + "epoch": 2.191780821917808, + "eval_loss": 1.9303339719772339, + "eval_runtime": 37.6087, + "eval_samples_per_second": 26.59, + "eval_steps_per_second": 1.117, + "step": 2000 + }, + { + "epoch": 2.192876712328767, + "grad_norm": 2.181492567062378, + "learning_rate": 1.525182778229082e-06, + "loss": 1.738, + "step": 2001 + }, + { + "epoch": 2.193972602739726, + "grad_norm": 2.260737180709839, + "learning_rate": 1.5231519090170593e-06, + "loss": 1.6734, + "step": 2002 + }, + { + "epoch": 2.1950684931506848, + "grad_norm": 2.3129513263702393, + "learning_rate": 1.5211210398050365e-06, + "loss": 1.6546, + "step": 2003 + }, + { + "epoch": 2.196164383561644, + "grad_norm": 2.1104655265808105, + "learning_rate": 1.5190901705930138e-06, + "loss": 1.6633, + "step": 2004 + }, + { + "epoch": 2.197260273972603, + "grad_norm": 2.357048749923706, + "learning_rate": 1.517059301380991e-06, + "loss": 1.7127, + "step": 2005 + }, + { + "epoch": 2.198356164383562, + "grad_norm": 2.320582628250122, + "learning_rate": 1.5150284321689684e-06, + "loss": 1.658, + "step": 2006 + }, + { + "epoch": 2.1994520547945204, + "grad_norm": 2.2205357551574707, + "learning_rate": 1.5129975629569456e-06, + "loss": 1.6902, + "step": 2007 + }, + { + "epoch": 2.2005479452054795, + "grad_norm": 2.227670669555664, + "learning_rate": 1.510966693744923e-06, + "loss": 1.6909, + "step": 2008 + }, + { + "epoch": 2.2016438356164385, + "grad_norm": 2.275946855545044, + "learning_rate": 1.5089358245329002e-06, + "loss": 1.6908, + "step": 2009 + }, + { + "epoch": 2.202739726027397, + "grad_norm": 2.376431941986084, + "learning_rate": 1.5069049553208775e-06, + "loss": 1.6657, + "step": 2010 + }, + { + "epoch": 2.202739726027397, + "eval_loss": 1.9320796728134155, + "eval_runtime": 37.3249, + "eval_samples_per_second": 26.792, + "eval_steps_per_second": 1.125, + "step": 2010 + }, + { + "epoch": 2.203835616438356, + "grad_norm": 2.3358347415924072, + "learning_rate": 1.5048740861088547e-06, + "loss": 1.718, + "step": 2011 + }, + { + "epoch": 2.204931506849315, + "grad_norm": 2.139920234680176, + "learning_rate": 1.502843216896832e-06, + "loss": 1.6973, + "step": 2012 + }, + { + "epoch": 2.206027397260274, + "grad_norm": 2.344691753387451, + "learning_rate": 1.500812347684809e-06, + "loss": 1.6242, + "step": 2013 + }, + { + "epoch": 2.207123287671233, + "grad_norm": 2.217684268951416, + "learning_rate": 1.4987814784727864e-06, + "loss": 1.6833, + "step": 2014 + }, + { + "epoch": 2.208219178082192, + "grad_norm": 2.2504189014434814, + "learning_rate": 1.4967506092607636e-06, + "loss": 1.7009, + "step": 2015 + }, + { + "epoch": 2.209315068493151, + "grad_norm": 2.3367316722869873, + "learning_rate": 1.494719740048741e-06, + "loss": 1.6433, + "step": 2016 + }, + { + "epoch": 2.2104109589041094, + "grad_norm": 2.2345099449157715, + "learning_rate": 1.4926888708367182e-06, + "loss": 1.713, + "step": 2017 + }, + { + "epoch": 2.2115068493150685, + "grad_norm": 2.3112707138061523, + "learning_rate": 1.4906580016246955e-06, + "loss": 1.7164, + "step": 2018 + }, + { + "epoch": 2.2126027397260275, + "grad_norm": 2.2484328746795654, + "learning_rate": 1.4886271324126727e-06, + "loss": 1.735, + "step": 2019 + }, + { + "epoch": 2.213698630136986, + "grad_norm": 2.184624195098877, + "learning_rate": 1.48659626320065e-06, + "loss": 1.7567, + "step": 2020 + }, + { + "epoch": 2.213698630136986, + "eval_loss": 1.9320781230926514, + "eval_runtime": 37.505, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 2020 + }, + { + "epoch": 2.214794520547945, + "grad_norm": 2.3959224224090576, + "learning_rate": 1.4845653939886273e-06, + "loss": 1.6049, + "step": 2021 + }, + { + "epoch": 2.215890410958904, + "grad_norm": 2.1809372901916504, + "learning_rate": 1.4825345247766046e-06, + "loss": 1.6103, + "step": 2022 + }, + { + "epoch": 2.216986301369863, + "grad_norm": 2.1426923274993896, + "learning_rate": 1.4805036555645818e-06, + "loss": 1.6102, + "step": 2023 + }, + { + "epoch": 2.2180821917808218, + "grad_norm": 2.3202199935913086, + "learning_rate": 1.4784727863525591e-06, + "loss": 1.6823, + "step": 2024 + }, + { + "epoch": 2.219178082191781, + "grad_norm": 2.2229015827178955, + "learning_rate": 1.4764419171405362e-06, + "loss": 1.6457, + "step": 2025 + }, + { + "epoch": 2.22027397260274, + "grad_norm": 2.3974742889404297, + "learning_rate": 1.4744110479285135e-06, + "loss": 1.721, + "step": 2026 + }, + { + "epoch": 2.2213698630136984, + "grad_norm": 2.3097574710845947, + "learning_rate": 1.4723801787164907e-06, + "loss": 1.7252, + "step": 2027 + }, + { + "epoch": 2.2224657534246575, + "grad_norm": 2.1884419918060303, + "learning_rate": 1.470349309504468e-06, + "loss": 1.7245, + "step": 2028 + }, + { + "epoch": 2.2235616438356165, + "grad_norm": 2.486947774887085, + "learning_rate": 1.4683184402924453e-06, + "loss": 1.6892, + "step": 2029 + }, + { + "epoch": 2.2246575342465755, + "grad_norm": 2.349576234817505, + "learning_rate": 1.4662875710804226e-06, + "loss": 1.6981, + "step": 2030 + }, + { + "epoch": 2.2246575342465755, + "eval_loss": 1.925842523574829, + "eval_runtime": 37.6142, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 2030 + }, + { + "epoch": 2.225753424657534, + "grad_norm": 2.218215227127075, + "learning_rate": 1.4642567018683998e-06, + "loss": 1.7772, + "step": 2031 + }, + { + "epoch": 2.226849315068493, + "grad_norm": 2.3655383586883545, + "learning_rate": 1.4622258326563771e-06, + "loss": 1.6822, + "step": 2032 + }, + { + "epoch": 2.227945205479452, + "grad_norm": 2.249873161315918, + "learning_rate": 1.4601949634443544e-06, + "loss": 1.6556, + "step": 2033 + }, + { + "epoch": 2.2290410958904108, + "grad_norm": 2.200159788131714, + "learning_rate": 1.4581640942323317e-06, + "loss": 1.7017, + "step": 2034 + }, + { + "epoch": 2.23013698630137, + "grad_norm": 2.2711470127105713, + "learning_rate": 1.456133225020309e-06, + "loss": 1.7838, + "step": 2035 + }, + { + "epoch": 2.231232876712329, + "grad_norm": 2.16304874420166, + "learning_rate": 1.4541023558082862e-06, + "loss": 1.6971, + "step": 2036 + }, + { + "epoch": 2.232328767123288, + "grad_norm": 2.18495774269104, + "learning_rate": 1.4520714865962635e-06, + "loss": 1.6839, + "step": 2037 + }, + { + "epoch": 2.2334246575342465, + "grad_norm": 2.2053871154785156, + "learning_rate": 1.4500406173842406e-06, + "loss": 1.7283, + "step": 2038 + }, + { + "epoch": 2.2345205479452055, + "grad_norm": 2.2265138626098633, + "learning_rate": 1.4480097481722178e-06, + "loss": 1.6967, + "step": 2039 + }, + { + "epoch": 2.2356164383561645, + "grad_norm": 2.1966476440429688, + "learning_rate": 1.4459788789601951e-06, + "loss": 1.676, + "step": 2040 + }, + { + "epoch": 2.2356164383561645, + "eval_loss": 1.9303417205810547, + "eval_runtime": 37.6334, + "eval_samples_per_second": 26.572, + "eval_steps_per_second": 1.116, + "step": 2040 + }, + { + "epoch": 2.236712328767123, + "grad_norm": 2.251950979232788, + "learning_rate": 1.4439480097481724e-06, + "loss": 1.6519, + "step": 2041 + }, + { + "epoch": 2.237808219178082, + "grad_norm": 2.4104197025299072, + "learning_rate": 1.4419171405361497e-06, + "loss": 1.6417, + "step": 2042 + }, + { + "epoch": 2.238904109589041, + "grad_norm": 2.324711561203003, + "learning_rate": 1.439886271324127e-06, + "loss": 1.6972, + "step": 2043 + }, + { + "epoch": 2.24, + "grad_norm": 2.2258901596069336, + "learning_rate": 1.4378554021121042e-06, + "loss": 1.7012, + "step": 2044 + }, + { + "epoch": 2.241095890410959, + "grad_norm": 2.256211757659912, + "learning_rate": 1.4358245329000815e-06, + "loss": 1.6653, + "step": 2045 + }, + { + "epoch": 2.242191780821918, + "grad_norm": 2.795642614364624, + "learning_rate": 1.4337936636880588e-06, + "loss": 1.7368, + "step": 2046 + }, + { + "epoch": 2.243287671232877, + "grad_norm": 2.337639570236206, + "learning_rate": 1.431762794476036e-06, + "loss": 1.6995, + "step": 2047 + }, + { + "epoch": 2.2443835616438355, + "grad_norm": 2.224888324737549, + "learning_rate": 1.4297319252640133e-06, + "loss": 1.646, + "step": 2048 + }, + { + "epoch": 2.2454794520547945, + "grad_norm": 2.1721208095550537, + "learning_rate": 1.4277010560519906e-06, + "loss": 1.6916, + "step": 2049 + }, + { + "epoch": 2.2465753424657535, + "grad_norm": 2.5284039974212646, + "learning_rate": 1.4256701868399674e-06, + "loss": 1.6601, + "step": 2050 + }, + { + "epoch": 2.2465753424657535, + "eval_loss": 1.930783748626709, + "eval_runtime": 37.6129, + "eval_samples_per_second": 26.587, + "eval_steps_per_second": 1.117, + "step": 2050 + }, + { + "epoch": 2.247671232876712, + "grad_norm": 2.485682487487793, + "learning_rate": 1.4236393176279447e-06, + "loss": 1.7586, + "step": 2051 + }, + { + "epoch": 2.248767123287671, + "grad_norm": 2.275346517562866, + "learning_rate": 1.421608448415922e-06, + "loss": 1.6837, + "step": 2052 + }, + { + "epoch": 2.24986301369863, + "grad_norm": 2.2854814529418945, + "learning_rate": 1.4195775792038993e-06, + "loss": 1.7784, + "step": 2053 + }, + { + "epoch": 2.250958904109589, + "grad_norm": 2.6571497917175293, + "learning_rate": 1.4175467099918765e-06, + "loss": 1.6594, + "step": 2054 + }, + { + "epoch": 2.252054794520548, + "grad_norm": 2.1349689960479736, + "learning_rate": 1.4155158407798538e-06, + "loss": 1.6959, + "step": 2055 + }, + { + "epoch": 2.253150684931507, + "grad_norm": 2.302368402481079, + "learning_rate": 1.413484971567831e-06, + "loss": 1.7109, + "step": 2056 + }, + { + "epoch": 2.254246575342466, + "grad_norm": 2.410618782043457, + "learning_rate": 1.4114541023558084e-06, + "loss": 1.6357, + "step": 2057 + }, + { + "epoch": 2.255342465753425, + "grad_norm": 2.5003485679626465, + "learning_rate": 1.4094232331437856e-06, + "loss": 1.6666, + "step": 2058 + }, + { + "epoch": 2.2564383561643835, + "grad_norm": 2.1961967945098877, + "learning_rate": 1.407392363931763e-06, + "loss": 1.7251, + "step": 2059 + }, + { + "epoch": 2.2575342465753425, + "grad_norm": 2.287523031234741, + "learning_rate": 1.4053614947197402e-06, + "loss": 1.7918, + "step": 2060 + }, + { + "epoch": 2.2575342465753425, + "eval_loss": 1.9307024478912354, + "eval_runtime": 37.62, + "eval_samples_per_second": 26.582, + "eval_steps_per_second": 1.116, + "step": 2060 + }, + { + "epoch": 2.2586301369863016, + "grad_norm": 2.5906260013580322, + "learning_rate": 1.4033306255077175e-06, + "loss": 1.7049, + "step": 2061 + }, + { + "epoch": 2.25972602739726, + "grad_norm": 2.1343653202056885, + "learning_rate": 1.4012997562956945e-06, + "loss": 1.6062, + "step": 2062 + }, + { + "epoch": 2.260821917808219, + "grad_norm": 2.210479974746704, + "learning_rate": 1.3992688870836718e-06, + "loss": 1.7367, + "step": 2063 + }, + { + "epoch": 2.261917808219178, + "grad_norm": 2.3736038208007812, + "learning_rate": 1.397238017871649e-06, + "loss": 1.7067, + "step": 2064 + }, + { + "epoch": 2.263013698630137, + "grad_norm": 2.4008090496063232, + "learning_rate": 1.3952071486596264e-06, + "loss": 1.6839, + "step": 2065 + }, + { + "epoch": 2.264109589041096, + "grad_norm": 2.2778053283691406, + "learning_rate": 1.3931762794476036e-06, + "loss": 1.6468, + "step": 2066 + }, + { + "epoch": 2.265205479452055, + "grad_norm": 2.373081922531128, + "learning_rate": 1.391145410235581e-06, + "loss": 1.7308, + "step": 2067 + }, + { + "epoch": 2.266301369863014, + "grad_norm": 2.2529540061950684, + "learning_rate": 1.3891145410235582e-06, + "loss": 1.68, + "step": 2068 + }, + { + "epoch": 2.2673972602739725, + "grad_norm": 2.4584438800811768, + "learning_rate": 1.3870836718115355e-06, + "loss": 1.6124, + "step": 2069 + }, + { + "epoch": 2.2684931506849315, + "grad_norm": 2.263357400894165, + "learning_rate": 1.3850528025995127e-06, + "loss": 1.6737, + "step": 2070 + }, + { + "epoch": 2.2684931506849315, + "eval_loss": 1.927953839302063, + "eval_runtime": 37.5973, + "eval_samples_per_second": 26.598, + "eval_steps_per_second": 1.117, + "step": 2070 + }, + { + "epoch": 2.2695890410958905, + "grad_norm": 2.078960657119751, + "learning_rate": 1.38302193338749e-06, + "loss": 1.6737, + "step": 2071 + }, + { + "epoch": 2.270684931506849, + "grad_norm": 2.4686176776885986, + "learning_rate": 1.3809910641754673e-06, + "loss": 1.6067, + "step": 2072 + }, + { + "epoch": 2.271780821917808, + "grad_norm": 2.7145605087280273, + "learning_rate": 1.3789601949634446e-06, + "loss": 1.7064, + "step": 2073 + }, + { + "epoch": 2.272876712328767, + "grad_norm": 2.1989216804504395, + "learning_rate": 1.3769293257514218e-06, + "loss": 1.6177, + "step": 2074 + }, + { + "epoch": 2.2739726027397262, + "grad_norm": 2.446955442428589, + "learning_rate": 1.374898456539399e-06, + "loss": 1.7187, + "step": 2075 + }, + { + "epoch": 2.275068493150685, + "grad_norm": 2.2136058807373047, + "learning_rate": 1.3728675873273762e-06, + "loss": 1.6481, + "step": 2076 + }, + { + "epoch": 2.276164383561644, + "grad_norm": 2.247084617614746, + "learning_rate": 1.3708367181153535e-06, + "loss": 1.7457, + "step": 2077 + }, + { + "epoch": 2.277260273972603, + "grad_norm": 2.2474679946899414, + "learning_rate": 1.3688058489033307e-06, + "loss": 1.7301, + "step": 2078 + }, + { + "epoch": 2.2783561643835615, + "grad_norm": 2.1973447799682617, + "learning_rate": 1.366774979691308e-06, + "loss": 1.6929, + "step": 2079 + }, + { + "epoch": 2.2794520547945205, + "grad_norm": 2.1621947288513184, + "learning_rate": 1.3647441104792853e-06, + "loss": 1.724, + "step": 2080 + }, + { + "epoch": 2.2794520547945205, + "eval_loss": 1.925129771232605, + "eval_runtime": 37.5704, + "eval_samples_per_second": 26.617, + "eval_steps_per_second": 1.118, + "step": 2080 + }, + { + "epoch": 2.2805479452054795, + "grad_norm": 2.237715482711792, + "learning_rate": 1.3627132412672626e-06, + "loss": 1.6907, + "step": 2081 + }, + { + "epoch": 2.281643835616438, + "grad_norm": 2.1457691192626953, + "learning_rate": 1.3606823720552398e-06, + "loss": 1.6391, + "step": 2082 + }, + { + "epoch": 2.282739726027397, + "grad_norm": 2.1025590896606445, + "learning_rate": 1.3586515028432171e-06, + "loss": 1.691, + "step": 2083 + }, + { + "epoch": 2.283835616438356, + "grad_norm": 2.1200029850006104, + "learning_rate": 1.3566206336311944e-06, + "loss": 1.6154, + "step": 2084 + }, + { + "epoch": 2.2849315068493152, + "grad_norm": 2.1806960105895996, + "learning_rate": 1.3545897644191717e-06, + "loss": 1.6908, + "step": 2085 + }, + { + "epoch": 2.286027397260274, + "grad_norm": 2.2580533027648926, + "learning_rate": 1.352558895207149e-06, + "loss": 1.6613, + "step": 2086 + }, + { + "epoch": 2.287123287671233, + "grad_norm": 2.1397502422332764, + "learning_rate": 1.3505280259951258e-06, + "loss": 1.7114, + "step": 2087 + }, + { + "epoch": 2.288219178082192, + "grad_norm": 2.124181032180786, + "learning_rate": 1.348497156783103e-06, + "loss": 1.687, + "step": 2088 + }, + { + "epoch": 2.289315068493151, + "grad_norm": 2.1617271900177, + "learning_rate": 1.3464662875710804e-06, + "loss": 1.7197, + "step": 2089 + }, + { + "epoch": 2.2904109589041095, + "grad_norm": 2.1740005016326904, + "learning_rate": 1.3444354183590576e-06, + "loss": 1.623, + "step": 2090 + }, + { + "epoch": 2.2904109589041095, + "eval_loss": 1.9223949909210205, + "eval_runtime": 37.6059, + "eval_samples_per_second": 26.592, + "eval_steps_per_second": 1.117, + "step": 2090 + }, + { + "epoch": 2.2915068493150685, + "grad_norm": 2.0894129276275635, + "learning_rate": 1.342404549147035e-06, + "loss": 1.71, + "step": 2091 + }, + { + "epoch": 2.2926027397260276, + "grad_norm": 2.263033151626587, + "learning_rate": 1.3403736799350122e-06, + "loss": 1.6532, + "step": 2092 + }, + { + "epoch": 2.293698630136986, + "grad_norm": 2.3359904289245605, + "learning_rate": 1.3383428107229895e-06, + "loss": 1.7208, + "step": 2093 + }, + { + "epoch": 2.294794520547945, + "grad_norm": 2.20955228805542, + "learning_rate": 1.3363119415109667e-06, + "loss": 1.7366, + "step": 2094 + }, + { + "epoch": 2.2958904109589042, + "grad_norm": 2.1048381328582764, + "learning_rate": 1.334281072298944e-06, + "loss": 1.6946, + "step": 2095 + }, + { + "epoch": 2.296986301369863, + "grad_norm": 2.3098573684692383, + "learning_rate": 1.3322502030869213e-06, + "loss": 1.706, + "step": 2096 + }, + { + "epoch": 2.298082191780822, + "grad_norm": 2.190420389175415, + "learning_rate": 1.3302193338748986e-06, + "loss": 1.6673, + "step": 2097 + }, + { + "epoch": 2.299178082191781, + "grad_norm": 2.165050745010376, + "learning_rate": 1.3281884646628758e-06, + "loss": 1.7221, + "step": 2098 + }, + { + "epoch": 2.30027397260274, + "grad_norm": 2.107862710952759, + "learning_rate": 1.3261575954508531e-06, + "loss": 1.7208, + "step": 2099 + }, + { + "epoch": 2.3013698630136985, + "grad_norm": 2.1177818775177, + "learning_rate": 1.3241267262388302e-06, + "loss": 1.7321, + "step": 2100 + }, + { + "epoch": 2.3013698630136985, + "eval_loss": 1.9249632358551025, + "eval_runtime": 37.5613, + "eval_samples_per_second": 26.623, + "eval_steps_per_second": 1.118, + "step": 2100 + }, + { + "epoch": 2.3024657534246575, + "grad_norm": 2.1729538440704346, + "learning_rate": 1.3220958570268074e-06, + "loss": 1.6805, + "step": 2101 + }, + { + "epoch": 2.3035616438356166, + "grad_norm": 2.131054162979126, + "learning_rate": 1.3200649878147847e-06, + "loss": 1.6379, + "step": 2102 + }, + { + "epoch": 2.304657534246575, + "grad_norm": 2.251866340637207, + "learning_rate": 1.318034118602762e-06, + "loss": 1.6754, + "step": 2103 + }, + { + "epoch": 2.305753424657534, + "grad_norm": 2.2547497749328613, + "learning_rate": 1.3160032493907393e-06, + "loss": 1.7158, + "step": 2104 + }, + { + "epoch": 2.3068493150684932, + "grad_norm": 2.296887159347534, + "learning_rate": 1.3139723801787166e-06, + "loss": 1.6898, + "step": 2105 + }, + { + "epoch": 2.3079452054794523, + "grad_norm": 2.315321207046509, + "learning_rate": 1.3119415109666938e-06, + "loss": 1.77, + "step": 2106 + }, + { + "epoch": 2.309041095890411, + "grad_norm": 2.1939897537231445, + "learning_rate": 1.309910641754671e-06, + "loss": 1.6886, + "step": 2107 + }, + { + "epoch": 2.31013698630137, + "grad_norm": 2.274505615234375, + "learning_rate": 1.3078797725426484e-06, + "loss": 1.6469, + "step": 2108 + }, + { + "epoch": 2.311232876712329, + "grad_norm": 2.1344008445739746, + "learning_rate": 1.3058489033306257e-06, + "loss": 1.653, + "step": 2109 + }, + { + "epoch": 2.3123287671232875, + "grad_norm": 2.2148144245147705, + "learning_rate": 1.303818034118603e-06, + "loss": 1.6581, + "step": 2110 + }, + { + "epoch": 2.3123287671232875, + "eval_loss": 1.9226648807525635, + "eval_runtime": 37.2603, + "eval_samples_per_second": 26.838, + "eval_steps_per_second": 1.127, + "step": 2110 + }, + { + "epoch": 2.3134246575342465, + "grad_norm": 2.2608847618103027, + "learning_rate": 1.3017871649065802e-06, + "loss": 1.6897, + "step": 2111 + }, + { + "epoch": 2.3145205479452056, + "grad_norm": 2.225954532623291, + "learning_rate": 1.2997562956945573e-06, + "loss": 1.667, + "step": 2112 + }, + { + "epoch": 2.315616438356164, + "grad_norm": 2.2951343059539795, + "learning_rate": 1.2977254264825345e-06, + "loss": 1.6464, + "step": 2113 + }, + { + "epoch": 2.316712328767123, + "grad_norm": 2.5108482837677, + "learning_rate": 1.2956945572705118e-06, + "loss": 1.68, + "step": 2114 + }, + { + "epoch": 2.317808219178082, + "grad_norm": 2.220449209213257, + "learning_rate": 1.293663688058489e-06, + "loss": 1.6993, + "step": 2115 + }, + { + "epoch": 2.3189041095890413, + "grad_norm": 2.1419951915740967, + "learning_rate": 1.2916328188464664e-06, + "loss": 1.6583, + "step": 2116 + }, + { + "epoch": 2.32, + "grad_norm": 2.1963181495666504, + "learning_rate": 1.2896019496344436e-06, + "loss": 1.5972, + "step": 2117 + }, + { + "epoch": 2.321095890410959, + "grad_norm": 2.180011034011841, + "learning_rate": 1.287571080422421e-06, + "loss": 1.6582, + "step": 2118 + }, + { + "epoch": 2.322191780821918, + "grad_norm": 2.2337605953216553, + "learning_rate": 1.2855402112103982e-06, + "loss": 1.7245, + "step": 2119 + }, + { + "epoch": 2.323287671232877, + "grad_norm": 2.3071448802948, + "learning_rate": 1.2835093419983755e-06, + "loss": 1.6345, + "step": 2120 + }, + { + "epoch": 2.323287671232877, + "eval_loss": 1.923328161239624, + "eval_runtime": 37.4597, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 2120 + }, + { + "epoch": 2.3243835616438355, + "grad_norm": 2.3645429611206055, + "learning_rate": 1.2814784727863528e-06, + "loss": 1.6582, + "step": 2121 + }, + { + "epoch": 2.3254794520547946, + "grad_norm": 2.1578540802001953, + "learning_rate": 1.27944760357433e-06, + "loss": 1.6548, + "step": 2122 + }, + { + "epoch": 2.3265753424657536, + "grad_norm": 2.3412892818450928, + "learning_rate": 1.2774167343623073e-06, + "loss": 1.7246, + "step": 2123 + }, + { + "epoch": 2.327671232876712, + "grad_norm": 2.553150177001953, + "learning_rate": 1.2753858651502846e-06, + "loss": 1.7247, + "step": 2124 + }, + { + "epoch": 2.328767123287671, + "grad_norm": 2.3273262977600098, + "learning_rate": 1.2733549959382616e-06, + "loss": 1.7032, + "step": 2125 + }, + { + "epoch": 2.3298630136986302, + "grad_norm": 2.1439902782440186, + "learning_rate": 1.271324126726239e-06, + "loss": 1.6795, + "step": 2126 + }, + { + "epoch": 2.330958904109589, + "grad_norm": 2.156188726425171, + "learning_rate": 1.2692932575142162e-06, + "loss": 1.6755, + "step": 2127 + }, + { + "epoch": 2.332054794520548, + "grad_norm": 2.155975103378296, + "learning_rate": 1.2672623883021935e-06, + "loss": 1.6316, + "step": 2128 + }, + { + "epoch": 2.333150684931507, + "grad_norm": 2.2429726123809814, + "learning_rate": 1.2652315190901707e-06, + "loss": 1.6238, + "step": 2129 + }, + { + "epoch": 2.334246575342466, + "grad_norm": 2.2985239028930664, + "learning_rate": 1.263200649878148e-06, + "loss": 1.7557, + "step": 2130 + }, + { + "epoch": 2.334246575342466, + "eval_loss": 1.9246450662612915, + "eval_runtime": 37.5494, + "eval_samples_per_second": 26.632, + "eval_steps_per_second": 1.119, + "step": 2130 + }, + { + "epoch": 2.3353424657534245, + "grad_norm": 2.1231393814086914, + "learning_rate": 1.2611697806661253e-06, + "loss": 1.6735, + "step": 2131 + }, + { + "epoch": 2.3364383561643836, + "grad_norm": 2.3760006427764893, + "learning_rate": 1.2591389114541026e-06, + "loss": 1.6957, + "step": 2132 + }, + { + "epoch": 2.3375342465753426, + "grad_norm": 2.3295648097991943, + "learning_rate": 1.2571080422420798e-06, + "loss": 1.6824, + "step": 2133 + }, + { + "epoch": 2.338630136986301, + "grad_norm": 2.3547332286834717, + "learning_rate": 1.2550771730300571e-06, + "loss": 1.6682, + "step": 2134 + }, + { + "epoch": 2.33972602739726, + "grad_norm": 2.3630971908569336, + "learning_rate": 1.2530463038180344e-06, + "loss": 1.7386, + "step": 2135 + }, + { + "epoch": 2.3408219178082192, + "grad_norm": 2.3361897468566895, + "learning_rate": 1.2510154346060117e-06, + "loss": 1.6768, + "step": 2136 + }, + { + "epoch": 2.3419178082191783, + "grad_norm": 2.793766498565674, + "learning_rate": 1.2489845653939887e-06, + "loss": 1.6898, + "step": 2137 + }, + { + "epoch": 2.343013698630137, + "grad_norm": 2.2704460620880127, + "learning_rate": 1.246953696181966e-06, + "loss": 1.7238, + "step": 2138 + }, + { + "epoch": 2.344109589041096, + "grad_norm": 2.4243364334106445, + "learning_rate": 1.2449228269699433e-06, + "loss": 1.7602, + "step": 2139 + }, + { + "epoch": 2.345205479452055, + "grad_norm": 2.2841155529022217, + "learning_rate": 1.2428919577579204e-06, + "loss": 1.6903, + "step": 2140 + }, + { + "epoch": 2.345205479452055, + "eval_loss": 1.9237843751907349, + "eval_runtime": 37.6298, + "eval_samples_per_second": 26.575, + "eval_steps_per_second": 1.116, + "step": 2140 + }, + { + "epoch": 2.3463013698630135, + "grad_norm": 2.1986255645751953, + "learning_rate": 1.2408610885458976e-06, + "loss": 1.6437, + "step": 2141 + }, + { + "epoch": 2.3473972602739726, + "grad_norm": 2.304729700088501, + "learning_rate": 1.238830219333875e-06, + "loss": 1.6732, + "step": 2142 + }, + { + "epoch": 2.3484931506849316, + "grad_norm": 2.193451166152954, + "learning_rate": 1.2367993501218522e-06, + "loss": 1.6601, + "step": 2143 + }, + { + "epoch": 2.34958904109589, + "grad_norm": 2.2576236724853516, + "learning_rate": 1.2347684809098295e-06, + "loss": 1.7311, + "step": 2144 + }, + { + "epoch": 2.350684931506849, + "grad_norm": 2.256216287612915, + "learning_rate": 1.2327376116978067e-06, + "loss": 1.7014, + "step": 2145 + }, + { + "epoch": 2.3517808219178082, + "grad_norm": 2.1790428161621094, + "learning_rate": 1.230706742485784e-06, + "loss": 1.6163, + "step": 2146 + }, + { + "epoch": 2.3528767123287673, + "grad_norm": 2.1548452377319336, + "learning_rate": 1.2286758732737613e-06, + "loss": 1.6294, + "step": 2147 + }, + { + "epoch": 2.353972602739726, + "grad_norm": 2.205404281616211, + "learning_rate": 1.2266450040617386e-06, + "loss": 1.7709, + "step": 2148 + }, + { + "epoch": 2.355068493150685, + "grad_norm": 2.1739368438720703, + "learning_rate": 1.2246141348497158e-06, + "loss": 1.7617, + "step": 2149 + }, + { + "epoch": 2.356164383561644, + "grad_norm": 2.0770647525787354, + "learning_rate": 1.2225832656376931e-06, + "loss": 1.7003, + "step": 2150 + }, + { + "epoch": 2.356164383561644, + "eval_loss": 1.9211101531982422, + "eval_runtime": 37.6076, + "eval_samples_per_second": 26.59, + "eval_steps_per_second": 1.117, + "step": 2150 + }, + { + "epoch": 2.357260273972603, + "grad_norm": 2.1482436656951904, + "learning_rate": 1.2205523964256704e-06, + "loss": 1.6897, + "step": 2151 + }, + { + "epoch": 2.3583561643835615, + "grad_norm": 2.354529857635498, + "learning_rate": 1.2185215272136477e-06, + "loss": 1.6658, + "step": 2152 + }, + { + "epoch": 2.3594520547945206, + "grad_norm": 2.168431520462036, + "learning_rate": 1.2164906580016247e-06, + "loss": 1.6869, + "step": 2153 + }, + { + "epoch": 2.3605479452054796, + "grad_norm": 2.1513350009918213, + "learning_rate": 1.214459788789602e-06, + "loss": 1.6882, + "step": 2154 + }, + { + "epoch": 2.361643835616438, + "grad_norm": 2.25950026512146, + "learning_rate": 1.2124289195775793e-06, + "loss": 1.6004, + "step": 2155 + }, + { + "epoch": 2.3627397260273972, + "grad_norm": 2.198716640472412, + "learning_rate": 1.2103980503655566e-06, + "loss": 1.64, + "step": 2156 + }, + { + "epoch": 2.3638356164383563, + "grad_norm": 2.166332960128784, + "learning_rate": 1.2083671811535338e-06, + "loss": 1.6732, + "step": 2157 + }, + { + "epoch": 2.364931506849315, + "grad_norm": 2.191387176513672, + "learning_rate": 1.2063363119415111e-06, + "loss": 1.6621, + "step": 2158 + }, + { + "epoch": 2.366027397260274, + "grad_norm": 2.15108060836792, + "learning_rate": 1.2043054427294882e-06, + "loss": 1.6496, + "step": 2159 + }, + { + "epoch": 2.367123287671233, + "grad_norm": 2.170804977416992, + "learning_rate": 1.2022745735174654e-06, + "loss": 1.6676, + "step": 2160 + }, + { + "epoch": 2.367123287671233, + "eval_loss": 1.9205737113952637, + "eval_runtime": 37.6409, + "eval_samples_per_second": 26.567, + "eval_steps_per_second": 1.116, + "step": 2160 + }, + { + "epoch": 2.368219178082192, + "grad_norm": 2.1614902019500732, + "learning_rate": 1.2002437043054427e-06, + "loss": 1.7079, + "step": 2161 + }, + { + "epoch": 2.3693150684931505, + "grad_norm": 2.0777454376220703, + "learning_rate": 1.19821283509342e-06, + "loss": 1.7309, + "step": 2162 + }, + { + "epoch": 2.3704109589041096, + "grad_norm": 2.065349817276001, + "learning_rate": 1.1961819658813973e-06, + "loss": 1.7245, + "step": 2163 + }, + { + "epoch": 2.3715068493150686, + "grad_norm": 2.285088062286377, + "learning_rate": 1.1941510966693746e-06, + "loss": 1.6947, + "step": 2164 + }, + { + "epoch": 2.3726027397260276, + "grad_norm": 2.2614521980285645, + "learning_rate": 1.1921202274573518e-06, + "loss": 1.6684, + "step": 2165 + }, + { + "epoch": 2.3736986301369862, + "grad_norm": 2.269033193588257, + "learning_rate": 1.190089358245329e-06, + "loss": 1.6648, + "step": 2166 + }, + { + "epoch": 2.3747945205479453, + "grad_norm": 2.1208066940307617, + "learning_rate": 1.1880584890333064e-06, + "loss": 1.7408, + "step": 2167 + }, + { + "epoch": 2.3758904109589043, + "grad_norm": 2.1376914978027344, + "learning_rate": 1.1860276198212837e-06, + "loss": 1.6658, + "step": 2168 + }, + { + "epoch": 2.376986301369863, + "grad_norm": 2.190903663635254, + "learning_rate": 1.183996750609261e-06, + "loss": 1.7041, + "step": 2169 + }, + { + "epoch": 2.378082191780822, + "grad_norm": 2.1453306674957275, + "learning_rate": 1.1819658813972382e-06, + "loss": 1.704, + "step": 2170 + }, + { + "epoch": 2.378082191780822, + "eval_loss": 1.921817660331726, + "eval_runtime": 37.6184, + "eval_samples_per_second": 26.583, + "eval_steps_per_second": 1.116, + "step": 2170 + }, + { + "epoch": 2.379178082191781, + "grad_norm": 2.2107887268066406, + "learning_rate": 1.1799350121852153e-06, + "loss": 1.6804, + "step": 2171 + }, + { + "epoch": 2.3802739726027395, + "grad_norm": 2.130927085876465, + "learning_rate": 1.1779041429731925e-06, + "loss": 1.7214, + "step": 2172 + }, + { + "epoch": 2.3813698630136986, + "grad_norm": 2.099144458770752, + "learning_rate": 1.1758732737611698e-06, + "loss": 1.7286, + "step": 2173 + }, + { + "epoch": 2.3824657534246576, + "grad_norm": 2.2131845951080322, + "learning_rate": 1.173842404549147e-06, + "loss": 1.6281, + "step": 2174 + }, + { + "epoch": 2.383561643835616, + "grad_norm": 2.2873342037200928, + "learning_rate": 1.1718115353371244e-06, + "loss": 1.7481, + "step": 2175 + }, + { + "epoch": 2.3846575342465752, + "grad_norm": 2.1060359477996826, + "learning_rate": 1.1697806661251016e-06, + "loss": 1.6651, + "step": 2176 + }, + { + "epoch": 2.3857534246575343, + "grad_norm": 2.1740171909332275, + "learning_rate": 1.167749796913079e-06, + "loss": 1.7324, + "step": 2177 + }, + { + "epoch": 2.3868493150684933, + "grad_norm": 2.212214708328247, + "learning_rate": 1.1657189277010562e-06, + "loss": 1.7403, + "step": 2178 + }, + { + "epoch": 2.387945205479452, + "grad_norm": 2.2824950218200684, + "learning_rate": 1.1636880584890335e-06, + "loss": 1.6755, + "step": 2179 + }, + { + "epoch": 2.389041095890411, + "grad_norm": 2.1406712532043457, + "learning_rate": 1.1616571892770108e-06, + "loss": 1.6637, + "step": 2180 + }, + { + "epoch": 2.389041095890411, + "eval_loss": 1.9189414978027344, + "eval_runtime": 37.6062, + "eval_samples_per_second": 26.591, + "eval_steps_per_second": 1.117, + "step": 2180 + }, + { + "epoch": 2.39013698630137, + "grad_norm": 2.2568323612213135, + "learning_rate": 1.159626320064988e-06, + "loss": 1.7238, + "step": 2181 + }, + { + "epoch": 2.391232876712329, + "grad_norm": 2.040121555328369, + "learning_rate": 1.1575954508529653e-06, + "loss": 1.6896, + "step": 2182 + }, + { + "epoch": 2.3923287671232876, + "grad_norm": 2.0916733741760254, + "learning_rate": 1.1555645816409426e-06, + "loss": 1.7133, + "step": 2183 + }, + { + "epoch": 2.3934246575342466, + "grad_norm": 2.1594905853271484, + "learning_rate": 1.1535337124289196e-06, + "loss": 1.6971, + "step": 2184 + }, + { + "epoch": 2.3945205479452056, + "grad_norm": 2.1951136589050293, + "learning_rate": 1.151502843216897e-06, + "loss": 1.6473, + "step": 2185 + }, + { + "epoch": 2.3956164383561642, + "grad_norm": 2.105931043624878, + "learning_rate": 1.1494719740048742e-06, + "loss": 1.6053, + "step": 2186 + }, + { + "epoch": 2.3967123287671233, + "grad_norm": 2.2354254722595215, + "learning_rate": 1.1474411047928515e-06, + "loss": 1.6831, + "step": 2187 + }, + { + "epoch": 2.3978082191780823, + "grad_norm": 2.2005128860473633, + "learning_rate": 1.1454102355808287e-06, + "loss": 1.6768, + "step": 2188 + }, + { + "epoch": 2.398904109589041, + "grad_norm": 2.3676326274871826, + "learning_rate": 1.143379366368806e-06, + "loss": 1.652, + "step": 2189 + }, + { + "epoch": 2.4, + "grad_norm": 2.119162082672119, + "learning_rate": 1.141348497156783e-06, + "loss": 1.6368, + "step": 2190 + }, + { + "epoch": 2.4, + "eval_loss": 1.9281524419784546, + "eval_runtime": 37.6184, + "eval_samples_per_second": 26.583, + "eval_steps_per_second": 1.116, + "step": 2190 + }, + { + "epoch": 2.401095890410959, + "grad_norm": 2.1221015453338623, + "learning_rate": 1.1393176279447604e-06, + "loss": 1.783, + "step": 2191 + }, + { + "epoch": 2.402191780821918, + "grad_norm": 2.1465466022491455, + "learning_rate": 1.1372867587327376e-06, + "loss": 1.7347, + "step": 2192 + }, + { + "epoch": 2.4032876712328766, + "grad_norm": 2.285583734512329, + "learning_rate": 1.135255889520715e-06, + "loss": 1.602, + "step": 2193 + }, + { + "epoch": 2.4043835616438356, + "grad_norm": 2.2676219940185547, + "learning_rate": 1.1332250203086922e-06, + "loss": 1.7105, + "step": 2194 + }, + { + "epoch": 2.4054794520547946, + "grad_norm": 2.2211203575134277, + "learning_rate": 1.1311941510966695e-06, + "loss": 1.6828, + "step": 2195 + }, + { + "epoch": 2.4065753424657537, + "grad_norm": 2.219998598098755, + "learning_rate": 1.1291632818846467e-06, + "loss": 1.6625, + "step": 2196 + }, + { + "epoch": 2.4076712328767123, + "grad_norm": 2.1806535720825195, + "learning_rate": 1.127132412672624e-06, + "loss": 1.6578, + "step": 2197 + }, + { + "epoch": 2.4087671232876713, + "grad_norm": 2.198413133621216, + "learning_rate": 1.1251015434606013e-06, + "loss": 1.6443, + "step": 2198 + }, + { + "epoch": 2.4098630136986303, + "grad_norm": 2.2120425701141357, + "learning_rate": 1.1230706742485786e-06, + "loss": 1.6861, + "step": 2199 + }, + { + "epoch": 2.410958904109589, + "grad_norm": 2.1654272079467773, + "learning_rate": 1.1210398050365558e-06, + "loss": 1.738, + "step": 2200 + }, + { + "epoch": 2.410958904109589, + "eval_loss": 1.9256032705307007, + "eval_runtime": 37.6204, + "eval_samples_per_second": 26.581, + "eval_steps_per_second": 1.116, + "step": 2200 + }, + { + "epoch": 2.412054794520548, + "grad_norm": 2.1871373653411865, + "learning_rate": 1.1190089358245331e-06, + "loss": 1.6806, + "step": 2201 + }, + { + "epoch": 2.413150684931507, + "grad_norm": 2.151883602142334, + "learning_rate": 1.1169780666125102e-06, + "loss": 1.6087, + "step": 2202 + }, + { + "epoch": 2.4142465753424656, + "grad_norm": 2.1763510704040527, + "learning_rate": 1.1149471974004875e-06, + "loss": 1.7085, + "step": 2203 + }, + { + "epoch": 2.4153424657534246, + "grad_norm": 2.206090211868286, + "learning_rate": 1.1129163281884647e-06, + "loss": 1.6576, + "step": 2204 + }, + { + "epoch": 2.4164383561643836, + "grad_norm": 2.198091745376587, + "learning_rate": 1.110885458976442e-06, + "loss": 1.7203, + "step": 2205 + }, + { + "epoch": 2.417534246575342, + "grad_norm": 2.081561803817749, + "learning_rate": 1.1088545897644193e-06, + "loss": 1.6447, + "step": 2206 + }, + { + "epoch": 2.4186301369863012, + "grad_norm": 2.1526243686676025, + "learning_rate": 1.1068237205523966e-06, + "loss": 1.7103, + "step": 2207 + }, + { + "epoch": 2.4197260273972603, + "grad_norm": 2.2836835384368896, + "learning_rate": 1.1047928513403736e-06, + "loss": 1.711, + "step": 2208 + }, + { + "epoch": 2.4208219178082193, + "grad_norm": 2.274479627609253, + "learning_rate": 1.102761982128351e-06, + "loss": 1.7186, + "step": 2209 + }, + { + "epoch": 2.421917808219178, + "grad_norm": 2.168281078338623, + "learning_rate": 1.1007311129163282e-06, + "loss": 1.6273, + "step": 2210 + }, + { + "epoch": 2.421917808219178, + "eval_loss": 1.9196020364761353, + "eval_runtime": 37.2362, + "eval_samples_per_second": 26.856, + "eval_steps_per_second": 1.128, + "step": 2210 + }, + { + "epoch": 2.423013698630137, + "grad_norm": 2.2091128826141357, + "learning_rate": 1.0987002437043055e-06, + "loss": 1.6756, + "step": 2211 + }, + { + "epoch": 2.424109589041096, + "grad_norm": 2.2397472858428955, + "learning_rate": 1.0966693744922827e-06, + "loss": 1.6637, + "step": 2212 + }, + { + "epoch": 2.425205479452055, + "grad_norm": 2.2397472858428955, + "learning_rate": 1.0966693744922827e-06, + "loss": 1.6172, + "step": 2213 + }, + { + "epoch": 2.4263013698630136, + "grad_norm": 2.17742919921875, + "learning_rate": 1.09463850528026e-06, + "loss": 1.6721, + "step": 2214 + }, + { + "epoch": 2.4273972602739726, + "grad_norm": 2.288299322128296, + "learning_rate": 1.0926076360682373e-06, + "loss": 1.649, + "step": 2215 + }, + { + "epoch": 2.4284931506849317, + "grad_norm": 2.217179298400879, + "learning_rate": 1.0905767668562146e-06, + "loss": 1.6632, + "step": 2216 + }, + { + "epoch": 2.4295890410958902, + "grad_norm": 2.22257661819458, + "learning_rate": 1.0885458976441918e-06, + "loss": 1.747, + "step": 2217 + }, + { + "epoch": 2.4306849315068493, + "grad_norm": 2.2523484230041504, + "learning_rate": 1.0865150284321691e-06, + "loss": 1.6839, + "step": 2218 + }, + { + "epoch": 2.4317808219178083, + "grad_norm": 2.1782052516937256, + "learning_rate": 1.0844841592201464e-06, + "loss": 1.7127, + "step": 2219 + }, + { + "epoch": 2.432876712328767, + "grad_norm": 2.270890235900879, + "learning_rate": 1.0824532900081237e-06, + "loss": 1.6766, + "step": 2220 + }, + { + "epoch": 2.432876712328767, + "eval_loss": 1.9283802509307861, + "eval_runtime": 37.4026, + "eval_samples_per_second": 26.736, + "eval_steps_per_second": 1.123, + "step": 2220 + }, + { + "epoch": 2.433972602739726, + "grad_norm": 2.153704881668091, + "learning_rate": 1.080422420796101e-06, + "loss": 1.6939, + "step": 2221 + }, + { + "epoch": 2.435068493150685, + "grad_norm": 2.175957202911377, + "learning_rate": 1.078391551584078e-06, + "loss": 1.7177, + "step": 2222 + }, + { + "epoch": 2.436164383561644, + "grad_norm": 2.194993734359741, + "learning_rate": 1.0763606823720553e-06, + "loss": 1.7507, + "step": 2223 + }, + { + "epoch": 2.4372602739726026, + "grad_norm": 2.2441670894622803, + "learning_rate": 1.0743298131600326e-06, + "loss": 1.7063, + "step": 2224 + }, + { + "epoch": 2.4383561643835616, + "grad_norm": 2.3121657371520996, + "learning_rate": 1.0722989439480098e-06, + "loss": 1.6674, + "step": 2225 + }, + { + "epoch": 2.4394520547945207, + "grad_norm": 2.1784026622772217, + "learning_rate": 1.070268074735987e-06, + "loss": 1.6558, + "step": 2226 + }, + { + "epoch": 2.4405479452054797, + "grad_norm": 2.222578287124634, + "learning_rate": 1.0682372055239644e-06, + "loss": 1.6589, + "step": 2227 + }, + { + "epoch": 2.4416438356164383, + "grad_norm": 2.332348108291626, + "learning_rate": 1.0662063363119414e-06, + "loss": 1.6551, + "step": 2228 + }, + { + "epoch": 2.4427397260273973, + "grad_norm": 2.1531267166137695, + "learning_rate": 1.0641754670999187e-06, + "loss": 1.6845, + "step": 2229 + }, + { + "epoch": 2.4438356164383563, + "grad_norm": 2.3486640453338623, + "learning_rate": 1.062144597887896e-06, + "loss": 1.6665, + "step": 2230 + }, + { + "epoch": 2.4438356164383563, + "eval_loss": 1.9210149049758911, + "eval_runtime": 37.4784, + "eval_samples_per_second": 26.682, + "eval_steps_per_second": 1.121, + "step": 2230 + }, + { + "epoch": 2.444931506849315, + "grad_norm": 2.2932095527648926, + "learning_rate": 1.0601137286758733e-06, + "loss": 1.7499, + "step": 2231 + }, + { + "epoch": 2.446027397260274, + "grad_norm": 2.1239523887634277, + "learning_rate": 1.0580828594638505e-06, + "loss": 1.588, + "step": 2232 + }, + { + "epoch": 2.447123287671233, + "grad_norm": 2.293843984603882, + "learning_rate": 1.0560519902518278e-06, + "loss": 1.7249, + "step": 2233 + }, + { + "epoch": 2.4482191780821916, + "grad_norm": 2.3284828662872314, + "learning_rate": 1.054021121039805e-06, + "loss": 1.6259, + "step": 2234 + }, + { + "epoch": 2.4493150684931506, + "grad_norm": 2.170203685760498, + "learning_rate": 1.0519902518277824e-06, + "loss": 1.6193, + "step": 2235 + }, + { + "epoch": 2.4504109589041096, + "grad_norm": 2.30904483795166, + "learning_rate": 1.0499593826157596e-06, + "loss": 1.698, + "step": 2236 + }, + { + "epoch": 2.4515068493150687, + "grad_norm": 2.183306932449341, + "learning_rate": 1.047928513403737e-06, + "loss": 1.6744, + "step": 2237 + }, + { + "epoch": 2.4526027397260273, + "grad_norm": 2.3230245113372803, + "learning_rate": 1.0458976441917142e-06, + "loss": 1.5969, + "step": 2238 + }, + { + "epoch": 2.4536986301369863, + "grad_norm": 2.194678783416748, + "learning_rate": 1.0438667749796915e-06, + "loss": 1.6106, + "step": 2239 + }, + { + "epoch": 2.4547945205479453, + "grad_norm": 2.1933088302612305, + "learning_rate": 1.0418359057676688e-06, + "loss": 1.6631, + "step": 2240 + }, + { + "epoch": 2.4547945205479453, + "eval_loss": 1.9211325645446777, + "eval_runtime": 37.5215, + "eval_samples_per_second": 26.651, + "eval_steps_per_second": 1.119, + "step": 2240 + }, + { + "epoch": 2.455890410958904, + "grad_norm": 2.304448366165161, + "learning_rate": 1.0398050365556458e-06, + "loss": 1.7081, + "step": 2241 + }, + { + "epoch": 2.456986301369863, + "grad_norm": 2.0046849250793457, + "learning_rate": 1.037774167343623e-06, + "loss": 1.6383, + "step": 2242 + }, + { + "epoch": 2.458082191780822, + "grad_norm": 2.0820834636688232, + "learning_rate": 1.0357432981316004e-06, + "loss": 1.6907, + "step": 2243 + }, + { + "epoch": 2.459178082191781, + "grad_norm": 2.2192306518554688, + "learning_rate": 1.0337124289195776e-06, + "loss": 1.7177, + "step": 2244 + }, + { + "epoch": 2.4602739726027396, + "grad_norm": 2.08642578125, + "learning_rate": 1.031681559707555e-06, + "loss": 1.7326, + "step": 2245 + }, + { + "epoch": 2.4613698630136986, + "grad_norm": 2.1756553649902344, + "learning_rate": 1.0296506904955322e-06, + "loss": 1.6746, + "step": 2246 + }, + { + "epoch": 2.4624657534246577, + "grad_norm": 2.3776814937591553, + "learning_rate": 1.0276198212835095e-06, + "loss": 1.683, + "step": 2247 + }, + { + "epoch": 2.4635616438356163, + "grad_norm": 2.3111648559570312, + "learning_rate": 1.0255889520714867e-06, + "loss": 1.6604, + "step": 2248 + }, + { + "epoch": 2.4646575342465753, + "grad_norm": 2.2604782581329346, + "learning_rate": 1.023558082859464e-06, + "loss": 1.64, + "step": 2249 + }, + { + "epoch": 2.4657534246575343, + "grad_norm": 2.1979877948760986, + "learning_rate": 1.0215272136474413e-06, + "loss": 1.7438, + "step": 2250 + }, + { + "epoch": 2.4657534246575343, + "eval_loss": 1.9184387922286987, + "eval_runtime": 37.509, + "eval_samples_per_second": 26.66, + "eval_steps_per_second": 1.12, + "step": 2250 + }, + { + "epoch": 2.466849315068493, + "grad_norm": 2.258877754211426, + "learning_rate": 1.0194963444354186e-06, + "loss": 1.6186, + "step": 2251 + }, + { + "epoch": 2.467945205479452, + "grad_norm": 2.323814868927002, + "learning_rate": 1.0174654752233958e-06, + "loss": 1.7172, + "step": 2252 + }, + { + "epoch": 2.469041095890411, + "grad_norm": 2.1510109901428223, + "learning_rate": 1.015434606011373e-06, + "loss": 1.6594, + "step": 2253 + }, + { + "epoch": 2.47013698630137, + "grad_norm": 2.115044355392456, + "learning_rate": 1.0134037367993502e-06, + "loss": 1.6818, + "step": 2254 + }, + { + "epoch": 2.4712328767123286, + "grad_norm": 2.2413828372955322, + "learning_rate": 1.0113728675873275e-06, + "loss": 1.6819, + "step": 2255 + }, + { + "epoch": 2.4723287671232876, + "grad_norm": 2.238022804260254, + "learning_rate": 1.0093419983753047e-06, + "loss": 1.6741, + "step": 2256 + }, + { + "epoch": 2.4734246575342467, + "grad_norm": 2.035796642303467, + "learning_rate": 1.007311129163282e-06, + "loss": 1.6322, + "step": 2257 + }, + { + "epoch": 2.4745205479452057, + "grad_norm": 2.0877630710601807, + "learning_rate": 1.0052802599512593e-06, + "loss": 1.6355, + "step": 2258 + }, + { + "epoch": 2.4756164383561643, + "grad_norm": 2.436210870742798, + "learning_rate": 1.0032493907392364e-06, + "loss": 1.6924, + "step": 2259 + }, + { + "epoch": 2.4767123287671233, + "grad_norm": 2.269679069519043, + "learning_rate": 1.0012185215272136e-06, + "loss": 1.611, + "step": 2260 + }, + { + "epoch": 2.4767123287671233, + "eval_loss": 1.9213950634002686, + "eval_runtime": 37.512, + "eval_samples_per_second": 26.658, + "eval_steps_per_second": 1.12, + "step": 2260 + }, + { + "epoch": 2.4778082191780824, + "grad_norm": 2.1426453590393066, + "learning_rate": 9.99187652315191e-07, + "loss": 1.7023, + "step": 2261 + }, + { + "epoch": 2.478904109589041, + "grad_norm": 2.217888116836548, + "learning_rate": 9.971567831031682e-07, + "loss": 1.6975, + "step": 2262 + }, + { + "epoch": 2.48, + "grad_norm": 2.394808530807495, + "learning_rate": 9.951259138911455e-07, + "loss": 1.6124, + "step": 2263 + }, + { + "epoch": 2.481095890410959, + "grad_norm": 2.1221909523010254, + "learning_rate": 9.930950446791227e-07, + "loss": 1.7324, + "step": 2264 + }, + { + "epoch": 2.4821917808219176, + "grad_norm": 2.2587404251098633, + "learning_rate": 9.910641754671e-07, + "loss": 1.7677, + "step": 2265 + }, + { + "epoch": 2.4832876712328766, + "grad_norm": 2.1692442893981934, + "learning_rate": 9.890333062550773e-07, + "loss": 1.6662, + "step": 2266 + }, + { + "epoch": 2.4843835616438357, + "grad_norm": 2.6075925827026367, + "learning_rate": 9.870024370430546e-07, + "loss": 1.7013, + "step": 2267 + }, + { + "epoch": 2.4854794520547947, + "grad_norm": 2.517238140106201, + "learning_rate": 9.849715678310318e-07, + "loss": 1.6809, + "step": 2268 + }, + { + "epoch": 2.4865753424657533, + "grad_norm": 2.1885604858398438, + "learning_rate": 9.829406986190091e-07, + "loss": 1.6818, + "step": 2269 + }, + { + "epoch": 2.4876712328767123, + "grad_norm": 2.3119800090789795, + "learning_rate": 9.809098294069864e-07, + "loss": 1.6104, + "step": 2270 + }, + { + "epoch": 2.4876712328767123, + "eval_loss": 1.9213876724243164, + "eval_runtime": 37.5254, + "eval_samples_per_second": 26.649, + "eval_steps_per_second": 1.119, + "step": 2270 + }, + { + "epoch": 2.4887671232876714, + "grad_norm": 2.6232759952545166, + "learning_rate": 9.788789601949637e-07, + "loss": 1.6887, + "step": 2271 + }, + { + "epoch": 2.48986301369863, + "grad_norm": 2.233729124069214, + "learning_rate": 9.768480909829407e-07, + "loss": 1.7409, + "step": 2272 + }, + { + "epoch": 2.490958904109589, + "grad_norm": 2.3070764541625977, + "learning_rate": 9.74817221770918e-07, + "loss": 1.6707, + "step": 2273 + }, + { + "epoch": 2.492054794520548, + "grad_norm": 2.7004380226135254, + "learning_rate": 9.727863525588953e-07, + "loss": 1.6829, + "step": 2274 + }, + { + "epoch": 2.493150684931507, + "grad_norm": 2.379331588745117, + "learning_rate": 9.707554833468726e-07, + "loss": 1.6785, + "step": 2275 + }, + { + "epoch": 2.4942465753424656, + "grad_norm": 2.264768123626709, + "learning_rate": 9.687246141348498e-07, + "loss": 1.666, + "step": 2276 + }, + { + "epoch": 2.4953424657534247, + "grad_norm": 3.7135798931121826, + "learning_rate": 9.666937449228271e-07, + "loss": 1.681, + "step": 2277 + }, + { + "epoch": 2.4964383561643837, + "grad_norm": 2.092416286468506, + "learning_rate": 9.646628757108042e-07, + "loss": 1.6206, + "step": 2278 + }, + { + "epoch": 2.4975342465753423, + "grad_norm": 2.2992608547210693, + "learning_rate": 9.626320064987814e-07, + "loss": 1.6997, + "step": 2279 + }, + { + "epoch": 2.4986301369863013, + "grad_norm": 2.2718007564544678, + "learning_rate": 9.606011372867587e-07, + "loss": 1.6421, + "step": 2280 + }, + { + "epoch": 2.4986301369863013, + "eval_loss": 1.9201171398162842, + "eval_runtime": 37.5634, + "eval_samples_per_second": 26.622, + "eval_steps_per_second": 1.118, + "step": 2280 + }, + { + "epoch": 2.4997260273972604, + "grad_norm": 2.708310127258301, + "learning_rate": 9.58570268074736e-07, + "loss": 1.6729, + "step": 2281 + }, + { + "epoch": 2.500821917808219, + "grad_norm": 2.2638731002807617, + "learning_rate": 9.565393988627133e-07, + "loss": 1.6731, + "step": 2282 + }, + { + "epoch": 2.501917808219178, + "grad_norm": 2.1231250762939453, + "learning_rate": 9.545085296506906e-07, + "loss": 1.6786, + "step": 2283 + }, + { + "epoch": 2.503013698630137, + "grad_norm": 2.38472843170166, + "learning_rate": 9.524776604386677e-07, + "loss": 1.6826, + "step": 2284 + }, + { + "epoch": 2.504109589041096, + "grad_norm": 2.485872983932495, + "learning_rate": 9.50446791226645e-07, + "loss": 1.6088, + "step": 2285 + }, + { + "epoch": 2.5052054794520546, + "grad_norm": 2.1208810806274414, + "learning_rate": 9.484159220146223e-07, + "loss": 1.718, + "step": 2286 + }, + { + "epoch": 2.5063013698630137, + "grad_norm": 2.219444513320923, + "learning_rate": 9.463850528025995e-07, + "loss": 1.7028, + "step": 2287 + }, + { + "epoch": 2.5073972602739727, + "grad_norm": 2.445141553878784, + "learning_rate": 9.443541835905768e-07, + "loss": 1.7477, + "step": 2288 + }, + { + "epoch": 2.5084931506849317, + "grad_norm": 2.33113956451416, + "learning_rate": 9.423233143785541e-07, + "loss": 1.7529, + "step": 2289 + }, + { + "epoch": 2.5095890410958903, + "grad_norm": 2.0926353931427, + "learning_rate": 9.402924451665313e-07, + "loss": 1.6419, + "step": 2290 + }, + { + "epoch": 2.5095890410958903, + "eval_loss": 1.9156067371368408, + "eval_runtime": 37.5683, + "eval_samples_per_second": 26.618, + "eval_steps_per_second": 1.118, + "step": 2290 + }, + { + "epoch": 2.5106849315068493, + "grad_norm": 2.1986207962036133, + "learning_rate": 9.382615759545085e-07, + "loss": 1.7222, + "step": 2291 + }, + { + "epoch": 2.5117808219178084, + "grad_norm": 2.4583442211151123, + "learning_rate": 9.362307067424858e-07, + "loss": 1.6753, + "step": 2292 + }, + { + "epoch": 2.512876712328767, + "grad_norm": 2.3618457317352295, + "learning_rate": 9.341998375304631e-07, + "loss": 1.7296, + "step": 2293 + }, + { + "epoch": 2.513972602739726, + "grad_norm": 2.289635419845581, + "learning_rate": 9.321689683184404e-07, + "loss": 1.6285, + "step": 2294 + }, + { + "epoch": 2.515068493150685, + "grad_norm": 2.252786159515381, + "learning_rate": 9.301380991064176e-07, + "loss": 1.6153, + "step": 2295 + }, + { + "epoch": 2.5161643835616436, + "grad_norm": 2.2542104721069336, + "learning_rate": 9.281072298943949e-07, + "loss": 1.6595, + "step": 2296 + }, + { + "epoch": 2.5172602739726027, + "grad_norm": 2.2887966632843018, + "learning_rate": 9.260763606823721e-07, + "loss": 1.6983, + "step": 2297 + }, + { + "epoch": 2.5183561643835617, + "grad_norm": 2.24477481842041, + "learning_rate": 9.240454914703494e-07, + "loss": 1.7007, + "step": 2298 + }, + { + "epoch": 2.5194520547945203, + "grad_norm": 2.0758657455444336, + "learning_rate": 9.220146222583266e-07, + "loss": 1.6479, + "step": 2299 + }, + { + "epoch": 2.5205479452054793, + "grad_norm": 2.2607831954956055, + "learning_rate": 9.199837530463039e-07, + "loss": 1.6781, + "step": 2300 + }, + { + "epoch": 2.5205479452054793, + "eval_loss": 1.9147255420684814, + "eval_runtime": 37.5612, + "eval_samples_per_second": 26.623, + "eval_steps_per_second": 1.118, + "step": 2300 + }, + { + "epoch": 2.5216438356164383, + "grad_norm": 2.140364170074463, + "learning_rate": 9.179528838342812e-07, + "loss": 1.6534, + "step": 2301 + }, + { + "epoch": 2.5227397260273974, + "grad_norm": 2.3352675437927246, + "learning_rate": 9.159220146222585e-07, + "loss": 1.6935, + "step": 2302 + }, + { + "epoch": 2.5238356164383564, + "grad_norm": 2.275895357131958, + "learning_rate": 9.138911454102356e-07, + "loss": 1.573, + "step": 2303 + }, + { + "epoch": 2.524931506849315, + "grad_norm": 2.193140983581543, + "learning_rate": 9.118602761982129e-07, + "loss": 1.6815, + "step": 2304 + }, + { + "epoch": 2.526027397260274, + "grad_norm": 2.325774669647217, + "learning_rate": 9.098294069861902e-07, + "loss": 1.6781, + "step": 2305 + }, + { + "epoch": 2.527123287671233, + "grad_norm": 2.2886908054351807, + "learning_rate": 9.077985377741675e-07, + "loss": 1.6385, + "step": 2306 + }, + { + "epoch": 2.5282191780821917, + "grad_norm": 2.1221580505371094, + "learning_rate": 9.057676685621447e-07, + "loss": 1.6052, + "step": 2307 + }, + { + "epoch": 2.5293150684931507, + "grad_norm": 2.2330124378204346, + "learning_rate": 9.03736799350122e-07, + "loss": 1.7295, + "step": 2308 + }, + { + "epoch": 2.5304109589041097, + "grad_norm": 2.5243897438049316, + "learning_rate": 9.017059301380991e-07, + "loss": 1.7084, + "step": 2309 + }, + { + "epoch": 2.5315068493150683, + "grad_norm": 2.1599924564361572, + "learning_rate": 8.996750609260764e-07, + "loss": 1.7427, + "step": 2310 + }, + { + "epoch": 2.5315068493150683, + "eval_loss": 1.9132241010665894, + "eval_runtime": 37.2999, + "eval_samples_per_second": 26.81, + "eval_steps_per_second": 1.126, + "step": 2310 + }, + { + "epoch": 2.5326027397260273, + "grad_norm": 2.195263624191284, + "learning_rate": 8.976441917140536e-07, + "loss": 1.7042, + "step": 2311 + }, + { + "epoch": 2.5336986301369864, + "grad_norm": 2.250441074371338, + "learning_rate": 8.956133225020309e-07, + "loss": 1.7438, + "step": 2312 + }, + { + "epoch": 2.534794520547945, + "grad_norm": 2.1880240440368652, + "learning_rate": 8.935824532900082e-07, + "loss": 1.6439, + "step": 2313 + }, + { + "epoch": 2.535890410958904, + "grad_norm": 2.1684138774871826, + "learning_rate": 8.915515840779855e-07, + "loss": 1.6225, + "step": 2314 + }, + { + "epoch": 2.536986301369863, + "grad_norm": 2.205507755279541, + "learning_rate": 8.895207148659626e-07, + "loss": 1.7052, + "step": 2315 + }, + { + "epoch": 2.538082191780822, + "grad_norm": 2.278517484664917, + "learning_rate": 8.874898456539399e-07, + "loss": 1.6825, + "step": 2316 + }, + { + "epoch": 2.539178082191781, + "grad_norm": 2.268779993057251, + "learning_rate": 8.854589764419172e-07, + "loss": 1.6872, + "step": 2317 + }, + { + "epoch": 2.5402739726027397, + "grad_norm": 2.1956069469451904, + "learning_rate": 8.834281072298945e-07, + "loss": 1.6334, + "step": 2318 + }, + { + "epoch": 2.5413698630136987, + "grad_norm": 2.166696310043335, + "learning_rate": 8.813972380178717e-07, + "loss": 1.6731, + "step": 2319 + }, + { + "epoch": 2.5424657534246577, + "grad_norm": 2.070521831512451, + "learning_rate": 8.79366368805849e-07, + "loss": 1.6905, + "step": 2320 + }, + { + "epoch": 2.5424657534246577, + "eval_loss": 1.9162200689315796, + "eval_runtime": 37.442, + "eval_samples_per_second": 26.708, + "eval_steps_per_second": 1.122, + "step": 2320 + }, + { + "epoch": 2.5435616438356163, + "grad_norm": 2.252784013748169, + "learning_rate": 8.773354995938262e-07, + "loss": 1.6331, + "step": 2321 + }, + { + "epoch": 2.5446575342465754, + "grad_norm": 2.200810432434082, + "learning_rate": 8.753046303818035e-07, + "loss": 1.6316, + "step": 2322 + }, + { + "epoch": 2.5457534246575344, + "grad_norm": 2.325618028640747, + "learning_rate": 8.732737611697807e-07, + "loss": 1.6734, + "step": 2323 + }, + { + "epoch": 2.546849315068493, + "grad_norm": 2.1776297092437744, + "learning_rate": 8.71242891957758e-07, + "loss": 1.6928, + "step": 2324 + }, + { + "epoch": 2.547945205479452, + "grad_norm": 2.1836626529693604, + "learning_rate": 8.692120227457353e-07, + "loss": 1.7099, + "step": 2325 + }, + { + "epoch": 2.549041095890411, + "grad_norm": 2.1865885257720947, + "learning_rate": 8.671811535337126e-07, + "loss": 1.6542, + "step": 2326 + }, + { + "epoch": 2.5501369863013696, + "grad_norm": 2.1699092388153076, + "learning_rate": 8.651502843216898e-07, + "loss": 1.7535, + "step": 2327 + }, + { + "epoch": 2.5512328767123287, + "grad_norm": 2.1458663940429688, + "learning_rate": 8.63119415109667e-07, + "loss": 1.7013, + "step": 2328 + }, + { + "epoch": 2.5523287671232877, + "grad_norm": 2.135432720184326, + "learning_rate": 8.610885458976443e-07, + "loss": 1.6924, + "step": 2329 + }, + { + "epoch": 2.5534246575342463, + "grad_norm": 2.2062275409698486, + "learning_rate": 8.590576766856216e-07, + "loss": 1.6503, + "step": 2330 + }, + { + "epoch": 2.5534246575342463, + "eval_loss": 1.9146372079849243, + "eval_runtime": 37.5755, + "eval_samples_per_second": 26.613, + "eval_steps_per_second": 1.118, + "step": 2330 + }, + { + "epoch": 2.5545205479452053, + "grad_norm": 2.1375393867492676, + "learning_rate": 8.570268074735988e-07, + "loss": 1.6297, + "step": 2331 + }, + { + "epoch": 2.5556164383561644, + "grad_norm": 2.126147747039795, + "learning_rate": 8.549959382615761e-07, + "loss": 1.6601, + "step": 2332 + }, + { + "epoch": 2.5567123287671234, + "grad_norm": 2.202545642852783, + "learning_rate": 8.529650690495534e-07, + "loss": 1.6783, + "step": 2333 + }, + { + "epoch": 2.5578082191780824, + "grad_norm": 2.222506284713745, + "learning_rate": 8.509341998375305e-07, + "loss": 1.6106, + "step": 2334 + }, + { + "epoch": 2.558904109589041, + "grad_norm": 2.083980083465576, + "learning_rate": 8.489033306255077e-07, + "loss": 1.5909, + "step": 2335 + }, + { + "epoch": 2.56, + "grad_norm": 2.1030898094177246, + "learning_rate": 8.46872461413485e-07, + "loss": 1.6371, + "step": 2336 + }, + { + "epoch": 2.561095890410959, + "grad_norm": 2.192338705062866, + "learning_rate": 8.448415922014623e-07, + "loss": 1.6557, + "step": 2337 + }, + { + "epoch": 2.5621917808219177, + "grad_norm": 2.1797902584075928, + "learning_rate": 8.428107229894396e-07, + "loss": 1.6899, + "step": 2338 + }, + { + "epoch": 2.5632876712328767, + "grad_norm": 2.2627408504486084, + "learning_rate": 8.407798537774168e-07, + "loss": 1.6723, + "step": 2339 + }, + { + "epoch": 2.5643835616438357, + "grad_norm": 2.270190477371216, + "learning_rate": 8.38748984565394e-07, + "loss": 1.6442, + "step": 2340 + }, + { + "epoch": 2.5643835616438357, + "eval_loss": 1.9147305488586426, + "eval_runtime": 37.5642, + "eval_samples_per_second": 26.621, + "eval_steps_per_second": 1.118, + "step": 2340 + }, + { + "epoch": 2.5654794520547943, + "grad_norm": 2.1523330211639404, + "learning_rate": 8.367181153533713e-07, + "loss": 1.6614, + "step": 2341 + }, + { + "epoch": 2.5665753424657534, + "grad_norm": 2.1988766193389893, + "learning_rate": 8.346872461413486e-07, + "loss": 1.655, + "step": 2342 + }, + { + "epoch": 2.5676712328767124, + "grad_norm": 2.2519428730010986, + "learning_rate": 8.326563769293258e-07, + "loss": 1.6576, + "step": 2343 + }, + { + "epoch": 2.568767123287671, + "grad_norm": 2.122844696044922, + "learning_rate": 8.306255077173031e-07, + "loss": 1.7568, + "step": 2344 + }, + { + "epoch": 2.56986301369863, + "grad_norm": 2.21121883392334, + "learning_rate": 8.285946385052804e-07, + "loss": 1.6139, + "step": 2345 + }, + { + "epoch": 2.570958904109589, + "grad_norm": 2.2146055698394775, + "learning_rate": 8.265637692932575e-07, + "loss": 1.6947, + "step": 2346 + }, + { + "epoch": 2.572054794520548, + "grad_norm": 2.220987558364868, + "learning_rate": 8.245329000812348e-07, + "loss": 1.6813, + "step": 2347 + }, + { + "epoch": 2.573150684931507, + "grad_norm": 2.124702215194702, + "learning_rate": 8.225020308692121e-07, + "loss": 1.6212, + "step": 2348 + }, + { + "epoch": 2.5742465753424657, + "grad_norm": 2.3794567584991455, + "learning_rate": 8.204711616571894e-07, + "loss": 1.6762, + "step": 2349 + }, + { + "epoch": 2.5753424657534247, + "grad_norm": 2.263453722000122, + "learning_rate": 8.184402924451667e-07, + "loss": 1.619, + "step": 2350 + }, + { + "epoch": 2.5753424657534247, + "eval_loss": 1.91135573387146, + "eval_runtime": 37.5946, + "eval_samples_per_second": 26.6, + "eval_steps_per_second": 1.117, + "step": 2350 + }, + { + "epoch": 2.5764383561643838, + "grad_norm": 2.0131776332855225, + "learning_rate": 8.164094232331439e-07, + "loss": 1.6982, + "step": 2351 + }, + { + "epoch": 2.5775342465753424, + "grad_norm": 2.0582032203674316, + "learning_rate": 8.14378554021121e-07, + "loss": 1.6791, + "step": 2352 + }, + { + "epoch": 2.5786301369863014, + "grad_norm": 2.2664265632629395, + "learning_rate": 8.123476848090983e-07, + "loss": 1.6464, + "step": 2353 + }, + { + "epoch": 2.5797260273972604, + "grad_norm": 2.2123348712921143, + "learning_rate": 8.103168155970755e-07, + "loss": 1.6536, + "step": 2354 + }, + { + "epoch": 2.580821917808219, + "grad_norm": 2.2500033378601074, + "learning_rate": 8.082859463850528e-07, + "loss": 1.6847, + "step": 2355 + }, + { + "epoch": 2.581917808219178, + "grad_norm": 2.2028374671936035, + "learning_rate": 8.062550771730301e-07, + "loss": 1.6575, + "step": 2356 + }, + { + "epoch": 2.583013698630137, + "grad_norm": 2.2103688716888428, + "learning_rate": 8.042242079610074e-07, + "loss": 1.6471, + "step": 2357 + }, + { + "epoch": 2.5841095890410957, + "grad_norm": 2.1425282955169678, + "learning_rate": 8.021933387489846e-07, + "loss": 1.6495, + "step": 2358 + }, + { + "epoch": 2.5852054794520547, + "grad_norm": 2.0925207138061523, + "learning_rate": 8.001624695369618e-07, + "loss": 1.6749, + "step": 2359 + }, + { + "epoch": 2.5863013698630137, + "grad_norm": 2.415372848510742, + "learning_rate": 7.981316003249391e-07, + "loss": 1.7333, + "step": 2360 + }, + { + "epoch": 2.5863013698630137, + "eval_loss": 1.9135847091674805, + "eval_runtime": 37.5734, + "eval_samples_per_second": 26.615, + "eval_steps_per_second": 1.118, + "step": 2360 + }, + { + "epoch": 2.5873972602739728, + "grad_norm": 2.320563316345215, + "learning_rate": 7.961007311129164e-07, + "loss": 1.609, + "step": 2361 + }, + { + "epoch": 2.5884931506849314, + "grad_norm": 2.2953672409057617, + "learning_rate": 7.940698619008936e-07, + "loss": 1.6928, + "step": 2362 + }, + { + "epoch": 2.5895890410958904, + "grad_norm": 2.7120959758758545, + "learning_rate": 7.920389926888709e-07, + "loss": 1.7727, + "step": 2363 + }, + { + "epoch": 2.5906849315068494, + "grad_norm": 2.1923747062683105, + "learning_rate": 7.900081234768482e-07, + "loss": 1.6342, + "step": 2364 + }, + { + "epoch": 2.5917808219178085, + "grad_norm": 2.1920101642608643, + "learning_rate": 7.879772542648254e-07, + "loss": 1.6886, + "step": 2365 + }, + { + "epoch": 2.592876712328767, + "grad_norm": 3.0836801528930664, + "learning_rate": 7.859463850528026e-07, + "loss": 1.6844, + "step": 2366 + }, + { + "epoch": 2.593972602739726, + "grad_norm": 2.223613739013672, + "learning_rate": 7.839155158407799e-07, + "loss": 1.6819, + "step": 2367 + }, + { + "epoch": 2.595068493150685, + "grad_norm": 2.1360442638397217, + "learning_rate": 7.818846466287572e-07, + "loss": 1.6829, + "step": 2368 + }, + { + "epoch": 2.5961643835616437, + "grad_norm": 2.620152473449707, + "learning_rate": 7.798537774167345e-07, + "loss": 1.7054, + "step": 2369 + }, + { + "epoch": 2.5972602739726027, + "grad_norm": 2.255185604095459, + "learning_rate": 7.778229082047117e-07, + "loss": 1.6383, + "step": 2370 + }, + { + "epoch": 2.5972602739726027, + "eval_loss": 1.91743803024292, + "eval_runtime": 37.6008, + "eval_samples_per_second": 26.595, + "eval_steps_per_second": 1.117, + "step": 2370 + }, + { + "epoch": 2.5983561643835618, + "grad_norm": 2.223876714706421, + "learning_rate": 7.757920389926889e-07, + "loss": 1.6836, + "step": 2371 + }, + { + "epoch": 2.5994520547945204, + "grad_norm": 2.147221088409424, + "learning_rate": 7.737611697806662e-07, + "loss": 1.6515, + "step": 2372 + }, + { + "epoch": 2.6005479452054794, + "grad_norm": 2.402515172958374, + "learning_rate": 7.717303005686435e-07, + "loss": 1.6994, + "step": 2373 + }, + { + "epoch": 2.6016438356164384, + "grad_norm": 2.3239173889160156, + "learning_rate": 7.696994313566207e-07, + "loss": 1.7392, + "step": 2374 + }, + { + "epoch": 2.602739726027397, + "grad_norm": 2.2326531410217285, + "learning_rate": 7.67668562144598e-07, + "loss": 1.6284, + "step": 2375 + }, + { + "epoch": 2.603835616438356, + "grad_norm": 2.0969064235687256, + "learning_rate": 7.656376929325753e-07, + "loss": 1.6205, + "step": 2376 + }, + { + "epoch": 2.604931506849315, + "grad_norm": 2.111938238143921, + "learning_rate": 7.636068237205524e-07, + "loss": 1.7074, + "step": 2377 + }, + { + "epoch": 2.606027397260274, + "grad_norm": 2.159001111984253, + "learning_rate": 7.615759545085296e-07, + "loss": 1.6288, + "step": 2378 + }, + { + "epoch": 2.607123287671233, + "grad_norm": 2.138695001602173, + "learning_rate": 7.595450852965069e-07, + "loss": 1.6489, + "step": 2379 + }, + { + "epoch": 2.6082191780821917, + "grad_norm": 2.127330780029297, + "learning_rate": 7.575142160844842e-07, + "loss": 1.6911, + "step": 2380 + }, + { + "epoch": 2.6082191780821917, + "eval_loss": 1.91731595993042, + "eval_runtime": 37.6324, + "eval_samples_per_second": 26.573, + "eval_steps_per_second": 1.116, + "step": 2380 + }, + { + "epoch": 2.6093150684931508, + "grad_norm": 2.2218921184539795, + "learning_rate": 7.554833468724615e-07, + "loss": 1.6333, + "step": 2381 + }, + { + "epoch": 2.61041095890411, + "grad_norm": 2.299722671508789, + "learning_rate": 7.534524776604387e-07, + "loss": 1.6435, + "step": 2382 + }, + { + "epoch": 2.6115068493150684, + "grad_norm": 2.387270212173462, + "learning_rate": 7.51421608448416e-07, + "loss": 1.6874, + "step": 2383 + }, + { + "epoch": 2.6126027397260274, + "grad_norm": 2.0951170921325684, + "learning_rate": 7.493907392363932e-07, + "loss": 1.6419, + "step": 2384 + }, + { + "epoch": 2.6136986301369864, + "grad_norm": 2.1200127601623535, + "learning_rate": 7.473598700243705e-07, + "loss": 1.598, + "step": 2385 + }, + { + "epoch": 2.614794520547945, + "grad_norm": 2.372738838195801, + "learning_rate": 7.453290008123477e-07, + "loss": 1.6631, + "step": 2386 + }, + { + "epoch": 2.615890410958904, + "grad_norm": 2.1989450454711914, + "learning_rate": 7.43298131600325e-07, + "loss": 1.6707, + "step": 2387 + }, + { + "epoch": 2.616986301369863, + "grad_norm": 2.139631509780884, + "learning_rate": 7.412672623883023e-07, + "loss": 1.6698, + "step": 2388 + }, + { + "epoch": 2.6180821917808217, + "grad_norm": 2.2012128829956055, + "learning_rate": 7.392363931762796e-07, + "loss": 1.6113, + "step": 2389 + }, + { + "epoch": 2.6191780821917807, + "grad_norm": 2.193068265914917, + "learning_rate": 7.372055239642567e-07, + "loss": 1.6687, + "step": 2390 + }, + { + "epoch": 2.6191780821917807, + "eval_loss": 1.9159072637557983, + "eval_runtime": 37.6046, + "eval_samples_per_second": 26.593, + "eval_steps_per_second": 1.117, + "step": 2390 + }, + { + "epoch": 2.6202739726027398, + "grad_norm": 2.2286741733551025, + "learning_rate": 7.35174654752234e-07, + "loss": 1.6887, + "step": 2391 + }, + { + "epoch": 2.621369863013699, + "grad_norm": 2.159212589263916, + "learning_rate": 7.331437855402113e-07, + "loss": 1.7208, + "step": 2392 + }, + { + "epoch": 2.6224657534246574, + "grad_norm": 2.2144219875335693, + "learning_rate": 7.311129163281886e-07, + "loss": 1.6324, + "step": 2393 + }, + { + "epoch": 2.6235616438356164, + "grad_norm": 2.1681270599365234, + "learning_rate": 7.290820471161658e-07, + "loss": 1.6934, + "step": 2394 + }, + { + "epoch": 2.6246575342465754, + "grad_norm": 2.2149975299835205, + "learning_rate": 7.270511779041431e-07, + "loss": 1.6905, + "step": 2395 + }, + { + "epoch": 2.6257534246575345, + "grad_norm": 2.1379663944244385, + "learning_rate": 7.250203086921203e-07, + "loss": 1.6117, + "step": 2396 + }, + { + "epoch": 2.626849315068493, + "grad_norm": 2.127993583679199, + "learning_rate": 7.229894394800976e-07, + "loss": 1.6665, + "step": 2397 + }, + { + "epoch": 2.627945205479452, + "grad_norm": 2.156416416168213, + "learning_rate": 7.209585702680748e-07, + "loss": 1.6928, + "step": 2398 + }, + { + "epoch": 2.629041095890411, + "grad_norm": 2.118269443511963, + "learning_rate": 7.189277010560521e-07, + "loss": 1.6213, + "step": 2399 + }, + { + "epoch": 2.6301369863013697, + "grad_norm": 2.1322832107543945, + "learning_rate": 7.168968318440294e-07, + "loss": 1.6417, + "step": 2400 + }, + { + "epoch": 2.6301369863013697, + "eval_loss": 1.9078015089035034, + "eval_runtime": 37.6016, + "eval_samples_per_second": 26.595, + "eval_steps_per_second": 1.117, + "step": 2400 + }, + { + "epoch": 2.6312328767123287, + "grad_norm": 2.217540979385376, + "learning_rate": 7.148659626320067e-07, + "loss": 1.6142, + "step": 2401 + }, + { + "epoch": 2.632328767123288, + "grad_norm": 2.2981984615325928, + "learning_rate": 7.128350934199837e-07, + "loss": 1.7361, + "step": 2402 + }, + { + "epoch": 2.6334246575342464, + "grad_norm": 2.043351173400879, + "learning_rate": 7.10804224207961e-07, + "loss": 1.7298, + "step": 2403 + }, + { + "epoch": 2.6345205479452054, + "grad_norm": 2.1623260974884033, + "learning_rate": 7.087733549959383e-07, + "loss": 1.6638, + "step": 2404 + }, + { + "epoch": 2.6356164383561644, + "grad_norm": 2.2542014122009277, + "learning_rate": 7.067424857839155e-07, + "loss": 1.6509, + "step": 2405 + }, + { + "epoch": 2.636712328767123, + "grad_norm": 2.2501258850097656, + "learning_rate": 7.047116165718928e-07, + "loss": 1.6586, + "step": 2406 + }, + { + "epoch": 2.637808219178082, + "grad_norm": 2.2709550857543945, + "learning_rate": 7.026807473598701e-07, + "loss": 1.6625, + "step": 2407 + }, + { + "epoch": 2.638904109589041, + "grad_norm": 2.2488975524902344, + "learning_rate": 7.006498781478473e-07, + "loss": 1.7293, + "step": 2408 + }, + { + "epoch": 2.64, + "grad_norm": 2.3725569248199463, + "learning_rate": 6.986190089358245e-07, + "loss": 1.6214, + "step": 2409 + }, + { + "epoch": 2.641095890410959, + "grad_norm": 2.1760759353637695, + "learning_rate": 6.965881397238018e-07, + "loss": 1.6795, + "step": 2410 + }, + { + "epoch": 2.641095890410959, + "eval_loss": 1.90324068069458, + "eval_runtime": 37.3373, + "eval_samples_per_second": 26.783, + "eval_steps_per_second": 1.125, + "step": 2410 + }, + { + "epoch": 2.6421917808219177, + "grad_norm": 2.073680877685547, + "learning_rate": 6.945572705117791e-07, + "loss": 1.6773, + "step": 2411 + }, + { + "epoch": 2.643287671232877, + "grad_norm": 2.1407957077026367, + "learning_rate": 6.925264012997564e-07, + "loss": 1.6734, + "step": 2412 + }, + { + "epoch": 2.644383561643836, + "grad_norm": 2.128048896789551, + "learning_rate": 6.904955320877336e-07, + "loss": 1.6853, + "step": 2413 + }, + { + "epoch": 2.6454794520547944, + "grad_norm": 2.2551848888397217, + "learning_rate": 6.884646628757109e-07, + "loss": 1.6868, + "step": 2414 + }, + { + "epoch": 2.6465753424657534, + "grad_norm": 2.2478785514831543, + "learning_rate": 6.864337936636881e-07, + "loss": 1.7482, + "step": 2415 + }, + { + "epoch": 2.6476712328767125, + "grad_norm": 2.2035670280456543, + "learning_rate": 6.844029244516654e-07, + "loss": 1.721, + "step": 2416 + }, + { + "epoch": 2.648767123287671, + "grad_norm": 2.1180107593536377, + "learning_rate": 6.823720552396426e-07, + "loss": 1.6264, + "step": 2417 + }, + { + "epoch": 2.64986301369863, + "grad_norm": 2.172999858856201, + "learning_rate": 6.803411860276199e-07, + "loss": 1.6825, + "step": 2418 + }, + { + "epoch": 2.650958904109589, + "grad_norm": 2.282914638519287, + "learning_rate": 6.783103168155972e-07, + "loss": 1.6677, + "step": 2419 + }, + { + "epoch": 2.6520547945205477, + "grad_norm": 2.21195387840271, + "learning_rate": 6.762794476035745e-07, + "loss": 1.655, + "step": 2420 + }, + { + "epoch": 2.6520547945205477, + "eval_loss": 1.902004361152649, + "eval_runtime": 37.5055, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 2420 + }, + { + "epoch": 2.6531506849315067, + "grad_norm": 2.1227028369903564, + "learning_rate": 6.742485783915515e-07, + "loss": 1.7057, + "step": 2421 + }, + { + "epoch": 2.6542465753424658, + "grad_norm": 2.1736481189727783, + "learning_rate": 6.722177091795288e-07, + "loss": 1.6776, + "step": 2422 + }, + { + "epoch": 2.655342465753425, + "grad_norm": 2.242764711380005, + "learning_rate": 6.701868399675061e-07, + "loss": 1.6788, + "step": 2423 + }, + { + "epoch": 2.6564383561643834, + "grad_norm": 2.2311532497406006, + "learning_rate": 6.681559707554834e-07, + "loss": 1.6318, + "step": 2424 + }, + { + "epoch": 2.6575342465753424, + "grad_norm": 2.1354873180389404, + "learning_rate": 6.661251015434606e-07, + "loss": 1.6435, + "step": 2425 + }, + { + "epoch": 2.6586301369863015, + "grad_norm": 2.1092300415039062, + "learning_rate": 6.640942323314379e-07, + "loss": 1.6318, + "step": 2426 + }, + { + "epoch": 2.6597260273972605, + "grad_norm": 2.161104917526245, + "learning_rate": 6.620633631194151e-07, + "loss": 1.6516, + "step": 2427 + }, + { + "epoch": 2.660821917808219, + "grad_norm": 2.2475714683532715, + "learning_rate": 6.600324939073924e-07, + "loss": 1.6783, + "step": 2428 + }, + { + "epoch": 2.661917808219178, + "grad_norm": 2.2516846656799316, + "learning_rate": 6.580016246953696e-07, + "loss": 1.6604, + "step": 2429 + }, + { + "epoch": 2.663013698630137, + "grad_norm": 2.3203659057617188, + "learning_rate": 6.559707554833469e-07, + "loss": 1.7081, + "step": 2430 + }, + { + "epoch": 2.663013698630137, + "eval_loss": 1.9063129425048828, + "eval_runtime": 37.6131, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 2430 + }, + { + "epoch": 2.6641095890410957, + "grad_norm": 2.1518654823303223, + "learning_rate": 6.539398862713242e-07, + "loss": 1.6486, + "step": 2431 + }, + { + "epoch": 2.6652054794520548, + "grad_norm": 2.1912600994110107, + "learning_rate": 6.519090170593015e-07, + "loss": 1.7074, + "step": 2432 + }, + { + "epoch": 2.666301369863014, + "grad_norm": 2.2298574447631836, + "learning_rate": 6.498781478472786e-07, + "loss": 1.6775, + "step": 2433 + }, + { + "epoch": 2.6673972602739724, + "grad_norm": 2.3464322090148926, + "learning_rate": 6.478472786352559e-07, + "loss": 1.7186, + "step": 2434 + }, + { + "epoch": 2.6684931506849314, + "grad_norm": 2.271078586578369, + "learning_rate": 6.458164094232332e-07, + "loss": 1.6688, + "step": 2435 + }, + { + "epoch": 2.6695890410958905, + "grad_norm": 2.299139976501465, + "learning_rate": 6.437855402112105e-07, + "loss": 1.6011, + "step": 2436 + }, + { + "epoch": 2.670684931506849, + "grad_norm": 2.263948678970337, + "learning_rate": 6.417546709991877e-07, + "loss": 1.636, + "step": 2437 + }, + { + "epoch": 2.671780821917808, + "grad_norm": 2.3219411373138428, + "learning_rate": 6.39723801787165e-07, + "loss": 1.7189, + "step": 2438 + }, + { + "epoch": 2.672876712328767, + "grad_norm": 2.2561123371124268, + "learning_rate": 6.376929325751423e-07, + "loss": 1.6524, + "step": 2439 + }, + { + "epoch": 2.673972602739726, + "grad_norm": 2.0937557220458984, + "learning_rate": 6.356620633631195e-07, + "loss": 1.6517, + "step": 2440 + }, + { + "epoch": 2.673972602739726, + "eval_loss": 1.907504916191101, + "eval_runtime": 37.6378, + "eval_samples_per_second": 26.569, + "eval_steps_per_second": 1.116, + "step": 2440 + }, + { + "epoch": 2.675068493150685, + "grad_norm": 2.284665822982788, + "learning_rate": 6.336311941510967e-07, + "loss": 1.6488, + "step": 2441 + }, + { + "epoch": 2.6761643835616438, + "grad_norm": 2.3389320373535156, + "learning_rate": 6.31600324939074e-07, + "loss": 1.6763, + "step": 2442 + }, + { + "epoch": 2.677260273972603, + "grad_norm": 2.224381446838379, + "learning_rate": 6.295694557270513e-07, + "loss": 1.7135, + "step": 2443 + }, + { + "epoch": 2.678356164383562, + "grad_norm": 2.1719634532928467, + "learning_rate": 6.275385865150286e-07, + "loss": 1.6865, + "step": 2444 + }, + { + "epoch": 2.6794520547945204, + "grad_norm": 2.220154047012329, + "learning_rate": 6.255077173030058e-07, + "loss": 1.6695, + "step": 2445 + }, + { + "epoch": 2.6805479452054795, + "grad_norm": 2.134202480316162, + "learning_rate": 6.23476848090983e-07, + "loss": 1.6711, + "step": 2446 + }, + { + "epoch": 2.6816438356164385, + "grad_norm": 2.333765983581543, + "learning_rate": 6.214459788789602e-07, + "loss": 1.6577, + "step": 2447 + }, + { + "epoch": 2.682739726027397, + "grad_norm": 2.2139508724212646, + "learning_rate": 6.194151096669375e-07, + "loss": 1.6946, + "step": 2448 + }, + { + "epoch": 2.683835616438356, + "grad_norm": 2.263981819152832, + "learning_rate": 6.173842404549147e-07, + "loss": 1.6586, + "step": 2449 + }, + { + "epoch": 2.684931506849315, + "grad_norm": 2.2153079509735107, + "learning_rate": 6.15353371242892e-07, + "loss": 1.6666, + "step": 2450 + }, + { + "epoch": 2.684931506849315, + "eval_loss": 1.9090298414230347, + "eval_runtime": 37.6472, + "eval_samples_per_second": 26.562, + "eval_steps_per_second": 1.116, + "step": 2450 + }, + { + "epoch": 2.6860273972602737, + "grad_norm": 2.166443109512329, + "learning_rate": 6.133225020308693e-07, + "loss": 1.5851, + "step": 2451 + }, + { + "epoch": 2.6871232876712328, + "grad_norm": 2.0849578380584717, + "learning_rate": 6.112916328188466e-07, + "loss": 1.6605, + "step": 2452 + }, + { + "epoch": 2.688219178082192, + "grad_norm": 2.262051582336426, + "learning_rate": 6.092607636068238e-07, + "loss": 1.7255, + "step": 2453 + }, + { + "epoch": 2.689315068493151, + "grad_norm": 2.2105391025543213, + "learning_rate": 6.07229894394801e-07, + "loss": 1.6394, + "step": 2454 + }, + { + "epoch": 2.69041095890411, + "grad_norm": 2.0528581142425537, + "learning_rate": 6.051990251827783e-07, + "loss": 1.6459, + "step": 2455 + }, + { + "epoch": 2.6915068493150685, + "grad_norm": 2.111941337585449, + "learning_rate": 6.031681559707556e-07, + "loss": 1.6892, + "step": 2456 + }, + { + "epoch": 2.6926027397260275, + "grad_norm": 2.2374067306518555, + "learning_rate": 6.011372867587327e-07, + "loss": 1.6721, + "step": 2457 + }, + { + "epoch": 2.6936986301369865, + "grad_norm": 2.0855772495269775, + "learning_rate": 5.9910641754671e-07, + "loss": 1.6401, + "step": 2458 + }, + { + "epoch": 2.694794520547945, + "grad_norm": 2.3623082637786865, + "learning_rate": 5.970755483346873e-07, + "loss": 1.6512, + "step": 2459 + }, + { + "epoch": 2.695890410958904, + "grad_norm": 2.182384490966797, + "learning_rate": 5.950446791226646e-07, + "loss": 1.6478, + "step": 2460 + }, + { + "epoch": 2.695890410958904, + "eval_loss": 1.9096838235855103, + "eval_runtime": 37.632, + "eval_samples_per_second": 26.573, + "eval_steps_per_second": 1.116, + "step": 2460 + }, + { + "epoch": 2.696986301369863, + "grad_norm": 2.181840419769287, + "learning_rate": 5.930138099106418e-07, + "loss": 1.667, + "step": 2461 + }, + { + "epoch": 2.6980821917808218, + "grad_norm": 2.157594680786133, + "learning_rate": 5.909829406986191e-07, + "loss": 1.6392, + "step": 2462 + }, + { + "epoch": 2.699178082191781, + "grad_norm": 2.0908291339874268, + "learning_rate": 5.889520714865963e-07, + "loss": 1.672, + "step": 2463 + }, + { + "epoch": 2.70027397260274, + "grad_norm": 2.060753345489502, + "learning_rate": 5.869212022745735e-07, + "loss": 1.6199, + "step": 2464 + }, + { + "epoch": 2.7013698630136984, + "grad_norm": 2.1930088996887207, + "learning_rate": 5.848903330625508e-07, + "loss": 1.6728, + "step": 2465 + }, + { + "epoch": 2.7024657534246574, + "grad_norm": 2.054527759552002, + "learning_rate": 5.828594638505281e-07, + "loss": 1.6783, + "step": 2466 + }, + { + "epoch": 2.7035616438356165, + "grad_norm": 2.160984516143799, + "learning_rate": 5.808285946385054e-07, + "loss": 1.652, + "step": 2467 + }, + { + "epoch": 2.704657534246575, + "grad_norm": 2.248612880706787, + "learning_rate": 5.787977254264827e-07, + "loss": 1.6827, + "step": 2468 + }, + { + "epoch": 2.705753424657534, + "grad_norm": 2.293546199798584, + "learning_rate": 5.767668562144598e-07, + "loss": 1.7021, + "step": 2469 + }, + { + "epoch": 2.706849315068493, + "grad_norm": 2.2298977375030518, + "learning_rate": 5.747359870024371e-07, + "loss": 1.6541, + "step": 2470 + }, + { + "epoch": 2.706849315068493, + "eval_loss": 1.9131715297698975, + "eval_runtime": 37.6667, + "eval_samples_per_second": 26.549, + "eval_steps_per_second": 1.115, + "step": 2470 + }, + { + "epoch": 2.707945205479452, + "grad_norm": 2.1208701133728027, + "learning_rate": 5.727051177904144e-07, + "loss": 1.6577, + "step": 2471 + }, + { + "epoch": 2.709041095890411, + "grad_norm": 2.1882405281066895, + "learning_rate": 5.706742485783915e-07, + "loss": 1.6995, + "step": 2472 + }, + { + "epoch": 2.71013698630137, + "grad_norm": 2.2345633506774902, + "learning_rate": 5.686433793663688e-07, + "loss": 1.6345, + "step": 2473 + }, + { + "epoch": 2.711232876712329, + "grad_norm": 2.1731173992156982, + "learning_rate": 5.666125101543461e-07, + "loss": 1.6066, + "step": 2474 + }, + { + "epoch": 2.712328767123288, + "grad_norm": 2.116769552230835, + "learning_rate": 5.645816409423234e-07, + "loss": 1.7164, + "step": 2475 + }, + { + "epoch": 2.7134246575342464, + "grad_norm": 2.104316473007202, + "learning_rate": 5.625507717303006e-07, + "loss": 1.6722, + "step": 2476 + }, + { + "epoch": 2.7145205479452055, + "grad_norm": 2.294095277786255, + "learning_rate": 5.605199025182779e-07, + "loss": 1.6491, + "step": 2477 + }, + { + "epoch": 2.7156164383561645, + "grad_norm": 2.224013090133667, + "learning_rate": 5.584890333062551e-07, + "loss": 1.6415, + "step": 2478 + }, + { + "epoch": 2.716712328767123, + "grad_norm": 2.131150007247925, + "learning_rate": 5.564581640942324e-07, + "loss": 1.6996, + "step": 2479 + }, + { + "epoch": 2.717808219178082, + "grad_norm": 2.246748208999634, + "learning_rate": 5.544272948822096e-07, + "loss": 1.6479, + "step": 2480 + }, + { + "epoch": 2.717808219178082, + "eval_loss": 1.9096418619155884, + "eval_runtime": 37.6177, + "eval_samples_per_second": 26.583, + "eval_steps_per_second": 1.116, + "step": 2480 + }, + { + "epoch": 2.718904109589041, + "grad_norm": 2.2861924171447754, + "learning_rate": 5.523964256701868e-07, + "loss": 1.653, + "step": 2481 + }, + { + "epoch": 2.7199999999999998, + "grad_norm": 2.17336368560791, + "learning_rate": 5.503655564581641e-07, + "loss": 1.677, + "step": 2482 + }, + { + "epoch": 2.721095890410959, + "grad_norm": 2.13470721244812, + "learning_rate": 5.483346872461414e-07, + "loss": 1.6476, + "step": 2483 + }, + { + "epoch": 2.722191780821918, + "grad_norm": 2.082038640975952, + "learning_rate": 5.463038180341186e-07, + "loss": 1.6623, + "step": 2484 + }, + { + "epoch": 2.723287671232877, + "grad_norm": 2.1547327041625977, + "learning_rate": 5.442729488220959e-07, + "loss": 1.6354, + "step": 2485 + }, + { + "epoch": 2.724383561643836, + "grad_norm": 2.061166286468506, + "learning_rate": 5.422420796100732e-07, + "loss": 1.6859, + "step": 2486 + }, + { + "epoch": 2.7254794520547945, + "grad_norm": 2.2619354724884033, + "learning_rate": 5.402112103980505e-07, + "loss": 1.6525, + "step": 2487 + }, + { + "epoch": 2.7265753424657535, + "grad_norm": 2.2121572494506836, + "learning_rate": 5.381803411860276e-07, + "loss": 1.694, + "step": 2488 + }, + { + "epoch": 2.7276712328767125, + "grad_norm": 2.2862884998321533, + "learning_rate": 5.361494719740049e-07, + "loss": 1.7317, + "step": 2489 + }, + { + "epoch": 2.728767123287671, + "grad_norm": 2.166372537612915, + "learning_rate": 5.341186027619822e-07, + "loss": 1.6221, + "step": 2490 + }, + { + "epoch": 2.728767123287671, + "eval_loss": 1.9126579761505127, + "eval_runtime": 37.6516, + "eval_samples_per_second": 26.559, + "eval_steps_per_second": 1.115, + "step": 2490 + }, + { + "epoch": 2.72986301369863, + "grad_norm": 2.13952374458313, + "learning_rate": 5.320877335499594e-07, + "loss": 1.6439, + "step": 2491 + }, + { + "epoch": 2.730958904109589, + "grad_norm": 2.1364922523498535, + "learning_rate": 5.300568643379366e-07, + "loss": 1.7299, + "step": 2492 + }, + { + "epoch": 2.732054794520548, + "grad_norm": 2.166541576385498, + "learning_rate": 5.280259951259139e-07, + "loss": 1.6759, + "step": 2493 + }, + { + "epoch": 2.733150684931507, + "grad_norm": 2.1914117336273193, + "learning_rate": 5.259951259138912e-07, + "loss": 1.6331, + "step": 2494 + }, + { + "epoch": 2.734246575342466, + "grad_norm": 2.220278263092041, + "learning_rate": 5.239642567018685e-07, + "loss": 1.6359, + "step": 2495 + }, + { + "epoch": 2.7353424657534244, + "grad_norm": 2.1812047958374023, + "learning_rate": 5.219333874898457e-07, + "loss": 1.6596, + "step": 2496 + }, + { + "epoch": 2.7364383561643835, + "grad_norm": 2.1469738483428955, + "learning_rate": 5.199025182778229e-07, + "loss": 1.6744, + "step": 2497 + }, + { + "epoch": 2.7375342465753425, + "grad_norm": 2.1015233993530273, + "learning_rate": 5.178716490658002e-07, + "loss": 1.6338, + "step": 2498 + }, + { + "epoch": 2.7386301369863015, + "grad_norm": 2.124433994293213, + "learning_rate": 5.158407798537775e-07, + "loss": 1.5929, + "step": 2499 + }, + { + "epoch": 2.73972602739726, + "grad_norm": 2.203584671020508, + "learning_rate": 5.138099106417547e-07, + "loss": 1.5634, + "step": 2500 + }, + { + "epoch": 2.73972602739726, + "eval_loss": 1.913787841796875, + "eval_runtime": 37.6527, + "eval_samples_per_second": 26.559, + "eval_steps_per_second": 1.115, + "step": 2500 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 7.7932735234048e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..12d8f134d0ace768aa9a4fc9a1c7787b2ce68b40 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1fd745605ae8f150c3645104631b2d3c3b452eecda3ab1f66d0a3f474af022ff +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..7255d4b96382a76a01d62b0f3fb8ed7ba214a8cf --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cab5d049f3d3fc102f2df0c1651bc9d13107a3c0c772fd671f399b9d7d1de545 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..cf2513c6be393ffd7f18b9966efe1ea94961ce76 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e03f5804d1958af28b457d1c95eadeccb6d55a564daf236124a0bd65854c268 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..f9364c61dc2402c3ef00b25f3eebc99d4b5efc9c --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1d19a82202c388e47cd72603f375077a8199bb61e6280a7031db20863a17d02 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0c27d91076a4f01fc26da3f3eb57e643a5b17b77 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2700/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e484bef0de9dbc2a1854d7263015c073f363105ee8aeafa103cb8b7f04d74337 +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..d5b3eaa75e01e0564b8641c8f89a308c359a8a6c --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf3509a33c2e615b42fa452a51f3b448cd738037d76a4830d79b89c785555ed0 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..f31c91e6fc0dcc1eb5c101a4c40e35f5b315d26f --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c32df687c2bf9ce5fb1beafdcec2d9cfb49bc02e3a06d089279c235d40a82ca8 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..c1492af50e3afb56d1aecd7362d9e876ea1c01da --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fef8236d6b1951c0a5c3e8cf66c0be4ae93551b780648507e1981a0117386378 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..f112648d34d1e0b57734c6b897ca15aae7f8469c --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c944c83ac9404d75bf961215384ef1a49e6fa6754464d41468636bdbf52af4c9 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0c27d91076a4f01fc26da3f3eb57e643a5b17b77 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-400/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e484bef0de9dbc2a1854d7263015c073f363105ee8aeafa103cb8b7f04d74337 +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..10b2f5e4f6c9f32023941f5b8cf39645a86289fd --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ccffad8722b9db65bab96085aca5b7293a1ef8d534c7b8817b948837c87f6b8 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..89cef672a6b7edc3804190d221d1a77cd6a2d146 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c200394d27cbd1e2e417e51ec9d687fba67a2bda4d806d7a86762d3cf2c3f77 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..5bea917185d04a65ba4e81124131216312c82932 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18cc081206615fa75c71f8517f9b2a08419efc32934364e7e1fc0555a304cfe3 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-800/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_local10_10M_seed0/runs/checkpoint-2500/global_step2500/mp_rank_00_model_states.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_local10_10M_seed0/runs/checkpoint-2500/global_step2500/mp_rank_00_model_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..626180e8a3946dd2158d74b3ca72ad539f23af22 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_local10_10M_seed0/runs/checkpoint-2500/global_step2500/mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11f08aeb5b3e90449139bfc45ee9f2094b7bb0a8f9aa746235254357556bc4e6 +size 6425576195 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/latest new file mode 100644 index 0000000000000000000000000000000000000000..744ae7dbad571b6f37ec6c7066549494261bb59e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/latest @@ -0,0 +1 @@ +global_step100 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..2f56e4d7e7ec24cbd048e15163553d6bac03d9af --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/trainer_state.json @@ -0,0 +1,813 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.1095890410958904, + "eval_steps": 10, + "global_step": 100, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 3.11730940936192e+16, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/latest new file mode 100644 index 0000000000000000000000000000000000000000..22cd5c3402316b70299aed2025d7943595f5d495 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/latest @@ -0,0 +1 @@ +global_step1100 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..7eae876b4eacfdf0d00e862020df831ae17e457d --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/trainer_state.json @@ -0,0 +1,8613 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.2054794520547945, + "eval_steps": 10, + "global_step": 1100, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9794018268585205, + "learning_rate": 1.551094890510949e-06, + "loss": 2.3047, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.396454334259033, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.3671, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 2.0826210975646973, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.4105, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.3590970039367676, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.3405, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 2.3945844173431396, + "learning_rate": 1.624087591240876e-06, + "loss": 2.2922, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 2.185126781463623, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.3399, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.2687461376190186, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.3547, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.920499324798584, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.28, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.127774715423584, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.3409, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 1.8677122592926025, + "learning_rate": 1.715328467153285e-06, + "loss": 2.3406, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.4506173133850098, + "eval_runtime": 37.5008, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9803308248519897, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.3642, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 1.7733794450759888, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.3666, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 1.7639747858047485, + "learning_rate": 1.77007299270073e-06, + "loss": 2.3926, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 1.9172815084457397, + "learning_rate": 1.788321167883212e-06, + "loss": 2.3431, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 1.580621361732483, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.2967, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.0889124870300293, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.331, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.6659903526306152, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.3181, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.028651237487793, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.3237, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.681212306022644, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.32, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 1.9986696243286133, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.3425, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.4408469200134277, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 1.832582950592041, + "learning_rate": 1.916058394160584e-06, + "loss": 2.2548, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 1.8609144687652588, + "learning_rate": 1.934306569343066e-06, + "loss": 2.3024, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 1.8699817657470703, + "learning_rate": 1.952554744525548e-06, + "loss": 2.3044, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 1.81354558467865, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.3098, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 1.7789452075958252, + "learning_rate": 1.989051094890511e-06, + "loss": 2.3623, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 1.6322330236434937, + "learning_rate": 2.007299270072993e-06, + "loss": 2.303, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 1.8742916584014893, + "learning_rate": 2.025547445255475e-06, + "loss": 2.3183, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.7790557146072388, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.3168, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 1.7624462842941284, + "learning_rate": 2.062043795620438e-06, + "loss": 2.3629, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.781114101409912, + "learning_rate": 2.08029197080292e-06, + "loss": 2.2813, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.426335334777832, + "eval_runtime": 37.7146, + "eval_samples_per_second": 26.515, + "eval_steps_per_second": 1.114, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.769988775253296, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.3396, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.0640275478363037, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.3688, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.7884094715118408, + "learning_rate": 2.135036496350365e-06, + "loss": 2.2775, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.831847071647644, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.2966, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9044525623321533, + "learning_rate": 2.171532846715329e-06, + "loss": 2.3359, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.9899096488952637, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.3099, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8570183515548706, + "learning_rate": 2.208029197080292e-06, + "loss": 2.3016, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 1.8255733251571655, + "learning_rate": 2.226277372262774e-06, + "loss": 2.335, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.8759729862213135, + "learning_rate": 2.244525547445256e-06, + "loss": 2.3357, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.9031389951705933, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.3014, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.4213266372680664, + "eval_runtime": 37.6757, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.1044118404388428, + "learning_rate": 2.281021897810219e-06, + "loss": 2.2946, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.005735158920288, + "learning_rate": 2.299270072992701e-06, + "loss": 2.3132, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.8821312189102173, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.2816, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7866829633712769, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.309, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.8452801704406738, + "learning_rate": 2.354014598540146e-06, + "loss": 2.3222, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.5781638622283936, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.2688, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 1.9746729135513306, + "learning_rate": 2.39051094890511e-06, + "loss": 2.2864, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 2.0312447547912598, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.2914, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.675674557685852, + "learning_rate": 2.427007299270073e-06, + "loss": 2.3441, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 1.8987798690795898, + "learning_rate": 2.445255474452555e-06, + "loss": 2.2589, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.4149818420410156, + "eval_runtime": 37.6627, + "eval_samples_per_second": 26.551, + "eval_steps_per_second": 1.115, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 2.041752338409424, + "learning_rate": 2.463503649635037e-06, + "loss": 2.2457, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.7845569849014282, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.3376, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 2.489330530166626, + "learning_rate": 2.5e-06, + "loss": 2.258, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 2.3793296813964844, + "learning_rate": 2.518248175182482e-06, + "loss": 2.2561, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 2.0474460124969482, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.2992, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.7181849479675293, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.3003, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 2.3854947090148926, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.2866, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.7228457927703857, + "learning_rate": 2.591240875912409e-06, + "loss": 2.3027, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.3928253650665283, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.3233, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.136934757232666, + "learning_rate": 2.627737226277373e-06, + "loss": 2.2949, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.410449266433716, + "eval_runtime": 37.6837, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.8906159400939941, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.2693, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 1.8668863773345947, + "learning_rate": 2.664233576642336e-06, + "loss": 2.2944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.319363594055176, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.2791, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0869688987731934, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.32, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 1.8910208940505981, + "learning_rate": 2.718978102189781e-06, + "loss": 2.3059, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.157984495162964, + "learning_rate": 2.737226277372263e-06, + "loss": 2.2756, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.387314558029175, + "learning_rate": 2.755474452554745e-06, + "loss": 2.3597, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.183333158493042, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.322, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.158015489578247, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.2663, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 1.9594391584396362, + "learning_rate": 2.81021897810219e-06, + "loss": 2.3, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.411637306213379, + "eval_runtime": 37.6428, + "eval_samples_per_second": 26.565, + "eval_steps_per_second": 1.116, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.28957462310791, + "learning_rate": 2.828467153284672e-06, + "loss": 2.2258, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.8085683584213257, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.2739, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 1.9662787914276123, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.2398, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 2.337609052658081, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.3216, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 1.6008479595184326, + "learning_rate": 2.901459854014599e-06, + "loss": 2.2997, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 2.177690029144287, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.2893, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8746743202209473, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.3303, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 1.7862998247146606, + "learning_rate": 2.956204379562044e-06, + "loss": 2.3007, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.7578285932540894, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.2591, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.1532702445983887, + "learning_rate": 2.992700729927008e-06, + "loss": 2.2544, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.3992533683776855, + "eval_runtime": 37.6571, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.0168659687042236, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.3108, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 1.8336135149002075, + "learning_rate": 3.029197080291971e-06, + "loss": 2.3044, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.6675384044647217, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.2978, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 1.8866558074951172, + "learning_rate": 3.065693430656935e-06, + "loss": 2.3169, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0588760375976562, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.3041, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.9300280809402466, + "learning_rate": 3.102189781021898e-06, + "loss": 2.2494, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 1.837496042251587, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.2339, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.155080795288086, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.324, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 1.9562510251998901, + "learning_rate": 3.156934306569343e-06, + "loss": 2.2937, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 2.522388219833374, + "learning_rate": 3.175182481751825e-06, + "loss": 2.2683, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.3936216831207275, + "eval_runtime": 37.6499, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 1.7409396171569824, + "learning_rate": 3.193430656934307e-06, + "loss": 2.2833, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.176687002182007, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.2748, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 1.8576542139053345, + "learning_rate": 3.2299270072992707e-06, + "loss": 2.2144, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.9564870595932007, + "learning_rate": 3.248175182481752e-06, + "loss": 2.2898, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.896238088607788, + "learning_rate": 3.266423357664234e-06, + "loss": 2.2793, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.735960841178894, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.2869, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.018303394317627, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.2775, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 1.7448145151138306, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.2736, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8245689868927002, + "learning_rate": 3.339416058394161e-06, + "loss": 2.2799, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 1.821118950843811, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.2823, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.3893656730651855, + "eval_runtime": 37.676, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 1.9112446308135986, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.287, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 1.6978867053985596, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.2865, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9376300573349, + "learning_rate": 3.412408759124088e-06, + "loss": 2.2326, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.8067563772201538, + "learning_rate": 3.43065693430657e-06, + "loss": 2.2739, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.8476370573043823, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.2328, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.927923321723938, + "learning_rate": 3.467153284671533e-06, + "loss": 2.2679, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7956453561782837, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.2607, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 1.857413649559021, + "learning_rate": 3.503649635036497e-06, + "loss": 2.2854, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 1.8384093046188354, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.3021, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.7346371412277222, + "learning_rate": 3.54014598540146e-06, + "loss": 2.3402, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.390045642852783, + "eval_runtime": 37.4358, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.7546677589416504, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.3214, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 1.7437881231307983, + "learning_rate": 3.576642335766424e-06, + "loss": 2.2004, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9509227275848389, + "learning_rate": 3.5948905109489054e-06, + "loss": 2.2044, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 1.6918139457702637, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.2369, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9664885997772217, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.254, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 1.7942848205566406, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.2296, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 1.8610197305679321, + "learning_rate": 3.667883211678833e-06, + "loss": 2.2273, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 1.8405956029891968, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.3034, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 1.7151877880096436, + "learning_rate": 3.704379562043796e-06, + "loss": 2.2689, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 1.8795795440673828, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.2363, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.384481430053711, + "eval_runtime": 37.6512, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 1.7494690418243408, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.257, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 1.7700639963150024, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.2457, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 1.846092700958252, + "learning_rate": 3.777372262773723e-06, + "loss": 2.2425, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 1.6699751615524292, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.2734, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 1.8978508710861206, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.2435, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.6387466192245483, + "learning_rate": 3.832116788321168e-06, + "loss": 2.2241, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 1.6999541521072388, + "learning_rate": 3.85036496350365e-06, + "loss": 2.2468, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 1.6322710514068604, + "learning_rate": 3.868613138686132e-06, + "loss": 2.2642, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.5633572340011597, + "learning_rate": 3.886861313868613e-06, + "loss": 2.2862, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 1.7642648220062256, + "learning_rate": 3.905109489051096e-06, + "loss": 2.2414, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.3775289058685303, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.76724374294281, + "learning_rate": 3.923357664233577e-06, + "loss": 2.2976, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.6049190759658813, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.3034, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.6591981649398804, + "learning_rate": 3.95985401459854e-06, + "loss": 2.1957, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.650676965713501, + "learning_rate": 3.978102189781022e-06, + "loss": 2.275, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.6665960550308228, + "learning_rate": 3.996350364963504e-06, + "loss": 2.1853, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.5785901546478271, + "learning_rate": 4.014598540145986e-06, + "loss": 2.2399, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.976162075996399, + "learning_rate": 4.032846715328467e-06, + "loss": 2.2499, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.642327904701233, + "learning_rate": 4.05109489051095e-06, + "loss": 2.235, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 2.0126261711120605, + "learning_rate": 4.069343065693431e-06, + "loss": 2.2466, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.6737161874771118, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.2984, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.375901222229004, + "eval_runtime": 37.7072, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 2.094257354736328, + "learning_rate": 4.105839416058395e-06, + "loss": 2.251, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.655393123626709, + "learning_rate": 4.124087591240876e-06, + "loss": 2.2366, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 2.5909082889556885, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.2986, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.8278532028198242, + "learning_rate": 4.16058394160584e-06, + "loss": 2.1977, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9404895305633545, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.2578, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.6525582075119019, + "learning_rate": 4.197080291970803e-06, + "loss": 2.2263, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 2.021822929382324, + "learning_rate": 4.215328467153285e-06, + "loss": 2.2326, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 1.9136033058166504, + "learning_rate": 4.233576642335767e-06, + "loss": 2.2457, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.389714002609253, + "learning_rate": 4.251824817518249e-06, + "loss": 2.2256, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.7886695861816406, + "learning_rate": 4.27007299270073e-06, + "loss": 2.2222, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.3729844093322754, + "eval_runtime": 37.7638, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.9499074220657349, + "learning_rate": 4.288321167883212e-06, + "loss": 2.2026, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 1.834797978401184, + "learning_rate": 4.306569343065693e-06, + "loss": 2.2564, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 2.0415048599243164, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.2462, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 1.6527886390686035, + "learning_rate": 4.343065693430658e-06, + "loss": 2.2493, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.0504889488220215, + "learning_rate": 4.361313868613139e-06, + "loss": 2.2734, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.004772424697876, + "learning_rate": 4.379562043795621e-06, + "loss": 2.2469, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.1973347663879395, + "learning_rate": 4.397810218978102e-06, + "loss": 2.2479, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.8222625255584717, + "learning_rate": 4.416058394160584e-06, + "loss": 2.2206, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 2.1759912967681885, + "learning_rate": 4.434306569343066e-06, + "loss": 2.261, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.8344624042510986, + "learning_rate": 4.452554744525548e-06, + "loss": 2.2605, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.3706893920898438, + "eval_runtime": 37.734, + "eval_samples_per_second": 26.501, + "eval_steps_per_second": 1.113, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.7783704996109009, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.284, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 2.101193428039551, + "learning_rate": 4.489051094890512e-06, + "loss": 2.2524, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 1.7923569679260254, + "learning_rate": 4.507299270072993e-06, + "loss": 2.2434, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 2.170883893966675, + "learning_rate": 4.525547445255475e-06, + "loss": 2.2173, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.6749192476272583, + "learning_rate": 4.543795620437956e-06, + "loss": 2.283, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.0484230518341064, + "learning_rate": 4.562043795620438e-06, + "loss": 2.2931, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.9212749004364014, + "learning_rate": 4.580291970802921e-06, + "loss": 2.2661, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.6729028224945068, + "learning_rate": 4.598540145985402e-06, + "loss": 2.2269, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 2.1190619468688965, + "learning_rate": 4.6167883211678835e-06, + "loss": 2.2278, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.7444697618484497, + "learning_rate": 4.635036496350365e-06, + "loss": 2.2634, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.3706092834472656, + "eval_runtime": 37.7391, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 1.8921526670455933, + "learning_rate": 4.653284671532847e-06, + "loss": 2.2697, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7347339391708374, + "learning_rate": 4.671532846715329e-06, + "loss": 2.2406, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 2.1617953777313232, + "learning_rate": 4.689781021897811e-06, + "loss": 2.2891, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.873384714126587, + "learning_rate": 4.708029197080292e-06, + "loss": 2.2554, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.801356554031372, + "learning_rate": 4.726277372262774e-06, + "loss": 2.2073, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.9250633716583252, + "learning_rate": 4.744525547445255e-06, + "loss": 2.2501, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 1.8360860347747803, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.2561, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 1.8711498975753784, + "learning_rate": 4.78102189781022e-06, + "loss": 2.2247, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.6604810953140259, + "learning_rate": 4.799270072992701e-06, + "loss": 2.2448, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 1.689201831817627, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.2494, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.3682427406311035, + "eval_runtime": 37.7233, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.720961093902588, + "learning_rate": 4.835766423357665e-06, + "loss": 2.2013, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.7863153219223022, + "learning_rate": 4.854014598540146e-06, + "loss": 2.2183, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 1.8583060503005981, + "learning_rate": 4.872262773722628e-06, + "loss": 2.2013, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 1.722428560256958, + "learning_rate": 4.89051094890511e-06, + "loss": 2.2399, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.8408993482589722, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.2489, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.887604832649231, + "learning_rate": 4.927007299270074e-06, + "loss": 2.2145, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 1.7112189531326294, + "learning_rate": 4.945255474452555e-06, + "loss": 2.2915, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 1.8314064741134644, + "learning_rate": 4.963503649635037e-06, + "loss": 2.2786, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.7001186609268188, + "learning_rate": 4.981751824817518e-06, + "loss": 2.249, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.5820505619049072, + "learning_rate": 5e-06, + "loss": 2.254, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.3624653816223145, + "eval_runtime": 37.7283, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 1.8516303300857544, + "learning_rate": 4.997969130787978e-06, + "loss": 2.1894, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 1.8591808080673218, + "learning_rate": 4.995938261575955e-06, + "loss": 2.2537, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.6822415590286255, + "learning_rate": 4.993907392363932e-06, + "loss": 2.1978, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 2.0779759883880615, + "learning_rate": 4.9918765231519095e-06, + "loss": 2.2286, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.6937508583068848, + "learning_rate": 4.989845653939887e-06, + "loss": 2.2339, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.9342437982559204, + "learning_rate": 4.987814784727864e-06, + "loss": 2.2145, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.773258924484253, + "learning_rate": 4.985783915515841e-06, + "loss": 2.2368, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.8196488618850708, + "learning_rate": 4.983753046303819e-06, + "loss": 2.2224, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.6786775588989258, + "learning_rate": 4.981722177091796e-06, + "loss": 2.2439, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.7761427164077759, + "learning_rate": 4.979691307879773e-06, + "loss": 2.236, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.3614282608032227, + "eval_runtime": 37.7314, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.8389159440994263, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.2745, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.7774744033813477, + "learning_rate": 4.975629569455728e-06, + "loss": 2.2122, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.7862017154693604, + "learning_rate": 4.973598700243705e-06, + "loss": 2.2097, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.8851722478866577, + "learning_rate": 4.971567831031681e-06, + "loss": 2.118, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.068608045578003, + "learning_rate": 4.9695369618196595e-06, + "loss": 2.1884, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.8885847330093384, + "learning_rate": 4.967506092607636e-06, + "loss": 2.2097, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.9441415071487427, + "learning_rate": 4.965475223395614e-06, + "loss": 2.2068, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 2.313577890396118, + "learning_rate": 4.9634443541835905e-06, + "loss": 2.2355, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 1.8542472124099731, + "learning_rate": 4.961413484971569e-06, + "loss": 2.2299, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9382838010787964, + "learning_rate": 4.959382615759545e-06, + "loss": 2.2072, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.361767053604126, + "eval_runtime": 37.4603, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.081242322921753, + "learning_rate": 4.957351746547523e-06, + "loss": 2.3167, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 1.9109065532684326, + "learning_rate": 4.9553208773355e-06, + "loss": 2.2336, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.124897003173828, + "learning_rate": 4.953290008123478e-06, + "loss": 2.184, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.8820433616638184, + "learning_rate": 4.951259138911454e-06, + "loss": 2.2836, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.016204595565796, + "learning_rate": 4.9492282696994315e-06, + "loss": 2.2221, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.0679776668548584, + "learning_rate": 4.947197400487409e-06, + "loss": 2.2113, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 1.8040891885757446, + "learning_rate": 4.945166531275386e-06, + "loss": 2.2292, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.0315871238708496, + "learning_rate": 4.943135662063363e-06, + "loss": 2.2657, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.71120023727417, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.2517, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0443193912506104, + "learning_rate": 4.939073923639318e-06, + "loss": 2.2105, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.3582265377044678, + "eval_runtime": 37.6749, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.9363837242126465, + "learning_rate": 4.937043054427295e-06, + "loss": 2.1944, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 2.0878167152404785, + "learning_rate": 4.935012185215272e-06, + "loss": 2.2477, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.9059010744094849, + "learning_rate": 4.93298131600325e-06, + "loss": 2.2278, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 2.220851421356201, + "learning_rate": 4.930950446791227e-06, + "loss": 2.2391, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.7555570602416992, + "learning_rate": 4.928919577579204e-06, + "loss": 2.2206, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 2.1546502113342285, + "learning_rate": 4.9268887083671815e-06, + "loss": 2.2216, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 1.9688911437988281, + "learning_rate": 4.924857839155159e-06, + "loss": 2.2423, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 1.983587622642517, + "learning_rate": 4.922826969943136e-06, + "loss": 2.1738, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 2.1990034580230713, + "learning_rate": 4.920796100731113e-06, + "loss": 2.2341, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 1.9690128564834595, + "learning_rate": 4.918765231519091e-06, + "loss": 2.1914, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.36171817779541, + "eval_runtime": 37.727, + "eval_samples_per_second": 26.506, + "eval_steps_per_second": 1.113, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 2.5950212478637695, + "learning_rate": 4.916734362307068e-06, + "loss": 2.1921, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 2.023130416870117, + "learning_rate": 4.914703493095045e-06, + "loss": 2.1932, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2486252784729004, + "learning_rate": 4.912672623883022e-06, + "loss": 2.2174, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 1.723677635192871, + "learning_rate": 4.910641754671e-06, + "loss": 2.2173, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.94509756565094, + "learning_rate": 4.908610885458977e-06, + "loss": 2.2539, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 1.6934322118759155, + "learning_rate": 4.906580016246954e-06, + "loss": 2.2345, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 1.6882652044296265, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.231, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.6730620861053467, + "learning_rate": 4.902518277822909e-06, + "loss": 2.2394, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.7261624336242676, + "learning_rate": 4.900487408610886e-06, + "loss": 2.2524, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.8310920000076294, + "learning_rate": 4.898456539398863e-06, + "loss": 2.3004, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.3536295890808105, + "eval_runtime": 37.7461, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.7025809288024902, + "learning_rate": 4.89642567018684e-06, + "loss": 2.1923, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.7285082340240479, + "learning_rate": 4.894394800974818e-06, + "loss": 2.241, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 1.7589807510375977, + "learning_rate": 4.892363931762794e-06, + "loss": 2.2019, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.6664931774139404, + "learning_rate": 4.8903330625507725e-06, + "loss": 2.1822, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 1.8308027982711792, + "learning_rate": 4.888302193338749e-06, + "loss": 2.2651, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.6426894664764404, + "learning_rate": 4.886271324126727e-06, + "loss": 2.2514, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.9765205383300781, + "learning_rate": 4.8842404549147034e-06, + "loss": 2.226, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.7872203588485718, + "learning_rate": 4.8822095857026816e-06, + "loss": 2.18, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 2.0645813941955566, + "learning_rate": 4.880178716490658e-06, + "loss": 2.2023, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 1.9823166131973267, + "learning_rate": 4.878147847278636e-06, + "loss": 2.1801, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.353776454925537, + "eval_runtime": 37.771, + "eval_samples_per_second": 26.475, + "eval_steps_per_second": 1.112, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.831152319908142, + "learning_rate": 4.8761169780666125e-06, + "loss": 2.1858, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 1.9321790933609009, + "learning_rate": 4.874086108854591e-06, + "loss": 2.1638, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.9097596406936646, + "learning_rate": 4.872055239642567e-06, + "loss": 2.2206, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8821815252304077, + "learning_rate": 4.870024370430544e-06, + "loss": 2.2338, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.7959749698638916, + "learning_rate": 4.867993501218522e-06, + "loss": 2.1882, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 1.625320315361023, + "learning_rate": 4.865962632006499e-06, + "loss": 2.2137, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 1.6914536952972412, + "learning_rate": 4.863931762794476e-06, + "loss": 2.2374, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0801584720611572, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.2517, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 1.6874804496765137, + "learning_rate": 4.859870024370431e-06, + "loss": 2.2621, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 1.9347832202911377, + "learning_rate": 4.857839155158408e-06, + "loss": 2.2326, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.3468236923217773, + "eval_runtime": 37.7184, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.6798334121704102, + "learning_rate": 4.855808285946385e-06, + "loss": 2.2222, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 1.8993884325027466, + "learning_rate": 4.8537774167343626e-06, + "loss": 2.2178, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 1.697892189025879, + "learning_rate": 4.85174654752234e-06, + "loss": 2.2217, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8484156131744385, + "learning_rate": 4.849715678310317e-06, + "loss": 2.2069, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 1.8125075101852417, + "learning_rate": 4.847684809098294e-06, + "loss": 2.2285, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 1.8215030431747437, + "learning_rate": 4.845653939886272e-06, + "loss": 2.2029, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 1.777129054069519, + "learning_rate": 4.843623070674249e-06, + "loss": 2.2512, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.7327351570129395, + "learning_rate": 4.841592201462226e-06, + "loss": 2.1767, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.7363202571868896, + "learning_rate": 4.8395613322502035e-06, + "loss": 2.2174, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9183942079544067, + "learning_rate": 4.837530463038181e-06, + "loss": 2.2383, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.345961332321167, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.7290767431259155, + "learning_rate": 4.835499593826158e-06, + "loss": 2.22, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 1.7892694473266602, + "learning_rate": 4.833468724614135e-06, + "loss": 2.1403, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.6671900749206543, + "learning_rate": 4.831437855402113e-06, + "loss": 2.1989, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 1.848751425743103, + "learning_rate": 4.82940698619009e-06, + "loss": 2.245, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.6957710981369019, + "learning_rate": 4.827376116978067e-06, + "loss": 2.2239, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.650583267211914, + "learning_rate": 4.8253452477660444e-06, + "loss": 2.2101, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.6510974168777466, + "learning_rate": 4.823314378554022e-06, + "loss": 2.204, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 1.9469026327133179, + "learning_rate": 4.821283509341998e-06, + "loss": 2.221, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 1.6031194925308228, + "learning_rate": 4.819252640129976e-06, + "loss": 2.1926, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.120325803756714, + "learning_rate": 4.817221770917953e-06, + "loss": 2.2166, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.3419017791748047, + "eval_runtime": 37.6854, + "eval_samples_per_second": 26.535, + "eval_steps_per_second": 1.114, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.5533722639083862, + "learning_rate": 4.815190901705931e-06, + "loss": 2.1981, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9650007486343384, + "learning_rate": 4.813160032493907e-06, + "loss": 2.2319, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 1.6874849796295166, + "learning_rate": 4.811129163281885e-06, + "loss": 2.2263, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.787996768951416, + "learning_rate": 4.809098294069862e-06, + "loss": 2.2546, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 1.7853440046310425, + "learning_rate": 4.80706742485784e-06, + "loss": 2.1994, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.7882641553878784, + "learning_rate": 4.805036555645816e-06, + "loss": 2.2449, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 1.8206003904342651, + "learning_rate": 4.8030056864337945e-06, + "loss": 2.2144, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.7012749910354614, + "learning_rate": 4.800974817221771e-06, + "loss": 2.2122, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 1.8711670637130737, + "learning_rate": 4.798943948009749e-06, + "loss": 2.2132, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 1.6880611181259155, + "learning_rate": 4.7969130787977254e-06, + "loss": 2.2336, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.340393543243408, + "eval_runtime": 37.6885, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 2.0543065071105957, + "learning_rate": 4.794882209585703e-06, + "loss": 2.1899, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 1.5774480104446411, + "learning_rate": 4.79285134037368e-06, + "loss": 2.1982, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 1.7154620885849, + "learning_rate": 4.790820471161657e-06, + "loss": 2.2096, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.8200151920318604, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.2391, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 1.6368298530578613, + "learning_rate": 4.786758732737612e-06, + "loss": 2.1855, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 1.6851340532302856, + "learning_rate": 4.784727863525589e-06, + "loss": 2.2114, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 1.8293818235397339, + "learning_rate": 4.782696994313566e-06, + "loss": 2.1636, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 1.6213513612747192, + "learning_rate": 4.780666125101544e-06, + "loss": 2.1884, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.7791825532913208, + "learning_rate": 4.778635255889521e-06, + "loss": 2.1977, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 1.6627994775772095, + "learning_rate": 4.776604386677498e-06, + "loss": 2.2444, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.3396637439727783, + "eval_runtime": 37.7069, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 1.7597277164459229, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.2369, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.7754322290420532, + "learning_rate": 4.772542648253453e-06, + "loss": 2.1859, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 1.7897601127624512, + "learning_rate": 4.77051177904143e-06, + "loss": 2.2245, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 1.7294282913208008, + "learning_rate": 4.768480909829407e-06, + "loss": 2.1998, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 1.8406974077224731, + "learning_rate": 4.766450040617385e-06, + "loss": 2.2353, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 1.849493384361267, + "learning_rate": 4.764419171405362e-06, + "loss": 2.2165, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.5568877458572388, + "learning_rate": 4.762388302193339e-06, + "loss": 2.1989, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 1.8816689252853394, + "learning_rate": 4.760357432981316e-06, + "loss": 2.2062, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 1.7648332118988037, + "learning_rate": 4.758326563769294e-06, + "loss": 2.2204, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.753944754600525, + "learning_rate": 4.756295694557271e-06, + "loss": 2.1748, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.339653253555298, + "eval_runtime": 37.4438, + "eval_samples_per_second": 26.707, + "eval_steps_per_second": 1.122, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.7343329191207886, + "learning_rate": 4.754264825345248e-06, + "loss": 2.1843, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 1.8125407695770264, + "learning_rate": 4.7522339561332255e-06, + "loss": 2.1492, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.5255184173583984, + "learning_rate": 4.750203086921203e-06, + "loss": 2.2178, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 1.6692545413970947, + "learning_rate": 4.74817221770918e-06, + "loss": 2.2516, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 1.7804819345474243, + "learning_rate": 4.7461413484971565e-06, + "loss": 2.1592, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8048933744430542, + "learning_rate": 4.744110479285135e-06, + "loss": 2.1869, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 1.6439207792282104, + "learning_rate": 4.742079610073111e-06, + "loss": 2.1995, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.7375173568725586, + "learning_rate": 4.740048740861089e-06, + "loss": 2.2067, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 1.9471659660339355, + "learning_rate": 4.738017871649066e-06, + "loss": 2.2513, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 1.6524564027786255, + "learning_rate": 4.735987002437044e-06, + "loss": 2.2643, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.337153434753418, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 1.8574563264846802, + "learning_rate": 4.73395613322502e-06, + "loss": 2.2373, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 1.836464524269104, + "learning_rate": 4.731925264012998e-06, + "loss": 2.1769, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 1.7012423276901245, + "learning_rate": 4.729894394800975e-06, + "loss": 2.1494, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 1.953180193901062, + "learning_rate": 4.727863525588953e-06, + "loss": 2.186, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 1.8468470573425293, + "learning_rate": 4.725832656376929e-06, + "loss": 2.2279, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 2.005077362060547, + "learning_rate": 4.723801787164907e-06, + "loss": 2.1798, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 1.8246952295303345, + "learning_rate": 4.721770917952885e-06, + "loss": 2.1916, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 1.8558839559555054, + "learning_rate": 4.719740048740861e-06, + "loss": 2.2237, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 1.7670999765396118, + "learning_rate": 4.717709179528839e-06, + "loss": 2.2179, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.8544219732284546, + "learning_rate": 4.715678310316816e-06, + "loss": 2.1811, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.333791732788086, + "eval_runtime": 37.7254, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 1.5792925357818604, + "learning_rate": 4.713647441104794e-06, + "loss": 2.1749, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 1.5678222179412842, + "learning_rate": 4.71161657189277e-06, + "loss": 2.2294, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 1.9467812776565552, + "learning_rate": 4.709585702680748e-06, + "loss": 2.2684, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 1.6647964715957642, + "learning_rate": 4.707554833468725e-06, + "loss": 2.1722, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 1.7037622928619385, + "learning_rate": 4.705523964256703e-06, + "loss": 2.1915, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.5857148170471191, + "learning_rate": 4.703493095044679e-06, + "loss": 2.1466, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 1.5792173147201538, + "learning_rate": 4.701462225832657e-06, + "loss": 2.2472, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 1.5998480319976807, + "learning_rate": 4.699431356620634e-06, + "loss": 2.2129, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.6728194952011108, + "learning_rate": 4.697400487408612e-06, + "loss": 2.2057, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 1.5402852296829224, + "learning_rate": 4.695369618196588e-06, + "loss": 2.1903, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.3325934410095215, + "eval_runtime": 37.7463, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 1.5890164375305176, + "learning_rate": 4.693338748984566e-06, + "loss": 2.1584, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 1.805923581123352, + "learning_rate": 4.691307879772543e-06, + "loss": 2.2263, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.5932700634002686, + "learning_rate": 4.68927701056052e-06, + "loss": 2.2109, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 1.5611685514450073, + "learning_rate": 4.6872461413484975e-06, + "loss": 2.192, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 1.6976194381713867, + "learning_rate": 4.685215272136475e-06, + "loss": 2.1998, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.7149630784988403, + "learning_rate": 4.683184402924452e-06, + "loss": 2.2599, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 1.7597423791885376, + "learning_rate": 4.681153533712429e-06, + "loss": 2.1765, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 1.539652943611145, + "learning_rate": 4.679122664500407e-06, + "loss": 2.1635, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.690402865409851, + "learning_rate": 4.677091795288384e-06, + "loss": 2.2417, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.7267673015594482, + "learning_rate": 4.675060926076361e-06, + "loss": 2.166, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.331728935241699, + "eval_runtime": 37.7133, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.6383073329925537, + "learning_rate": 4.673030056864338e-06, + "loss": 2.1624, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.5364140272140503, + "learning_rate": 4.670999187652316e-06, + "loss": 2.1696, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.591477870941162, + "learning_rate": 4.668968318440293e-06, + "loss": 2.2306, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 1.8121449947357178, + "learning_rate": 4.66693744922827e-06, + "loss": 2.1449, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.515539288520813, + "learning_rate": 4.6649065800162475e-06, + "loss": 2.2108, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 1.6521453857421875, + "learning_rate": 4.662875710804225e-06, + "loss": 2.1398, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 1.7074251174926758, + "learning_rate": 4.660844841592202e-06, + "loss": 2.2342, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.7688119411468506, + "learning_rate": 4.658813972380179e-06, + "loss": 2.1287, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.7170679569244385, + "learning_rate": 4.656783103168157e-06, + "loss": 2.2143, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 1.7401350736618042, + "learning_rate": 4.654752233956134e-06, + "loss": 2.1855, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.335416555404663, + "eval_runtime": 37.7115, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.6344956159591675, + "learning_rate": 4.652721364744111e-06, + "loss": 2.2205, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.629184603691101, + "learning_rate": 4.6506904955320885e-06, + "loss": 2.2271, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.769491195678711, + "learning_rate": 4.648659626320066e-06, + "loss": 2.1945, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.6513251066207886, + "learning_rate": 4.646628757108043e-06, + "loss": 2.1705, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9357609748840332, + "learning_rate": 4.6445978878960194e-06, + "loss": 2.1968, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 1.75972318649292, + "learning_rate": 4.6425670186839976e-06, + "loss": 2.2229, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 1.686574935913086, + "learning_rate": 4.640536149471974e-06, + "loss": 2.2361, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 1.655661702156067, + "learning_rate": 4.638505280259952e-06, + "loss": 2.2064, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.7583094835281372, + "learning_rate": 4.6364744110479285e-06, + "loss": 2.2352, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.5906383991241455, + "learning_rate": 4.634443541835907e-06, + "loss": 2.187, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.332923412322998, + "eval_runtime": 37.7496, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 1.708019733428955, + "learning_rate": 4.632412672623883e-06, + "loss": 2.1032, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 1.6350252628326416, + "learning_rate": 4.630381803411861e-06, + "loss": 2.1976, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.9037389755249023, + "learning_rate": 4.628350934199838e-06, + "loss": 2.224, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 1.7126816511154175, + "learning_rate": 4.626320064987816e-06, + "loss": 2.1925, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 1.5828406810760498, + "learning_rate": 4.624289195775792e-06, + "loss": 2.1928, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.6797102689743042, + "learning_rate": 4.62225832656377e-06, + "loss": 2.1902, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 1.6871609687805176, + "learning_rate": 4.620227457351747e-06, + "loss": 2.2024, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.8785804510116577, + "learning_rate": 4.618196588139724e-06, + "loss": 2.1778, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.6175488233566284, + "learning_rate": 4.616165718927701e-06, + "loss": 2.2438, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.7724019289016724, + "learning_rate": 4.6141348497156786e-06, + "loss": 2.2272, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.328923225402832, + "eval_runtime": 37.7443, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.7230926752090454, + "learning_rate": 4.612103980503656e-06, + "loss": 2.2021, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.7899917364120483, + "learning_rate": 4.610073111291633e-06, + "loss": 2.1854, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.5864417552947998, + "learning_rate": 4.60804224207961e-06, + "loss": 2.2042, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.875375509262085, + "learning_rate": 4.606011372867588e-06, + "loss": 2.1844, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 1.830597162246704, + "learning_rate": 4.603980503655565e-06, + "loss": 2.1474, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 1.8624461889266968, + "learning_rate": 4.601949634443542e-06, + "loss": 2.158, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.7161756753921509, + "learning_rate": 4.5999187652315195e-06, + "loss": 2.1838, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.6212890148162842, + "learning_rate": 4.597887896019497e-06, + "loss": 2.1911, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.8586416244506836, + "learning_rate": 4.595857026807474e-06, + "loss": 2.1794, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.653976559638977, + "learning_rate": 4.593826157595451e-06, + "loss": 2.2428, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.3279106616973877, + "eval_runtime": 37.739, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 2.054537057876587, + "learning_rate": 4.591795288383429e-06, + "loss": 2.2003, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.730942964553833, + "learning_rate": 4.589764419171406e-06, + "loss": 2.1816, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.7264728546142578, + "learning_rate": 4.587733549959383e-06, + "loss": 2.1615, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 1.8276281356811523, + "learning_rate": 4.5857026807473604e-06, + "loss": 2.1442, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 1.9947556257247925, + "learning_rate": 4.583671811535338e-06, + "loss": 2.2192, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 1.770601511001587, + "learning_rate": 4.581640942323315e-06, + "loss": 2.233, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 1.8946491479873657, + "learning_rate": 4.579610073111292e-06, + "loss": 2.2249, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 1.7839634418487549, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.2469, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.6298807859420776, + "learning_rate": 4.575548334687247e-06, + "loss": 2.1723, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 1.9310041666030884, + "learning_rate": 4.573517465475224e-06, + "loss": 2.2156, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.3249237537384033, + "eval_runtime": 37.7505, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 1.7901647090911865, + "learning_rate": 4.571486596263201e-06, + "loss": 2.2169, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.7739574909210205, + "learning_rate": 4.569455727051178e-06, + "loss": 2.2005, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 1.7310121059417725, + "learning_rate": 4.567424857839156e-06, + "loss": 2.2369, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.004469156265259, + "learning_rate": 4.565393988627132e-06, + "loss": 2.2051, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 1.7906666994094849, + "learning_rate": 4.5633631194151105e-06, + "loss": 2.1908, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 1.7224774360656738, + "learning_rate": 4.561332250203087e-06, + "loss": 2.1529, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 1.9950156211853027, + "learning_rate": 4.559301380991065e-06, + "loss": 2.1776, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 1.8257590532302856, + "learning_rate": 4.5572705117790414e-06, + "loss": 2.2267, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.038490056991577, + "learning_rate": 4.5552396425670196e-06, + "loss": 2.1703, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 1.7314448356628418, + "learning_rate": 4.553208773354996e-06, + "loss": 2.2409, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.327078104019165, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 2.0013606548309326, + "learning_rate": 4.551177904142974e-06, + "loss": 2.1732, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.7352690696716309, + "learning_rate": 4.5491470349309505e-06, + "loss": 2.2027, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9024865627288818, + "learning_rate": 4.547116165718929e-06, + "loss": 2.1806, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8870717287063599, + "learning_rate": 4.545085296506905e-06, + "loss": 2.1664, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.781226634979248, + "learning_rate": 4.543054427294882e-06, + "loss": 2.166, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.8994221687316895, + "learning_rate": 4.54102355808286e-06, + "loss": 2.1951, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.6999517679214478, + "learning_rate": 4.538992688870837e-06, + "loss": 2.1786, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.899590253829956, + "learning_rate": 4.536961819658814e-06, + "loss": 2.2051, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.8054970502853394, + "learning_rate": 4.5349309504467915e-06, + "loss": 2.1675, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.6732861995697021, + "learning_rate": 4.532900081234769e-06, + "loss": 2.1997, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.3242335319519043, + "eval_runtime": 37.6748, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 2.0063157081604004, + "learning_rate": 4.530869212022746e-06, + "loss": 2.1909, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.7750036716461182, + "learning_rate": 4.528838342810723e-06, + "loss": 2.1935, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9993029832839966, + "learning_rate": 4.526807473598701e-06, + "loss": 2.1753, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.6742994785308838, + "learning_rate": 4.524776604386678e-06, + "loss": 2.2205, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 1.9300428628921509, + "learning_rate": 4.522745735174655e-06, + "loss": 2.2081, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.8770017623901367, + "learning_rate": 4.520714865962632e-06, + "loss": 2.2074, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.847765326499939, + "learning_rate": 4.51868399675061e-06, + "loss": 2.1775, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.905117392539978, + "learning_rate": 4.516653127538587e-06, + "loss": 2.1511, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.7623670101165771, + "learning_rate": 4.514622258326564e-06, + "loss": 2.188, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.9232176542282104, + "learning_rate": 4.5125913891145415e-06, + "loss": 2.2109, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.3252475261688232, + "eval_runtime": 37.6923, + "eval_samples_per_second": 26.531, + "eval_steps_per_second": 1.114, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.6523535251617432, + "learning_rate": 4.510560519902519e-06, + "loss": 2.1499, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9934184551239014, + "learning_rate": 4.508529650690496e-06, + "loss": 2.1383, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.704522728919983, + "learning_rate": 4.506498781478473e-06, + "loss": 2.1892, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.9909989833831787, + "learning_rate": 4.504467912266451e-06, + "loss": 2.2422, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.7490800619125366, + "learning_rate": 4.502437043054428e-06, + "loss": 2.2008, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8620145320892334, + "learning_rate": 4.500406173842405e-06, + "loss": 2.1892, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.7069931030273438, + "learning_rate": 4.4983753046303824e-06, + "loss": 2.1816, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 2.152491807937622, + "learning_rate": 4.49634443541836e-06, + "loss": 2.2331, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.8526151180267334, + "learning_rate": 4.494313566206336e-06, + "loss": 2.1811, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.6865259408950806, + "learning_rate": 4.492282696994314e-06, + "loss": 2.1599, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.3265392780303955, + "eval_runtime": 37.7255, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.7710354328155518, + "learning_rate": 4.490251827782291e-06, + "loss": 2.1636, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.8444210290908813, + "learning_rate": 4.488220958570269e-06, + "loss": 2.1345, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.9095014333724976, + "learning_rate": 4.486190089358245e-06, + "loss": 2.2236, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.7406108379364014, + "learning_rate": 4.484159220146223e-06, + "loss": 2.2078, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.712534785270691, + "learning_rate": 4.4821283509342e-06, + "loss": 2.1809, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8921459913253784, + "learning_rate": 4.480097481722178e-06, + "loss": 2.158, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.805146336555481, + "learning_rate": 4.478066612510154e-06, + "loss": 2.1765, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 2.1620516777038574, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.2832, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.661618947982788, + "learning_rate": 4.474004874086109e-06, + "loss": 2.1397, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.732483983039856, + "learning_rate": 4.471974004874087e-06, + "loss": 2.225, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.3276846408843994, + "eval_runtime": 37.7156, + "eval_samples_per_second": 26.514, + "eval_steps_per_second": 1.114, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.8693495988845825, + "learning_rate": 4.4699431356620635e-06, + "loss": 2.1959, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8201690912246704, + "learning_rate": 4.467912266450041e-06, + "loss": 2.1882, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.9229968786239624, + "learning_rate": 4.465881397238018e-06, + "loss": 2.2006, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.6330770254135132, + "learning_rate": 4.463850528025995e-06, + "loss": 2.209, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9654744863510132, + "learning_rate": 4.4618196588139726e-06, + "loss": 2.1963, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.8808507919311523, + "learning_rate": 4.45978878960195e-06, + "loss": 2.1538, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 1.8481040000915527, + "learning_rate": 4.457757920389927e-06, + "loss": 2.1769, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 2.114439010620117, + "learning_rate": 4.455727051177904e-06, + "loss": 2.2241, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.5492331981658936, + "learning_rate": 4.453696181965882e-06, + "loss": 2.206, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 2.11787486076355, + "learning_rate": 4.451665312753859e-06, + "loss": 2.2328, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.3215034008026123, + "eval_runtime": 37.7471, + "eval_samples_per_second": 26.492, + "eval_steps_per_second": 1.113, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.6561832427978516, + "learning_rate": 4.449634443541836e-06, + "loss": 2.1607, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.7510966062545776, + "learning_rate": 4.4476035743298135e-06, + "loss": 2.1699, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.9424657821655273, + "learning_rate": 4.445572705117791e-06, + "loss": 2.2298, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 1.5836161375045776, + "learning_rate": 4.443541835905768e-06, + "loss": 2.1711, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.895216464996338, + "learning_rate": 4.441510966693745e-06, + "loss": 2.2252, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.3365776538848877, + "learning_rate": 4.439480097481723e-06, + "loss": 2.1309, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.6988815069198608, + "learning_rate": 4.4374492282697e-06, + "loss": 2.2023, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 2.228900671005249, + "learning_rate": 4.435418359057677e-06, + "loss": 2.2021, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.5536632537841797, + "learning_rate": 4.433387489845654e-06, + "loss": 2.1706, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.8463878631591797, + "learning_rate": 4.431356620633632e-06, + "loss": 2.1715, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.320664167404175, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.924741506576538, + "learning_rate": 4.429325751421609e-06, + "loss": 2.1894, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 1.5921828746795654, + "learning_rate": 4.427294882209586e-06, + "loss": 2.1358, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 2.4617693424224854, + "learning_rate": 4.4252640129975635e-06, + "loss": 2.1918, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.6239196062088013, + "learning_rate": 4.423233143785541e-06, + "loss": 2.1996, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.952112078666687, + "learning_rate": 4.421202274573518e-06, + "loss": 2.112, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.663409948348999, + "learning_rate": 4.4191714053614945e-06, + "loss": 2.149, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.831711769104004, + "learning_rate": 4.417140536149473e-06, + "loss": 2.209, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 1.9281738996505737, + "learning_rate": 4.415109666937449e-06, + "loss": 2.1637, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.5661600828170776, + "learning_rate": 4.413078797725427e-06, + "loss": 2.1832, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.6930522918701172, + "learning_rate": 4.411047928513404e-06, + "loss": 2.196, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.318861246109009, + "eval_runtime": 37.6994, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.74558424949646, + "learning_rate": 4.409017059301382e-06, + "loss": 2.2097, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.664880633354187, + "learning_rate": 4.406986190089358e-06, + "loss": 2.2076, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.5981497764587402, + "learning_rate": 4.404955320877336e-06, + "loss": 2.1706, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.6035112142562866, + "learning_rate": 4.402924451665313e-06, + "loss": 2.1538, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 1.6819255352020264, + "learning_rate": 4.400893582453291e-06, + "loss": 2.1816, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.7082457542419434, + "learning_rate": 4.398862713241267e-06, + "loss": 2.1804, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.6893151998519897, + "learning_rate": 4.396831844029245e-06, + "loss": 2.1783, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 1.8639923334121704, + "learning_rate": 4.394800974817222e-06, + "loss": 2.1989, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.6407119035720825, + "learning_rate": 4.392770105605199e-06, + "loss": 2.2013, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.6435848474502563, + "learning_rate": 4.390739236393176e-06, + "loss": 2.2049, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.318084478378296, + "eval_runtime": 37.6999, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.7974674701690674, + "learning_rate": 4.388708367181154e-06, + "loss": 2.188, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 1.8340431451797485, + "learning_rate": 4.386677497969131e-06, + "loss": 2.2305, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.55428946018219, + "learning_rate": 4.384646628757108e-06, + "loss": 2.1837, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.7258585691452026, + "learning_rate": 4.3826157595450855e-06, + "loss": 2.1426, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.7347694635391235, + "learning_rate": 4.380584890333063e-06, + "loss": 2.1161, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.6082992553710938, + "learning_rate": 4.37855402112104e-06, + "loss": 2.2284, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.9446425437927246, + "learning_rate": 4.376523151909017e-06, + "loss": 2.1571, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 1.8597826957702637, + "learning_rate": 4.3744922826969946e-06, + "loss": 2.1982, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.8589327335357666, + "learning_rate": 4.372461413484972e-06, + "loss": 2.2097, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.6825448274612427, + "learning_rate": 4.370430544272949e-06, + "loss": 2.2077, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.315889596939087, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 600 + }, + { + "epoch": 0.6586301369863014, + "grad_norm": 1.8007057905197144, + "learning_rate": 4.368399675060926e-06, + "loss": 2.1843, + "step": 601 + }, + { + "epoch": 0.6597260273972603, + "grad_norm": 2.057697057723999, + "learning_rate": 4.366368805848904e-06, + "loss": 2.13, + "step": 602 + }, + { + "epoch": 0.6608219178082192, + "grad_norm": 1.5714081525802612, + "learning_rate": 4.364337936636881e-06, + "loss": 2.1879, + "step": 603 + }, + { + "epoch": 0.6619178082191781, + "grad_norm": 1.855185866355896, + "learning_rate": 4.362307067424858e-06, + "loss": 2.1659, + "step": 604 + }, + { + "epoch": 0.663013698630137, + "grad_norm": 1.826479196548462, + "learning_rate": 4.3602761982128355e-06, + "loss": 2.1716, + "step": 605 + }, + { + "epoch": 0.6641095890410958, + "grad_norm": 1.6905351877212524, + "learning_rate": 4.358245329000813e-06, + "loss": 2.1825, + "step": 606 + }, + { + "epoch": 0.6652054794520548, + "grad_norm": 1.8157447576522827, + "learning_rate": 4.35621445978879e-06, + "loss": 2.1811, + "step": 607 + }, + { + "epoch": 0.6663013698630137, + "grad_norm": 1.6311850547790527, + "learning_rate": 4.354183590576767e-06, + "loss": 2.1507, + "step": 608 + }, + { + "epoch": 0.6673972602739726, + "grad_norm": 1.579039454460144, + "learning_rate": 4.352152721364745e-06, + "loss": 2.2031, + "step": 609 + }, + { + "epoch": 0.6684931506849315, + "grad_norm": 1.7773518562316895, + "learning_rate": 4.350121852152722e-06, + "loss": 2.1978, + "step": 610 + }, + { + "epoch": 0.6684931506849315, + "eval_loss": 2.314854383468628, + "eval_runtime": 37.4501, + "eval_samples_per_second": 26.702, + "eval_steps_per_second": 1.121, + "step": 610 + }, + { + "epoch": 0.6695890410958905, + "grad_norm": 1.7606587409973145, + "learning_rate": 4.348090982940699e-06, + "loss": 2.1891, + "step": 611 + }, + { + "epoch": 0.6706849315068493, + "grad_norm": 1.6145644187927246, + "learning_rate": 4.3460601137286764e-06, + "loss": 2.1799, + "step": 612 + }, + { + "epoch": 0.6717808219178082, + "grad_norm": 1.8070226907730103, + "learning_rate": 4.344029244516654e-06, + "loss": 2.241, + "step": 613 + }, + { + "epoch": 0.6728767123287671, + "grad_norm": 1.6185903549194336, + "learning_rate": 4.341998375304631e-06, + "loss": 2.1908, + "step": 614 + }, + { + "epoch": 0.673972602739726, + "grad_norm": 1.6690049171447754, + "learning_rate": 4.339967506092607e-06, + "loss": 2.1804, + "step": 615 + }, + { + "epoch": 0.675068493150685, + "grad_norm": 1.740003228187561, + "learning_rate": 4.3379366368805855e-06, + "loss": 2.1734, + "step": 616 + }, + { + "epoch": 0.6761643835616439, + "grad_norm": 1.7311010360717773, + "learning_rate": 4.335905767668562e-06, + "loss": 2.2015, + "step": 617 + }, + { + "epoch": 0.6772602739726027, + "grad_norm": 1.7491734027862549, + "learning_rate": 4.33387489845654e-06, + "loss": 2.1894, + "step": 618 + }, + { + "epoch": 0.6783561643835616, + "grad_norm": 1.6343051195144653, + "learning_rate": 4.3318440292445165e-06, + "loss": 2.1583, + "step": 619 + }, + { + "epoch": 0.6794520547945205, + "grad_norm": 1.584243893623352, + "learning_rate": 4.329813160032495e-06, + "loss": 2.1396, + "step": 620 + }, + { + "epoch": 0.6794520547945205, + "eval_loss": 2.3157904148101807, + "eval_runtime": 37.6734, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 620 + }, + { + "epoch": 0.6805479452054795, + "grad_norm": 1.6275745630264282, + "learning_rate": 4.327782290820471e-06, + "loss": 2.2064, + "step": 621 + }, + { + "epoch": 0.6816438356164384, + "grad_norm": 1.5353672504425049, + "learning_rate": 4.325751421608449e-06, + "loss": 2.1686, + "step": 622 + }, + { + "epoch": 0.6827397260273973, + "grad_norm": 1.6155133247375488, + "learning_rate": 4.323720552396426e-06, + "loss": 2.1652, + "step": 623 + }, + { + "epoch": 0.6838356164383562, + "grad_norm": 1.638383150100708, + "learning_rate": 4.321689683184404e-06, + "loss": 2.1747, + "step": 624 + }, + { + "epoch": 0.684931506849315, + "grad_norm": 1.6153279542922974, + "learning_rate": 4.31965881397238e-06, + "loss": 2.1975, + "step": 625 + }, + { + "epoch": 0.686027397260274, + "grad_norm": 2.023699998855591, + "learning_rate": 4.3176279447603574e-06, + "loss": 2.1904, + "step": 626 + }, + { + "epoch": 0.6871232876712329, + "grad_norm": 1.5969340801239014, + "learning_rate": 4.315597075548335e-06, + "loss": 2.2133, + "step": 627 + }, + { + "epoch": 0.6882191780821918, + "grad_norm": 2.138498067855835, + "learning_rate": 4.313566206336312e-06, + "loss": 2.1685, + "step": 628 + }, + { + "epoch": 0.6893150684931507, + "grad_norm": 1.7141013145446777, + "learning_rate": 4.311535337124289e-06, + "loss": 2.1818, + "step": 629 + }, + { + "epoch": 0.6904109589041096, + "grad_norm": 1.9214462041854858, + "learning_rate": 4.3095044679122665e-06, + "loss": 2.191, + "step": 630 + }, + { + "epoch": 0.6904109589041096, + "eval_loss": 2.3150644302368164, + "eval_runtime": 37.723, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 630 + }, + { + "epoch": 0.6915068493150685, + "grad_norm": 2.03241229057312, + "learning_rate": 4.307473598700244e-06, + "loss": 2.1398, + "step": 631 + }, + { + "epoch": 0.6926027397260274, + "grad_norm": 1.608492374420166, + "learning_rate": 4.305442729488221e-06, + "loss": 2.2113, + "step": 632 + }, + { + "epoch": 0.6936986301369863, + "grad_norm": 2.1223549842834473, + "learning_rate": 4.303411860276198e-06, + "loss": 2.2012, + "step": 633 + }, + { + "epoch": 0.6947945205479452, + "grad_norm": 1.5212339162826538, + "learning_rate": 4.301380991064176e-06, + "loss": 2.2201, + "step": 634 + }, + { + "epoch": 0.6958904109589041, + "grad_norm": 1.9565715789794922, + "learning_rate": 4.299350121852153e-06, + "loss": 2.2155, + "step": 635 + }, + { + "epoch": 0.6969863013698631, + "grad_norm": 1.700822114944458, + "learning_rate": 4.29731925264013e-06, + "loss": 2.1764, + "step": 636 + }, + { + "epoch": 0.6980821917808219, + "grad_norm": 1.5625321865081787, + "learning_rate": 4.2952883834281075e-06, + "loss": 2.1407, + "step": 637 + }, + { + "epoch": 0.6991780821917808, + "grad_norm": 2.3319756984710693, + "learning_rate": 4.293257514216085e-06, + "loss": 2.2058, + "step": 638 + }, + { + "epoch": 0.7002739726027397, + "grad_norm": 1.617644190788269, + "learning_rate": 4.291226645004062e-06, + "loss": 2.2157, + "step": 639 + }, + { + "epoch": 0.7013698630136986, + "grad_norm": 1.9357377290725708, + "learning_rate": 4.289195775792039e-06, + "loss": 2.1736, + "step": 640 + }, + { + "epoch": 0.7013698630136986, + "eval_loss": 2.310267925262451, + "eval_runtime": 37.731, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 640 + }, + { + "epoch": 0.7024657534246576, + "grad_norm": 1.6908962726593018, + "learning_rate": 4.287164906580017e-06, + "loss": 2.1523, + "step": 641 + }, + { + "epoch": 0.7035616438356165, + "grad_norm": 1.6051634550094604, + "learning_rate": 4.285134037367994e-06, + "loss": 2.1499, + "step": 642 + }, + { + "epoch": 0.7046575342465753, + "grad_norm": 1.760164499282837, + "learning_rate": 4.283103168155971e-06, + "loss": 2.1493, + "step": 643 + }, + { + "epoch": 0.7057534246575342, + "grad_norm": 1.6225790977478027, + "learning_rate": 4.281072298943948e-06, + "loss": 2.1737, + "step": 644 + }, + { + "epoch": 0.7068493150684931, + "grad_norm": 1.8128079175949097, + "learning_rate": 4.279041429731926e-06, + "loss": 2.1492, + "step": 645 + }, + { + "epoch": 0.707945205479452, + "grad_norm": 1.993295431137085, + "learning_rate": 4.277010560519903e-06, + "loss": 2.1702, + "step": 646 + }, + { + "epoch": 0.709041095890411, + "grad_norm": 1.7715190649032593, + "learning_rate": 4.27497969130788e-06, + "loss": 2.1939, + "step": 647 + }, + { + "epoch": 0.7101369863013699, + "grad_norm": 1.921738624572754, + "learning_rate": 4.2729488220958575e-06, + "loss": 2.201, + "step": 648 + }, + { + "epoch": 0.7112328767123288, + "grad_norm": 1.5670042037963867, + "learning_rate": 4.270917952883835e-06, + "loss": 2.152, + "step": 649 + }, + { + "epoch": 0.7123287671232876, + "grad_norm": 1.6405668258666992, + "learning_rate": 4.268887083671812e-06, + "loss": 2.2488, + "step": 650 + }, + { + "epoch": 0.7123287671232876, + "eval_loss": 2.3100948333740234, + "eval_runtime": 37.7393, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 650 + }, + { + "epoch": 0.7134246575342466, + "grad_norm": 1.79966402053833, + "learning_rate": 4.266856214459789e-06, + "loss": 2.1506, + "step": 651 + }, + { + "epoch": 0.7145205479452055, + "grad_norm": 1.6062074899673462, + "learning_rate": 4.264825345247766e-06, + "loss": 2.1818, + "step": 652 + }, + { + "epoch": 0.7156164383561644, + "grad_norm": 1.61439049243927, + "learning_rate": 4.262794476035744e-06, + "loss": 2.1269, + "step": 653 + }, + { + "epoch": 0.7167123287671233, + "grad_norm": 1.7649872303009033, + "learning_rate": 4.26076360682372e-06, + "loss": 2.1703, + "step": 654 + }, + { + "epoch": 0.7178082191780822, + "grad_norm": 1.6238528490066528, + "learning_rate": 4.2587327376116984e-06, + "loss": 2.1764, + "step": 655 + }, + { + "epoch": 0.718904109589041, + "grad_norm": 1.6847354173660278, + "learning_rate": 4.256701868399675e-06, + "loss": 2.1796, + "step": 656 + }, + { + "epoch": 0.72, + "grad_norm": 1.7208436727523804, + "learning_rate": 4.254670999187653e-06, + "loss": 2.1903, + "step": 657 + }, + { + "epoch": 0.7210958904109589, + "grad_norm": 1.5888099670410156, + "learning_rate": 4.252640129975629e-06, + "loss": 2.1686, + "step": 658 + }, + { + "epoch": 0.7221917808219178, + "grad_norm": 1.742118000984192, + "learning_rate": 4.2506092607636075e-06, + "loss": 2.1287, + "step": 659 + }, + { + "epoch": 0.7232876712328767, + "grad_norm": 1.8450883626937866, + "learning_rate": 4.248578391551584e-06, + "loss": 2.1761, + "step": 660 + }, + { + "epoch": 0.7232876712328767, + "eval_loss": 2.3134732246398926, + "eval_runtime": 37.7568, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 660 + }, + { + "epoch": 0.7243835616438357, + "grad_norm": 1.630985140800476, + "learning_rate": 4.246547522339562e-06, + "loss": 2.1743, + "step": 661 + }, + { + "epoch": 0.7254794520547945, + "grad_norm": 1.6984914541244507, + "learning_rate": 4.2445166531275385e-06, + "loss": 2.1356, + "step": 662 + }, + { + "epoch": 0.7265753424657534, + "grad_norm": 1.6732920408248901, + "learning_rate": 4.242485783915517e-06, + "loss": 2.1377, + "step": 663 + }, + { + "epoch": 0.7276712328767123, + "grad_norm": 1.6210153102874756, + "learning_rate": 4.240454914703493e-06, + "loss": 2.1455, + "step": 664 + }, + { + "epoch": 0.7287671232876712, + "grad_norm": 1.7643382549285889, + "learning_rate": 4.23842404549147e-06, + "loss": 2.1373, + "step": 665 + }, + { + "epoch": 0.7298630136986302, + "grad_norm": 1.6491565704345703, + "learning_rate": 4.236393176279448e-06, + "loss": 2.1897, + "step": 666 + }, + { + "epoch": 0.7309589041095891, + "grad_norm": 1.670323133468628, + "learning_rate": 4.234362307067425e-06, + "loss": 2.2081, + "step": 667 + }, + { + "epoch": 0.7320547945205479, + "grad_norm": 1.6350057125091553, + "learning_rate": 4.232331437855402e-06, + "loss": 2.2002, + "step": 668 + }, + { + "epoch": 0.7331506849315068, + "grad_norm": 1.6191266775131226, + "learning_rate": 4.2303005686433795e-06, + "loss": 2.1744, + "step": 669 + }, + { + "epoch": 0.7342465753424657, + "grad_norm": 1.5511432886123657, + "learning_rate": 4.228269699431357e-06, + "loss": 2.1797, + "step": 670 + }, + { + "epoch": 0.7342465753424657, + "eval_loss": 2.313626527786255, + "eval_runtime": 37.7763, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 670 + }, + { + "epoch": 0.7353424657534247, + "grad_norm": 1.5128833055496216, + "learning_rate": 4.226238830219334e-06, + "loss": 2.1729, + "step": 671 + }, + { + "epoch": 0.7364383561643836, + "grad_norm": 1.6752032041549683, + "learning_rate": 4.224207961007311e-06, + "loss": 2.1747, + "step": 672 + }, + { + "epoch": 0.7375342465753425, + "grad_norm": 1.8643760681152344, + "learning_rate": 4.2221770917952886e-06, + "loss": 2.174, + "step": 673 + }, + { + "epoch": 0.7386301369863014, + "grad_norm": 1.5659266710281372, + "learning_rate": 4.220146222583266e-06, + "loss": 2.1706, + "step": 674 + }, + { + "epoch": 0.7397260273972602, + "grad_norm": 1.9675986766815186, + "learning_rate": 4.218115353371243e-06, + "loss": 2.1558, + "step": 675 + }, + { + "epoch": 0.7408219178082192, + "grad_norm": 1.6091173887252808, + "learning_rate": 4.21608448415922e-06, + "loss": 2.2126, + "step": 676 + }, + { + "epoch": 0.7419178082191781, + "grad_norm": 1.6801071166992188, + "learning_rate": 4.214053614947198e-06, + "loss": 2.1352, + "step": 677 + }, + { + "epoch": 0.743013698630137, + "grad_norm": 1.6964598894119263, + "learning_rate": 4.212022745735175e-06, + "loss": 2.1863, + "step": 678 + }, + { + "epoch": 0.7441095890410959, + "grad_norm": 1.670597791671753, + "learning_rate": 4.209991876523152e-06, + "loss": 2.1478, + "step": 679 + }, + { + "epoch": 0.7452054794520548, + "grad_norm": 1.6591588258743286, + "learning_rate": 4.2079610073111295e-06, + "loss": 2.2198, + "step": 680 + }, + { + "epoch": 0.7452054794520548, + "eval_loss": 2.3091039657592773, + "eval_runtime": 37.7551, + "eval_samples_per_second": 26.486, + "eval_steps_per_second": 1.112, + "step": 680 + }, + { + "epoch": 0.7463013698630137, + "grad_norm": 1.5147838592529297, + "learning_rate": 4.205930138099107e-06, + "loss": 2.1836, + "step": 681 + }, + { + "epoch": 0.7473972602739726, + "grad_norm": 1.6278952360153198, + "learning_rate": 4.203899268887084e-06, + "loss": 2.196, + "step": 682 + }, + { + "epoch": 0.7484931506849315, + "grad_norm": 1.5964363813400269, + "learning_rate": 4.201868399675061e-06, + "loss": 2.1162, + "step": 683 + }, + { + "epoch": 0.7495890410958904, + "grad_norm": 1.6993104219436646, + "learning_rate": 4.199837530463039e-06, + "loss": 2.1731, + "step": 684 + }, + { + "epoch": 0.7506849315068493, + "grad_norm": 1.6196811199188232, + "learning_rate": 4.197806661251016e-06, + "loss": 2.16, + "step": 685 + }, + { + "epoch": 0.7517808219178083, + "grad_norm": 1.6826317310333252, + "learning_rate": 4.195775792038993e-06, + "loss": 2.1123, + "step": 686 + }, + { + "epoch": 0.7528767123287671, + "grad_norm": 1.7604600191116333, + "learning_rate": 4.19374492282697e-06, + "loss": 2.1962, + "step": 687 + }, + { + "epoch": 0.753972602739726, + "grad_norm": 1.5533616542816162, + "learning_rate": 4.191714053614948e-06, + "loss": 2.1593, + "step": 688 + }, + { + "epoch": 0.7550684931506849, + "grad_norm": 1.8652232885360718, + "learning_rate": 4.189683184402924e-06, + "loss": 2.2067, + "step": 689 + }, + { + "epoch": 0.7561643835616438, + "grad_norm": 1.6330097913742065, + "learning_rate": 4.187652315190902e-06, + "loss": 2.1483, + "step": 690 + }, + { + "epoch": 0.7561643835616438, + "eval_loss": 2.309835195541382, + "eval_runtime": 37.746, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 690 + }, + { + "epoch": 0.7572602739726028, + "grad_norm": 1.5989009141921997, + "learning_rate": 4.185621445978879e-06, + "loss": 2.1686, + "step": 691 + }, + { + "epoch": 0.7583561643835617, + "grad_norm": 1.7416167259216309, + "learning_rate": 4.183590576766857e-06, + "loss": 2.2111, + "step": 692 + }, + { + "epoch": 0.7594520547945205, + "grad_norm": 1.5840280055999756, + "learning_rate": 4.181559707554833e-06, + "loss": 2.1499, + "step": 693 + }, + { + "epoch": 0.7605479452054794, + "grad_norm": 1.839306354522705, + "learning_rate": 4.179528838342811e-06, + "loss": 2.1492, + "step": 694 + }, + { + "epoch": 0.7616438356164383, + "grad_norm": 1.54163658618927, + "learning_rate": 4.177497969130788e-06, + "loss": 2.2029, + "step": 695 + }, + { + "epoch": 0.7627397260273973, + "grad_norm": 1.8050168752670288, + "learning_rate": 4.175467099918766e-06, + "loss": 2.1615, + "step": 696 + }, + { + "epoch": 0.7638356164383562, + "grad_norm": 1.6971197128295898, + "learning_rate": 4.173436230706742e-06, + "loss": 2.1476, + "step": 697 + }, + { + "epoch": 0.7649315068493151, + "grad_norm": 1.6200546026229858, + "learning_rate": 4.1714053614947205e-06, + "loss": 2.1602, + "step": 698 + }, + { + "epoch": 0.766027397260274, + "grad_norm": 1.7440853118896484, + "learning_rate": 4.169374492282697e-06, + "loss": 2.1949, + "step": 699 + }, + { + "epoch": 0.7671232876712328, + "grad_norm": 1.5590331554412842, + "learning_rate": 4.167343623070675e-06, + "loss": 2.1656, + "step": 700 + }, + { + "epoch": 0.7671232876712328, + "eval_loss": 2.309752941131592, + "eval_runtime": 37.7663, + "eval_samples_per_second": 26.479, + "eval_steps_per_second": 1.112, + "step": 700 + }, + { + "epoch": 0.7682191780821918, + "grad_norm": 1.7314256429672241, + "learning_rate": 4.165312753858652e-06, + "loss": 2.1516, + "step": 701 + }, + { + "epoch": 0.7693150684931507, + "grad_norm": 1.5145938396453857, + "learning_rate": 4.163281884646629e-06, + "loss": 2.1795, + "step": 702 + }, + { + "epoch": 0.7704109589041096, + "grad_norm": 1.5300369262695312, + "learning_rate": 4.161251015434607e-06, + "loss": 2.1356, + "step": 703 + }, + { + "epoch": 0.7715068493150685, + "grad_norm": 1.6113280057907104, + "learning_rate": 4.159220146222583e-06, + "loss": 2.157, + "step": 704 + }, + { + "epoch": 0.7726027397260274, + "grad_norm": 1.5458502769470215, + "learning_rate": 4.157189277010561e-06, + "loss": 2.2017, + "step": 705 + }, + { + "epoch": 0.7736986301369863, + "grad_norm": 1.5456187725067139, + "learning_rate": 4.155158407798538e-06, + "loss": 2.1758, + "step": 706 + }, + { + "epoch": 0.7747945205479452, + "grad_norm": 1.6051454544067383, + "learning_rate": 4.153127538586516e-06, + "loss": 2.183, + "step": 707 + }, + { + "epoch": 0.7758904109589041, + "grad_norm": 1.565788745880127, + "learning_rate": 4.151096669374492e-06, + "loss": 2.1538, + "step": 708 + }, + { + "epoch": 0.776986301369863, + "grad_norm": 1.7484439611434937, + "learning_rate": 4.1490658001624705e-06, + "loss": 2.169, + "step": 709 + }, + { + "epoch": 0.7780821917808219, + "grad_norm": 1.7206509113311768, + "learning_rate": 4.147034930950447e-06, + "loss": 2.1661, + "step": 710 + }, + { + "epoch": 0.7780821917808219, + "eval_loss": 2.307927131652832, + "eval_runtime": 37.5145, + "eval_samples_per_second": 26.656, + "eval_steps_per_second": 1.12, + "step": 710 + }, + { + "epoch": 0.7791780821917809, + "grad_norm": 1.8865350484848022, + "learning_rate": 4.145004061738425e-06, + "loss": 2.093, + "step": 711 + }, + { + "epoch": 0.7802739726027397, + "grad_norm": 1.935110092163086, + "learning_rate": 4.1429731925264015e-06, + "loss": 2.1298, + "step": 712 + }, + { + "epoch": 0.7813698630136986, + "grad_norm": 1.6938892602920532, + "learning_rate": 4.140942323314379e-06, + "loss": 2.1508, + "step": 713 + }, + { + "epoch": 0.7824657534246575, + "grad_norm": 1.5774824619293213, + "learning_rate": 4.138911454102356e-06, + "loss": 2.1808, + "step": 714 + }, + { + "epoch": 0.7835616438356164, + "grad_norm": 1.8637802600860596, + "learning_rate": 4.136880584890333e-06, + "loss": 2.1822, + "step": 715 + }, + { + "epoch": 0.7846575342465754, + "grad_norm": 1.6216785907745361, + "learning_rate": 4.1348497156783106e-06, + "loss": 2.1807, + "step": 716 + }, + { + "epoch": 0.7857534246575343, + "grad_norm": 1.7688151597976685, + "learning_rate": 4.132818846466288e-06, + "loss": 2.1967, + "step": 717 + }, + { + "epoch": 0.7868493150684932, + "grad_norm": 1.6824712753295898, + "learning_rate": 4.130787977254265e-06, + "loss": 2.1952, + "step": 718 + }, + { + "epoch": 0.787945205479452, + "grad_norm": 1.911048412322998, + "learning_rate": 4.128757108042242e-06, + "loss": 2.2408, + "step": 719 + }, + { + "epoch": 0.7890410958904109, + "grad_norm": 1.62872314453125, + "learning_rate": 4.12672623883022e-06, + "loss": 2.1647, + "step": 720 + }, + { + "epoch": 0.7890410958904109, + "eval_loss": 2.3071532249450684, + "eval_runtime": 37.6838, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 720 + }, + { + "epoch": 0.7901369863013699, + "grad_norm": 2.080674171447754, + "learning_rate": 4.124695369618197e-06, + "loss": 2.1614, + "step": 721 + }, + { + "epoch": 0.7912328767123288, + "grad_norm": 1.7225316762924194, + "learning_rate": 4.122664500406174e-06, + "loss": 2.1784, + "step": 722 + }, + { + "epoch": 0.7923287671232877, + "grad_norm": 1.838957667350769, + "learning_rate": 4.1206336311941515e-06, + "loss": 2.1281, + "step": 723 + }, + { + "epoch": 0.7934246575342466, + "grad_norm": 1.9360777139663696, + "learning_rate": 4.118602761982129e-06, + "loss": 2.1168, + "step": 724 + }, + { + "epoch": 0.7945205479452054, + "grad_norm": 1.7712862491607666, + "learning_rate": 4.116571892770106e-06, + "loss": 2.2004, + "step": 725 + }, + { + "epoch": 0.7956164383561644, + "grad_norm": 1.6759192943572998, + "learning_rate": 4.114541023558083e-06, + "loss": 2.1588, + "step": 726 + }, + { + "epoch": 0.7967123287671233, + "grad_norm": 1.8396296501159668, + "learning_rate": 4.112510154346061e-06, + "loss": 2.1777, + "step": 727 + }, + { + "epoch": 0.7978082191780822, + "grad_norm": 1.6775752305984497, + "learning_rate": 4.110479285134038e-06, + "loss": 2.1923, + "step": 728 + }, + { + "epoch": 0.7989041095890411, + "grad_norm": 1.7111576795578003, + "learning_rate": 4.108448415922015e-06, + "loss": 2.1854, + "step": 729 + }, + { + "epoch": 0.8, + "grad_norm": 1.7400974035263062, + "learning_rate": 4.1064175467099924e-06, + "loss": 2.0897, + "step": 730 + }, + { + "epoch": 0.8, + "eval_loss": 2.3070170879364014, + "eval_runtime": 37.7602, + "eval_samples_per_second": 26.483, + "eval_steps_per_second": 1.112, + "step": 730 + }, + { + "epoch": 0.8010958904109589, + "grad_norm": 1.9166353940963745, + "learning_rate": 4.10438667749797e-06, + "loss": 2.2417, + "step": 731 + }, + { + "epoch": 0.8021917808219178, + "grad_norm": 1.7914725542068481, + "learning_rate": 4.102355808285947e-06, + "loss": 2.1677, + "step": 732 + }, + { + "epoch": 0.8032876712328767, + "grad_norm": 1.7602828741073608, + "learning_rate": 4.100324939073924e-06, + "loss": 2.1623, + "step": 733 + }, + { + "epoch": 0.8043835616438356, + "grad_norm": 1.6864640712738037, + "learning_rate": 4.0982940698619015e-06, + "loss": 2.1748, + "step": 734 + }, + { + "epoch": 0.8054794520547945, + "grad_norm": 1.564070463180542, + "learning_rate": 4.096263200649879e-06, + "loss": 2.1458, + "step": 735 + }, + { + "epoch": 0.8065753424657535, + "grad_norm": 1.9004014730453491, + "learning_rate": 4.094232331437856e-06, + "loss": 2.1536, + "step": 736 + }, + { + "epoch": 0.8076712328767123, + "grad_norm": 1.6670420169830322, + "learning_rate": 4.092201462225833e-06, + "loss": 2.1809, + "step": 737 + }, + { + "epoch": 0.8087671232876712, + "grad_norm": 1.6553101539611816, + "learning_rate": 4.090170593013811e-06, + "loss": 2.1343, + "step": 738 + }, + { + "epoch": 0.8098630136986301, + "grad_norm": 1.9083986282348633, + "learning_rate": 4.088139723801787e-06, + "loss": 2.1455, + "step": 739 + }, + { + "epoch": 0.810958904109589, + "grad_norm": 1.6069375276565552, + "learning_rate": 4.086108854589765e-06, + "loss": 2.1621, + "step": 740 + }, + { + "epoch": 0.810958904109589, + "eval_loss": 2.3045499324798584, + "eval_runtime": 37.7804, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 740 + }, + { + "epoch": 0.812054794520548, + "grad_norm": 1.6774693727493286, + "learning_rate": 4.084077985377742e-06, + "loss": 2.2079, + "step": 741 + }, + { + "epoch": 0.8131506849315069, + "grad_norm": 1.772355556488037, + "learning_rate": 4.08204711616572e-06, + "loss": 2.1276, + "step": 742 + }, + { + "epoch": 0.8142465753424658, + "grad_norm": 1.6998035907745361, + "learning_rate": 4.080016246953696e-06, + "loss": 2.1903, + "step": 743 + }, + { + "epoch": 0.8153424657534246, + "grad_norm": 1.850844383239746, + "learning_rate": 4.077985377741674e-06, + "loss": 2.1965, + "step": 744 + }, + { + "epoch": 0.8164383561643835, + "grad_norm": 1.5916879177093506, + "learning_rate": 4.075954508529651e-06, + "loss": 2.1982, + "step": 745 + }, + { + "epoch": 0.8175342465753425, + "grad_norm": 1.8195812702178955, + "learning_rate": 4.073923639317629e-06, + "loss": 2.1507, + "step": 746 + }, + { + "epoch": 0.8186301369863014, + "grad_norm": 1.5145621299743652, + "learning_rate": 4.071892770105605e-06, + "loss": 2.1236, + "step": 747 + }, + { + "epoch": 0.8197260273972603, + "grad_norm": 1.5750877857208252, + "learning_rate": 4.069861900893583e-06, + "loss": 2.142, + "step": 748 + }, + { + "epoch": 0.8208219178082192, + "grad_norm": 1.708666443824768, + "learning_rate": 4.06783103168156e-06, + "loss": 2.1577, + "step": 749 + }, + { + "epoch": 0.821917808219178, + "grad_norm": 1.5988762378692627, + "learning_rate": 4.065800162469538e-06, + "loss": 2.1795, + "step": 750 + }, + { + "epoch": 0.821917808219178, + "eval_loss": 2.3052196502685547, + "eval_runtime": 37.7983, + "eval_samples_per_second": 26.456, + "eval_steps_per_second": 1.111, + "step": 750 + }, + { + "epoch": 0.823013698630137, + "grad_norm": 1.608738899230957, + "learning_rate": 4.063769293257514e-06, + "loss": 2.1578, + "step": 751 + }, + { + "epoch": 0.8241095890410959, + "grad_norm": 1.5578334331512451, + "learning_rate": 4.061738424045492e-06, + "loss": 2.1724, + "step": 752 + }, + { + "epoch": 0.8252054794520548, + "grad_norm": 1.6957333087921143, + "learning_rate": 4.059707554833469e-06, + "loss": 2.1975, + "step": 753 + }, + { + "epoch": 0.8263013698630137, + "grad_norm": 1.691512107849121, + "learning_rate": 4.057676685621446e-06, + "loss": 2.1705, + "step": 754 + }, + { + "epoch": 0.8273972602739726, + "grad_norm": 1.6061747074127197, + "learning_rate": 4.0556458164094235e-06, + "loss": 2.1031, + "step": 755 + }, + { + "epoch": 0.8284931506849315, + "grad_norm": 1.6074447631835938, + "learning_rate": 4.053614947197401e-06, + "loss": 2.1738, + "step": 756 + }, + { + "epoch": 0.8295890410958904, + "grad_norm": 1.9722917079925537, + "learning_rate": 4.051584077985378e-06, + "loss": 2.2043, + "step": 757 + }, + { + "epoch": 0.8306849315068493, + "grad_norm": 1.597692847251892, + "learning_rate": 4.049553208773355e-06, + "loss": 2.1621, + "step": 758 + }, + { + "epoch": 0.8317808219178082, + "grad_norm": 1.85458242893219, + "learning_rate": 4.047522339561333e-06, + "loss": 2.1504, + "step": 759 + }, + { + "epoch": 0.8328767123287671, + "grad_norm": 1.8328344821929932, + "learning_rate": 4.04549147034931e-06, + "loss": 2.2541, + "step": 760 + }, + { + "epoch": 0.8328767123287671, + "eval_loss": 2.3033809661865234, + "eval_runtime": 37.8088, + "eval_samples_per_second": 26.449, + "eval_steps_per_second": 1.111, + "step": 760 + }, + { + "epoch": 0.8339726027397261, + "grad_norm": 1.6544418334960938, + "learning_rate": 4.043460601137287e-06, + "loss": 2.1605, + "step": 761 + }, + { + "epoch": 0.8350684931506849, + "grad_norm": 1.9443641901016235, + "learning_rate": 4.041429731925264e-06, + "loss": 2.1621, + "step": 762 + }, + { + "epoch": 0.8361643835616438, + "grad_norm": 1.731277346611023, + "learning_rate": 4.039398862713242e-06, + "loss": 2.2168, + "step": 763 + }, + { + "epoch": 0.8372602739726027, + "grad_norm": 1.671072244644165, + "learning_rate": 4.037367993501219e-06, + "loss": 2.1612, + "step": 764 + }, + { + "epoch": 0.8383561643835616, + "grad_norm": 1.8777449131011963, + "learning_rate": 4.035337124289196e-06, + "loss": 2.1339, + "step": 765 + }, + { + "epoch": 0.8394520547945206, + "grad_norm": 1.5941543579101562, + "learning_rate": 4.0333062550771735e-06, + "loss": 2.1545, + "step": 766 + }, + { + "epoch": 0.8405479452054795, + "grad_norm": 1.9245301485061646, + "learning_rate": 4.031275385865151e-06, + "loss": 2.187, + "step": 767 + }, + { + "epoch": 0.8416438356164384, + "grad_norm": 1.69217848777771, + "learning_rate": 4.029244516653128e-06, + "loss": 2.1168, + "step": 768 + }, + { + "epoch": 0.8427397260273972, + "grad_norm": 1.7042242288589478, + "learning_rate": 4.027213647441105e-06, + "loss": 2.1353, + "step": 769 + }, + { + "epoch": 0.8438356164383561, + "grad_norm": 1.6391927003860474, + "learning_rate": 4.025182778229083e-06, + "loss": 2.1568, + "step": 770 + }, + { + "epoch": 0.8438356164383561, + "eval_loss": 2.310056686401367, + "eval_runtime": 37.7573, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 770 + }, + { + "epoch": 0.8449315068493151, + "grad_norm": 1.7833209037780762, + "learning_rate": 4.02315190901706e-06, + "loss": 2.1664, + "step": 771 + }, + { + "epoch": 0.846027397260274, + "grad_norm": 1.8203891515731812, + "learning_rate": 4.021121039805037e-06, + "loss": 2.2342, + "step": 772 + }, + { + "epoch": 0.8471232876712329, + "grad_norm": 1.572940707206726, + "learning_rate": 4.0190901705930144e-06, + "loss": 2.1418, + "step": 773 + }, + { + "epoch": 0.8482191780821918, + "grad_norm": 1.6871955394744873, + "learning_rate": 4.017059301380992e-06, + "loss": 2.2023, + "step": 774 + }, + { + "epoch": 0.8493150684931506, + "grad_norm": 1.6343133449554443, + "learning_rate": 4.015028432168969e-06, + "loss": 2.1302, + "step": 775 + }, + { + "epoch": 0.8504109589041096, + "grad_norm": 1.4492065906524658, + "learning_rate": 4.012997562956945e-06, + "loss": 2.1427, + "step": 776 + }, + { + "epoch": 0.8515068493150685, + "grad_norm": 1.6599146127700806, + "learning_rate": 4.0109666937449235e-06, + "loss": 2.1479, + "step": 777 + }, + { + "epoch": 0.8526027397260274, + "grad_norm": 1.6836353540420532, + "learning_rate": 4.0089358245329e-06, + "loss": 2.148, + "step": 778 + }, + { + "epoch": 0.8536986301369863, + "grad_norm": 1.5977990627288818, + "learning_rate": 4.006904955320878e-06, + "loss": 2.1946, + "step": 779 + }, + { + "epoch": 0.8547945205479452, + "grad_norm": 1.6259957551956177, + "learning_rate": 4.0048740861088545e-06, + "loss": 2.1761, + "step": 780 + }, + { + "epoch": 0.8547945205479452, + "eval_loss": 2.303910970687866, + "eval_runtime": 37.7728, + "eval_samples_per_second": 26.474, + "eval_steps_per_second": 1.112, + "step": 780 + }, + { + "epoch": 0.8558904109589041, + "grad_norm": 1.6990300416946411, + "learning_rate": 4.002843216896833e-06, + "loss": 2.1309, + "step": 781 + }, + { + "epoch": 0.856986301369863, + "grad_norm": 1.6556029319763184, + "learning_rate": 4.000812347684809e-06, + "loss": 2.151, + "step": 782 + }, + { + "epoch": 0.8580821917808219, + "grad_norm": 1.5659290552139282, + "learning_rate": 3.998781478472787e-06, + "loss": 2.1317, + "step": 783 + }, + { + "epoch": 0.8591780821917808, + "grad_norm": 1.6553704738616943, + "learning_rate": 3.996750609260764e-06, + "loss": 2.2344, + "step": 784 + }, + { + "epoch": 0.8602739726027397, + "grad_norm": 1.6608418226242065, + "learning_rate": 3.994719740048742e-06, + "loss": 2.1807, + "step": 785 + }, + { + "epoch": 0.8613698630136987, + "grad_norm": 1.5862030982971191, + "learning_rate": 3.992688870836718e-06, + "loss": 2.1351, + "step": 786 + }, + { + "epoch": 0.8624657534246576, + "grad_norm": 1.767526388168335, + "learning_rate": 3.990658001624696e-06, + "loss": 2.1856, + "step": 787 + }, + { + "epoch": 0.8635616438356164, + "grad_norm": 1.6902577877044678, + "learning_rate": 3.988627132412673e-06, + "loss": 2.1592, + "step": 788 + }, + { + "epoch": 0.8646575342465753, + "grad_norm": 1.7760623693466187, + "learning_rate": 3.98659626320065e-06, + "loss": 2.1645, + "step": 789 + }, + { + "epoch": 0.8657534246575342, + "grad_norm": 1.6791038513183594, + "learning_rate": 3.984565393988627e-06, + "loss": 2.1611, + "step": 790 + }, + { + "epoch": 0.8657534246575342, + "eval_loss": 2.3027942180633545, + "eval_runtime": 37.736, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 790 + }, + { + "epoch": 0.8668493150684932, + "grad_norm": 1.6695773601531982, + "learning_rate": 3.9825345247766046e-06, + "loss": 2.1747, + "step": 791 + }, + { + "epoch": 0.8679452054794521, + "grad_norm": 1.8736445903778076, + "learning_rate": 3.980503655564582e-06, + "loss": 2.1806, + "step": 792 + }, + { + "epoch": 0.869041095890411, + "grad_norm": 1.6695518493652344, + "learning_rate": 3.978472786352559e-06, + "loss": 2.1631, + "step": 793 + }, + { + "epoch": 0.8701369863013698, + "grad_norm": 1.6782366037368774, + "learning_rate": 3.976441917140536e-06, + "loss": 2.2175, + "step": 794 + }, + { + "epoch": 0.8712328767123287, + "grad_norm": 1.678619623184204, + "learning_rate": 3.974411047928514e-06, + "loss": 2.1883, + "step": 795 + }, + { + "epoch": 0.8723287671232877, + "grad_norm": 1.6274909973144531, + "learning_rate": 3.972380178716491e-06, + "loss": 2.1866, + "step": 796 + }, + { + "epoch": 0.8734246575342466, + "grad_norm": 1.5609477758407593, + "learning_rate": 3.970349309504468e-06, + "loss": 2.1257, + "step": 797 + }, + { + "epoch": 0.8745205479452055, + "grad_norm": 1.7419973611831665, + "learning_rate": 3.9683184402924455e-06, + "loss": 2.1253, + "step": 798 + }, + { + "epoch": 0.8756164383561644, + "grad_norm": 1.6296330690383911, + "learning_rate": 3.966287571080423e-06, + "loss": 2.1755, + "step": 799 + }, + { + "epoch": 0.8767123287671232, + "grad_norm": 1.5995864868164062, + "learning_rate": 3.9642567018684e-06, + "loss": 2.169, + "step": 800 + }, + { + "epoch": 0.8767123287671232, + "eval_loss": 2.301098108291626, + "eval_runtime": 37.7893, + "eval_samples_per_second": 26.463, + "eval_steps_per_second": 1.111, + "step": 800 + }, + { + "epoch": 0.8778082191780822, + "grad_norm": 1.653707504272461, + "learning_rate": 3.962225832656377e-06, + "loss": 2.115, + "step": 801 + }, + { + "epoch": 0.8789041095890411, + "grad_norm": 1.5908581018447876, + "learning_rate": 3.960194963444355e-06, + "loss": 2.1659, + "step": 802 + }, + { + "epoch": 0.88, + "grad_norm": 1.6865545511245728, + "learning_rate": 3.958164094232332e-06, + "loss": 2.1666, + "step": 803 + }, + { + "epoch": 0.8810958904109589, + "grad_norm": 1.6234707832336426, + "learning_rate": 3.956133225020309e-06, + "loss": 2.1594, + "step": 804 + }, + { + "epoch": 0.8821917808219178, + "grad_norm": 1.7375986576080322, + "learning_rate": 3.954102355808286e-06, + "loss": 2.1596, + "step": 805 + }, + { + "epoch": 0.8832876712328767, + "grad_norm": 1.6529247760772705, + "learning_rate": 3.952071486596264e-06, + "loss": 2.1495, + "step": 806 + }, + { + "epoch": 0.8843835616438356, + "grad_norm": 1.517357349395752, + "learning_rate": 3.950040617384241e-06, + "loss": 2.1394, + "step": 807 + }, + { + "epoch": 0.8854794520547945, + "grad_norm": 1.7367429733276367, + "learning_rate": 3.948009748172218e-06, + "loss": 2.206, + "step": 808 + }, + { + "epoch": 0.8865753424657534, + "grad_norm": 1.5467783212661743, + "learning_rate": 3.9459788789601955e-06, + "loss": 2.1672, + "step": 809 + }, + { + "epoch": 0.8876712328767123, + "grad_norm": 1.5647791624069214, + "learning_rate": 3.943948009748173e-06, + "loss": 2.1695, + "step": 810 + }, + { + "epoch": 0.8876712328767123, + "eval_loss": 2.2975995540618896, + "eval_runtime": 37.4912, + "eval_samples_per_second": 26.673, + "eval_steps_per_second": 1.12, + "step": 810 + }, + { + "epoch": 0.8887671232876713, + "grad_norm": 1.597876787185669, + "learning_rate": 3.94191714053615e-06, + "loss": 2.1808, + "step": 811 + }, + { + "epoch": 0.8898630136986302, + "grad_norm": 1.7645189762115479, + "learning_rate": 3.939886271324127e-06, + "loss": 2.1291, + "step": 812 + }, + { + "epoch": 0.890958904109589, + "grad_norm": 1.5209035873413086, + "learning_rate": 3.937855402112104e-06, + "loss": 2.1269, + "step": 813 + }, + { + "epoch": 0.8920547945205479, + "grad_norm": 1.5804905891418457, + "learning_rate": 3.935824532900082e-06, + "loss": 2.1751, + "step": 814 + }, + { + "epoch": 0.8931506849315068, + "grad_norm": 1.687865972518921, + "learning_rate": 3.933793663688058e-06, + "loss": 2.1917, + "step": 815 + }, + { + "epoch": 0.8942465753424658, + "grad_norm": 1.6767957210540771, + "learning_rate": 3.9317627944760365e-06, + "loss": 2.118, + "step": 816 + }, + { + "epoch": 0.8953424657534247, + "grad_norm": 1.657046914100647, + "learning_rate": 3.929731925264013e-06, + "loss": 2.1479, + "step": 817 + }, + { + "epoch": 0.8964383561643836, + "grad_norm": 1.5638110637664795, + "learning_rate": 3.927701056051991e-06, + "loss": 2.1565, + "step": 818 + }, + { + "epoch": 0.8975342465753424, + "grad_norm": 1.6865980625152588, + "learning_rate": 3.9256701868399674e-06, + "loss": 2.1382, + "step": 819 + }, + { + "epoch": 0.8986301369863013, + "grad_norm": 1.6738817691802979, + "learning_rate": 3.9236393176279456e-06, + "loss": 2.1257, + "step": 820 + }, + { + "epoch": 0.8986301369863013, + "eval_loss": 2.2953474521636963, + "eval_runtime": 37.7016, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 820 + }, + { + "epoch": 0.8997260273972603, + "grad_norm": 1.568678617477417, + "learning_rate": 3.921608448415922e-06, + "loss": 2.169, + "step": 821 + }, + { + "epoch": 0.9008219178082192, + "grad_norm": 1.6803818941116333, + "learning_rate": 3.9195775792039e-06, + "loss": 2.1489, + "step": 822 + }, + { + "epoch": 0.9019178082191781, + "grad_norm": 1.7010858058929443, + "learning_rate": 3.9175467099918765e-06, + "loss": 2.1846, + "step": 823 + }, + { + "epoch": 0.903013698630137, + "grad_norm": 1.5782372951507568, + "learning_rate": 3.915515840779855e-06, + "loss": 2.1788, + "step": 824 + }, + { + "epoch": 0.9041095890410958, + "grad_norm": 1.7804685831069946, + "learning_rate": 3.913484971567831e-06, + "loss": 2.1058, + "step": 825 + }, + { + "epoch": 0.9052054794520548, + "grad_norm": 1.5909292697906494, + "learning_rate": 3.911454102355808e-06, + "loss": 2.1716, + "step": 826 + }, + { + "epoch": 0.9063013698630137, + "grad_norm": 1.7001307010650635, + "learning_rate": 3.909423233143786e-06, + "loss": 2.1647, + "step": 827 + }, + { + "epoch": 0.9073972602739726, + "grad_norm": 1.5576646327972412, + "learning_rate": 3.907392363931763e-06, + "loss": 2.2174, + "step": 828 + }, + { + "epoch": 0.9084931506849315, + "grad_norm": 1.6621390581130981, + "learning_rate": 3.90536149471974e-06, + "loss": 2.2344, + "step": 829 + }, + { + "epoch": 0.9095890410958904, + "grad_norm": 1.592091679573059, + "learning_rate": 3.9033306255077175e-06, + "loss": 2.2501, + "step": 830 + }, + { + "epoch": 0.9095890410958904, + "eval_loss": 2.294229507446289, + "eval_runtime": 37.7195, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.113, + "step": 830 + }, + { + "epoch": 0.9106849315068493, + "grad_norm": 1.5698535442352295, + "learning_rate": 3.901299756295695e-06, + "loss": 2.1534, + "step": 831 + }, + { + "epoch": 0.9117808219178082, + "grad_norm": 1.7159922122955322, + "learning_rate": 3.899268887083672e-06, + "loss": 2.1601, + "step": 832 + }, + { + "epoch": 0.9128767123287671, + "grad_norm": 1.5601210594177246, + "learning_rate": 3.897238017871649e-06, + "loss": 2.1598, + "step": 833 + }, + { + "epoch": 0.913972602739726, + "grad_norm": 1.5987435579299927, + "learning_rate": 3.8952071486596266e-06, + "loss": 2.1307, + "step": 834 + }, + { + "epoch": 0.915068493150685, + "grad_norm": 1.7568409442901611, + "learning_rate": 3.893176279447604e-06, + "loss": 2.2209, + "step": 835 + }, + { + "epoch": 0.9161643835616439, + "grad_norm": 1.5044634342193604, + "learning_rate": 3.891145410235581e-06, + "loss": 2.1699, + "step": 836 + }, + { + "epoch": 0.9172602739726028, + "grad_norm": 1.7287421226501465, + "learning_rate": 3.889114541023558e-06, + "loss": 2.1489, + "step": 837 + }, + { + "epoch": 0.9183561643835616, + "grad_norm": 1.6595203876495361, + "learning_rate": 3.887083671811536e-06, + "loss": 2.1167, + "step": 838 + }, + { + "epoch": 0.9194520547945205, + "grad_norm": 1.5620248317718506, + "learning_rate": 3.885052802599513e-06, + "loss": 2.2119, + "step": 839 + }, + { + "epoch": 0.9205479452054794, + "grad_norm": 1.6492353677749634, + "learning_rate": 3.88302193338749e-06, + "loss": 2.1725, + "step": 840 + }, + { + "epoch": 0.9205479452054794, + "eval_loss": 2.2956349849700928, + "eval_runtime": 37.779, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 840 + }, + { + "epoch": 0.9216438356164384, + "grad_norm": 1.6601520776748657, + "learning_rate": 3.8809910641754675e-06, + "loss": 2.1532, + "step": 841 + }, + { + "epoch": 0.9227397260273973, + "grad_norm": 1.6723994016647339, + "learning_rate": 3.878960194963445e-06, + "loss": 2.1931, + "step": 842 + }, + { + "epoch": 0.9238356164383562, + "grad_norm": 1.7594473361968994, + "learning_rate": 3.876929325751422e-06, + "loss": 2.1219, + "step": 843 + }, + { + "epoch": 0.924931506849315, + "grad_norm": 1.546398401260376, + "learning_rate": 3.874898456539399e-06, + "loss": 2.1679, + "step": 844 + }, + { + "epoch": 0.9260273972602739, + "grad_norm": 1.8171741962432861, + "learning_rate": 3.872867587327377e-06, + "loss": 2.1434, + "step": 845 + }, + { + "epoch": 0.9271232876712329, + "grad_norm": 1.703196406364441, + "learning_rate": 3.870836718115354e-06, + "loss": 2.1574, + "step": 846 + }, + { + "epoch": 0.9282191780821918, + "grad_norm": 1.6290628910064697, + "learning_rate": 3.868805848903331e-06, + "loss": 2.1657, + "step": 847 + }, + { + "epoch": 0.9293150684931507, + "grad_norm": 1.5868806838989258, + "learning_rate": 3.8667749796913084e-06, + "loss": 2.1659, + "step": 848 + }, + { + "epoch": 0.9304109589041096, + "grad_norm": 1.551581621170044, + "learning_rate": 3.864744110479286e-06, + "loss": 2.1916, + "step": 849 + }, + { + "epoch": 0.9315068493150684, + "grad_norm": 1.6634963750839233, + "learning_rate": 3.862713241267262e-06, + "loss": 2.2027, + "step": 850 + }, + { + "epoch": 0.9315068493150684, + "eval_loss": 2.2988946437835693, + "eval_runtime": 37.7841, + "eval_samples_per_second": 26.466, + "eval_steps_per_second": 1.112, + "step": 850 + }, + { + "epoch": 0.9326027397260274, + "grad_norm": 1.6138510704040527, + "learning_rate": 3.86068237205524e-06, + "loss": 2.1634, + "step": 851 + }, + { + "epoch": 0.9336986301369863, + "grad_norm": 1.6294615268707275, + "learning_rate": 3.858651502843217e-06, + "loss": 2.1175, + "step": 852 + }, + { + "epoch": 0.9347945205479452, + "grad_norm": 1.7515549659729004, + "learning_rate": 3.856620633631195e-06, + "loss": 2.1562, + "step": 853 + }, + { + "epoch": 0.9358904109589041, + "grad_norm": 1.6304186582565308, + "learning_rate": 3.854589764419171e-06, + "loss": 2.2256, + "step": 854 + }, + { + "epoch": 0.936986301369863, + "grad_norm": 1.6189597845077515, + "learning_rate": 3.852558895207149e-06, + "loss": 2.1742, + "step": 855 + }, + { + "epoch": 0.938082191780822, + "grad_norm": 1.665515661239624, + "learning_rate": 3.850528025995126e-06, + "loss": 2.1481, + "step": 856 + }, + { + "epoch": 0.9391780821917808, + "grad_norm": 1.7066419124603271, + "learning_rate": 3.848497156783104e-06, + "loss": 2.1522, + "step": 857 + }, + { + "epoch": 0.9402739726027397, + "grad_norm": 1.884599208831787, + "learning_rate": 3.84646628757108e-06, + "loss": 2.1493, + "step": 858 + }, + { + "epoch": 0.9413698630136986, + "grad_norm": 1.6224873065948486, + "learning_rate": 3.8444354183590585e-06, + "loss": 2.1461, + "step": 859 + }, + { + "epoch": 0.9424657534246575, + "grad_norm": 1.916416049003601, + "learning_rate": 3.842404549147035e-06, + "loss": 2.1265, + "step": 860 + }, + { + "epoch": 0.9424657534246575, + "eval_loss": 2.297154426574707, + "eval_runtime": 37.77, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 860 + }, + { + "epoch": 0.9435616438356165, + "grad_norm": 1.5281857252120972, + "learning_rate": 3.840373679935013e-06, + "loss": 2.1476, + "step": 861 + }, + { + "epoch": 0.9446575342465754, + "grad_norm": 1.7905207872390747, + "learning_rate": 3.8383428107229894e-06, + "loss": 2.1421, + "step": 862 + }, + { + "epoch": 0.9457534246575342, + "grad_norm": 1.8282849788665771, + "learning_rate": 3.836311941510967e-06, + "loss": 2.1807, + "step": 863 + }, + { + "epoch": 0.9468493150684931, + "grad_norm": 1.7861522436141968, + "learning_rate": 3.834281072298944e-06, + "loss": 2.1337, + "step": 864 + }, + { + "epoch": 0.947945205479452, + "grad_norm": 1.6107250452041626, + "learning_rate": 3.832250203086921e-06, + "loss": 2.1911, + "step": 865 + }, + { + "epoch": 0.949041095890411, + "grad_norm": 1.657909631729126, + "learning_rate": 3.8302193338748985e-06, + "loss": 2.2257, + "step": 866 + }, + { + "epoch": 0.9501369863013699, + "grad_norm": 1.8969722986221313, + "learning_rate": 3.828188464662876e-06, + "loss": 2.1778, + "step": 867 + }, + { + "epoch": 0.9512328767123288, + "grad_norm": 1.6498360633850098, + "learning_rate": 3.826157595450853e-06, + "loss": 2.127, + "step": 868 + }, + { + "epoch": 0.9523287671232876, + "grad_norm": 1.5750585794448853, + "learning_rate": 3.82412672623883e-06, + "loss": 2.1716, + "step": 869 + }, + { + "epoch": 0.9534246575342465, + "grad_norm": 1.6047886610031128, + "learning_rate": 3.822095857026808e-06, + "loss": 2.1354, + "step": 870 + }, + { + "epoch": 0.9534246575342465, + "eval_loss": 2.295785665512085, + "eval_runtime": 37.7419, + "eval_samples_per_second": 26.496, + "eval_steps_per_second": 1.113, + "step": 870 + }, + { + "epoch": 0.9545205479452055, + "grad_norm": 1.6426187753677368, + "learning_rate": 3.820064987814785e-06, + "loss": 2.1375, + "step": 871 + }, + { + "epoch": 0.9556164383561644, + "grad_norm": 1.5118632316589355, + "learning_rate": 3.818034118602762e-06, + "loss": 2.1491, + "step": 872 + }, + { + "epoch": 0.9567123287671233, + "grad_norm": 1.630476474761963, + "learning_rate": 3.8160032493907395e-06, + "loss": 2.1497, + "step": 873 + }, + { + "epoch": 0.9578082191780822, + "grad_norm": 1.6820838451385498, + "learning_rate": 3.813972380178717e-06, + "loss": 2.1746, + "step": 874 + }, + { + "epoch": 0.958904109589041, + "grad_norm": 1.5480482578277588, + "learning_rate": 3.811941510966694e-06, + "loss": 2.1226, + "step": 875 + }, + { + "epoch": 0.96, + "grad_norm": 1.601493239402771, + "learning_rate": 3.809910641754671e-06, + "loss": 2.0934, + "step": 876 + }, + { + "epoch": 0.9610958904109589, + "grad_norm": 1.5756622552871704, + "learning_rate": 3.8078797725426486e-06, + "loss": 2.1679, + "step": 877 + }, + { + "epoch": 0.9621917808219178, + "grad_norm": 1.6811070442199707, + "learning_rate": 3.8058489033306254e-06, + "loss": 2.1407, + "step": 878 + }, + { + "epoch": 0.9632876712328767, + "grad_norm": 1.5748940706253052, + "learning_rate": 3.803818034118603e-06, + "loss": 2.1985, + "step": 879 + }, + { + "epoch": 0.9643835616438357, + "grad_norm": 1.5412342548370361, + "learning_rate": 3.80178716490658e-06, + "loss": 2.1876, + "step": 880 + }, + { + "epoch": 0.9643835616438357, + "eval_loss": 2.291630268096924, + "eval_runtime": 37.7702, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 880 + }, + { + "epoch": 0.9654794520547946, + "grad_norm": 1.7080684900283813, + "learning_rate": 3.7997562956945577e-06, + "loss": 2.1453, + "step": 881 + }, + { + "epoch": 0.9665753424657534, + "grad_norm": 1.8705110549926758, + "learning_rate": 3.7977254264825345e-06, + "loss": 2.1298, + "step": 882 + }, + { + "epoch": 0.9676712328767123, + "grad_norm": 1.7332537174224854, + "learning_rate": 3.7956945572705122e-06, + "loss": 2.2117, + "step": 883 + }, + { + "epoch": 0.9687671232876712, + "grad_norm": 1.88144052028656, + "learning_rate": 3.793663688058489e-06, + "loss": 2.2005, + "step": 884 + }, + { + "epoch": 0.9698630136986301, + "grad_norm": 1.6263943910598755, + "learning_rate": 3.791632818846467e-06, + "loss": 2.1853, + "step": 885 + }, + { + "epoch": 0.9709589041095891, + "grad_norm": 1.5968067646026611, + "learning_rate": 3.7896019496344436e-06, + "loss": 2.1998, + "step": 886 + }, + { + "epoch": 0.972054794520548, + "grad_norm": 1.7149827480316162, + "learning_rate": 3.787571080422421e-06, + "loss": 2.1855, + "step": 887 + }, + { + "epoch": 0.9731506849315068, + "grad_norm": 1.7375026941299438, + "learning_rate": 3.785540211210398e-06, + "loss": 2.1664, + "step": 888 + }, + { + "epoch": 0.9742465753424657, + "grad_norm": 1.6411423683166504, + "learning_rate": 3.7835093419983755e-06, + "loss": 2.1492, + "step": 889 + }, + { + "epoch": 0.9753424657534246, + "grad_norm": 1.6208266019821167, + "learning_rate": 3.7814784727863527e-06, + "loss": 2.2341, + "step": 890 + }, + { + "epoch": 0.9753424657534246, + "eval_loss": 2.2929985523223877, + "eval_runtime": 37.6693, + "eval_samples_per_second": 26.547, + "eval_steps_per_second": 1.115, + "step": 890 + }, + { + "epoch": 0.9764383561643836, + "grad_norm": 1.6871188879013062, + "learning_rate": 3.77944760357433e-06, + "loss": 2.1985, + "step": 891 + }, + { + "epoch": 0.9775342465753425, + "grad_norm": 1.8475587368011475, + "learning_rate": 3.7774167343623073e-06, + "loss": 2.1423, + "step": 892 + }, + { + "epoch": 0.9786301369863014, + "grad_norm": 1.6115065813064575, + "learning_rate": 3.7753858651502846e-06, + "loss": 2.1944, + "step": 893 + }, + { + "epoch": 0.9797260273972602, + "grad_norm": 1.6232231855392456, + "learning_rate": 3.773354995938262e-06, + "loss": 2.1751, + "step": 894 + }, + { + "epoch": 0.9808219178082191, + "grad_norm": 1.8564099073410034, + "learning_rate": 3.771324126726239e-06, + "loss": 2.2021, + "step": 895 + }, + { + "epoch": 0.9819178082191781, + "grad_norm": 1.7270197868347168, + "learning_rate": 3.7692932575142164e-06, + "loss": 2.1351, + "step": 896 + }, + { + "epoch": 0.983013698630137, + "grad_norm": 1.7187668085098267, + "learning_rate": 3.7672623883021937e-06, + "loss": 2.0856, + "step": 897 + }, + { + "epoch": 0.9841095890410959, + "grad_norm": 1.7766387462615967, + "learning_rate": 3.765231519090171e-06, + "loss": 2.168, + "step": 898 + }, + { + "epoch": 0.9852054794520548, + "grad_norm": 1.8063374757766724, + "learning_rate": 3.7632006498781482e-06, + "loss": 2.1317, + "step": 899 + }, + { + "epoch": 0.9863013698630136, + "grad_norm": 1.7601871490478516, + "learning_rate": 3.761169780666125e-06, + "loss": 2.1742, + "step": 900 + }, + { + "epoch": 0.9863013698630136, + "eval_loss": 2.292574167251587, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 900 + }, + { + "epoch": 0.9873972602739726, + "grad_norm": 1.8501280546188354, + "learning_rate": 3.7591389114541028e-06, + "loss": 2.1856, + "step": 901 + }, + { + "epoch": 0.9884931506849315, + "grad_norm": 1.8786728382110596, + "learning_rate": 3.7571080422420796e-06, + "loss": 2.1432, + "step": 902 + }, + { + "epoch": 0.9895890410958904, + "grad_norm": 1.7662389278411865, + "learning_rate": 3.7550771730300573e-06, + "loss": 2.2011, + "step": 903 + }, + { + "epoch": 0.9906849315068493, + "grad_norm": 1.6810210943222046, + "learning_rate": 3.753046303818034e-06, + "loss": 2.1719, + "step": 904 + }, + { + "epoch": 0.9917808219178083, + "grad_norm": 2.034982681274414, + "learning_rate": 3.751015434606012e-06, + "loss": 2.2206, + "step": 905 + }, + { + "epoch": 0.9928767123287672, + "grad_norm": 1.804049015045166, + "learning_rate": 3.7489845653939887e-06, + "loss": 2.1248, + "step": 906 + }, + { + "epoch": 0.993972602739726, + "grad_norm": 1.6039106845855713, + "learning_rate": 3.7469536961819664e-06, + "loss": 2.1545, + "step": 907 + }, + { + "epoch": 0.9950684931506849, + "grad_norm": 1.9352415800094604, + "learning_rate": 3.7449228269699433e-06, + "loss": 2.1842, + "step": 908 + }, + { + "epoch": 0.9961643835616438, + "grad_norm": 1.5759649276733398, + "learning_rate": 3.742891957757921e-06, + "loss": 2.1588, + "step": 909 + }, + { + "epoch": 0.9972602739726028, + "grad_norm": 1.6169486045837402, + "learning_rate": 3.740861088545898e-06, + "loss": 2.1871, + "step": 910 + }, + { + "epoch": 0.9972602739726028, + "eval_loss": 2.2923426628112793, + "eval_runtime": 37.4584, + "eval_samples_per_second": 26.696, + "eval_steps_per_second": 1.121, + "step": 910 + }, + { + "epoch": 0.9983561643835617, + "grad_norm": 1.748138427734375, + "learning_rate": 3.7388302193338755e-06, + "loss": 2.144, + "step": 911 + }, + { + "epoch": 0.9994520547945206, + "grad_norm": 1.8182991743087769, + "learning_rate": 3.7367993501218524e-06, + "loss": 2.1474, + "step": 912 + }, + { + "epoch": 1.0005479452054795, + "grad_norm": 1.6532196998596191, + "learning_rate": 3.7347684809098297e-06, + "loss": 2.1146, + "step": 913 + }, + { + "epoch": 1.0016438356164383, + "grad_norm": 2.0338973999023438, + "learning_rate": 3.732737611697807e-06, + "loss": 2.0851, + "step": 914 + }, + { + "epoch": 1.0027397260273974, + "grad_norm": 1.71133553981781, + "learning_rate": 3.7307067424857842e-06, + "loss": 2.1197, + "step": 915 + }, + { + "epoch": 1.0038356164383562, + "grad_norm": 1.6081736087799072, + "learning_rate": 3.7286758732737615e-06, + "loss": 2.1106, + "step": 916 + }, + { + "epoch": 1.004931506849315, + "grad_norm": 2.100734233856201, + "learning_rate": 3.7266450040617388e-06, + "loss": 2.2125, + "step": 917 + }, + { + "epoch": 1.006027397260274, + "grad_norm": 1.758898377418518, + "learning_rate": 3.724614134849716e-06, + "loss": 2.0985, + "step": 918 + }, + { + "epoch": 1.0071232876712328, + "grad_norm": 1.7335586547851562, + "learning_rate": 3.7225832656376933e-06, + "loss": 2.1911, + "step": 919 + }, + { + "epoch": 1.0082191780821919, + "grad_norm": 1.7879023551940918, + "learning_rate": 3.7205523964256706e-06, + "loss": 2.1632, + "step": 920 + }, + { + "epoch": 1.0082191780821919, + "eval_loss": 2.294053554534912, + "eval_runtime": 37.6132, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 920 + }, + { + "epoch": 1.0093150684931507, + "grad_norm": 1.7201987504959106, + "learning_rate": 3.718521527213648e-06, + "loss": 2.1501, + "step": 921 + }, + { + "epoch": 1.0104109589041095, + "grad_norm": 1.7638986110687256, + "learning_rate": 3.716490658001625e-06, + "loss": 2.0661, + "step": 922 + }, + { + "epoch": 1.0115068493150685, + "grad_norm": 1.7376288175582886, + "learning_rate": 3.7144597887896024e-06, + "loss": 2.1613, + "step": 923 + }, + { + "epoch": 1.0126027397260273, + "grad_norm": 1.9208155870437622, + "learning_rate": 3.7124289195775797e-06, + "loss": 2.1562, + "step": 924 + }, + { + "epoch": 1.0136986301369864, + "grad_norm": 1.6709166765213013, + "learning_rate": 3.710398050365557e-06, + "loss": 2.1309, + "step": 925 + }, + { + "epoch": 1.0147945205479452, + "grad_norm": 1.7070904970169067, + "learning_rate": 3.708367181153534e-06, + "loss": 2.1538, + "step": 926 + }, + { + "epoch": 1.0158904109589042, + "grad_norm": 1.7832821607589722, + "learning_rate": 3.7063363119415115e-06, + "loss": 2.1262, + "step": 927 + }, + { + "epoch": 1.016986301369863, + "grad_norm": 1.8703486919403076, + "learning_rate": 3.7043054427294884e-06, + "loss": 2.1142, + "step": 928 + }, + { + "epoch": 1.0180821917808218, + "grad_norm": 1.692858338356018, + "learning_rate": 3.702274573517466e-06, + "loss": 2.1182, + "step": 929 + }, + { + "epoch": 1.0191780821917809, + "grad_norm": 1.7156164646148682, + "learning_rate": 3.700243704305443e-06, + "loss": 2.1189, + "step": 930 + }, + { + "epoch": 1.0191780821917809, + "eval_loss": 2.2969048023223877, + "eval_runtime": 37.6673, + "eval_samples_per_second": 26.548, + "eval_steps_per_second": 1.115, + "step": 930 + }, + { + "epoch": 1.0202739726027397, + "grad_norm": 1.7443044185638428, + "learning_rate": 3.6982128350934206e-06, + "loss": 2.1058, + "step": 931 + }, + { + "epoch": 1.0213698630136987, + "grad_norm": 2.00412917137146, + "learning_rate": 3.6961819658813975e-06, + "loss": 2.2071, + "step": 932 + }, + { + "epoch": 1.0224657534246575, + "grad_norm": 1.7170531749725342, + "learning_rate": 3.694151096669375e-06, + "loss": 2.1254, + "step": 933 + }, + { + "epoch": 1.0235616438356165, + "grad_norm": 1.679018497467041, + "learning_rate": 3.692120227457352e-06, + "loss": 2.1487, + "step": 934 + }, + { + "epoch": 1.0246575342465754, + "grad_norm": 1.8477305173873901, + "learning_rate": 3.6900893582453297e-06, + "loss": 2.062, + "step": 935 + }, + { + "epoch": 1.0257534246575342, + "grad_norm": 1.664095401763916, + "learning_rate": 3.6880584890333066e-06, + "loss": 2.1418, + "step": 936 + }, + { + "epoch": 1.0268493150684932, + "grad_norm": 1.8831602334976196, + "learning_rate": 3.6860276198212834e-06, + "loss": 2.1483, + "step": 937 + }, + { + "epoch": 1.027945205479452, + "grad_norm": 1.7625365257263184, + "learning_rate": 3.683996750609261e-06, + "loss": 2.146, + "step": 938 + }, + { + "epoch": 1.029041095890411, + "grad_norm": 1.8168137073516846, + "learning_rate": 3.681965881397238e-06, + "loss": 2.1594, + "step": 939 + }, + { + "epoch": 1.0301369863013699, + "grad_norm": 1.6803343296051025, + "learning_rate": 3.6799350121852157e-06, + "loss": 2.1168, + "step": 940 + }, + { + "epoch": 1.0301369863013699, + "eval_loss": 2.2941982746124268, + "eval_runtime": 37.7112, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 940 + }, + { + "epoch": 1.0312328767123287, + "grad_norm": 2.0829122066497803, + "learning_rate": 3.6779041429731925e-06, + "loss": 2.1399, + "step": 941 + }, + { + "epoch": 1.0323287671232877, + "grad_norm": 1.6813585758209229, + "learning_rate": 3.6758732737611702e-06, + "loss": 2.1312, + "step": 942 + }, + { + "epoch": 1.0334246575342465, + "grad_norm": 1.7511613368988037, + "learning_rate": 3.673842404549147e-06, + "loss": 2.1273, + "step": 943 + }, + { + "epoch": 1.0345205479452055, + "grad_norm": 1.8264555931091309, + "learning_rate": 3.671811535337125e-06, + "loss": 2.1343, + "step": 944 + }, + { + "epoch": 1.0356164383561643, + "grad_norm": 1.7675082683563232, + "learning_rate": 3.6697806661251016e-06, + "loss": 2.2153, + "step": 945 + }, + { + "epoch": 1.0367123287671234, + "grad_norm": 1.6589823961257935, + "learning_rate": 3.6677497969130793e-06, + "loss": 2.1575, + "step": 946 + }, + { + "epoch": 1.0378082191780822, + "grad_norm": 1.6730010509490967, + "learning_rate": 3.665718927701056e-06, + "loss": 2.1773, + "step": 947 + }, + { + "epoch": 1.038904109589041, + "grad_norm": 1.627983570098877, + "learning_rate": 3.663688058489034e-06, + "loss": 2.1394, + "step": 948 + }, + { + "epoch": 1.04, + "grad_norm": 1.6363394260406494, + "learning_rate": 3.6616571892770107e-06, + "loss": 2.0828, + "step": 949 + }, + { + "epoch": 1.0410958904109588, + "grad_norm": 1.6270133256912231, + "learning_rate": 3.659626320064988e-06, + "loss": 2.1308, + "step": 950 + }, + { + "epoch": 1.0410958904109588, + "eval_loss": 2.293911933898926, + "eval_runtime": 37.7095, + "eval_samples_per_second": 26.519, + "eval_steps_per_second": 1.114, + "step": 950 + }, + { + "epoch": 1.0421917808219179, + "grad_norm": 1.784549593925476, + "learning_rate": 3.6575954508529653e-06, + "loss": 2.1239, + "step": 951 + }, + { + "epoch": 1.0432876712328767, + "grad_norm": 1.7265311479568481, + "learning_rate": 3.6555645816409426e-06, + "loss": 2.1573, + "step": 952 + }, + { + "epoch": 1.0443835616438357, + "grad_norm": 1.8133327960968018, + "learning_rate": 3.65353371242892e-06, + "loss": 2.046, + "step": 953 + }, + { + "epoch": 1.0454794520547945, + "grad_norm": 1.7085450887680054, + "learning_rate": 3.651502843216897e-06, + "loss": 2.1471, + "step": 954 + }, + { + "epoch": 1.0465753424657533, + "grad_norm": 2.0099172592163086, + "learning_rate": 3.6494719740048744e-06, + "loss": 2.1193, + "step": 955 + }, + { + "epoch": 1.0476712328767124, + "grad_norm": 1.582085132598877, + "learning_rate": 3.6474411047928517e-06, + "loss": 2.1522, + "step": 956 + }, + { + "epoch": 1.0487671232876712, + "grad_norm": 1.7923028469085693, + "learning_rate": 3.645410235580829e-06, + "loss": 2.1358, + "step": 957 + }, + { + "epoch": 1.0498630136986302, + "grad_norm": 1.6636080741882324, + "learning_rate": 3.6433793663688062e-06, + "loss": 2.1263, + "step": 958 + }, + { + "epoch": 1.050958904109589, + "grad_norm": 1.681675672531128, + "learning_rate": 3.6413484971567835e-06, + "loss": 2.1369, + "step": 959 + }, + { + "epoch": 1.0520547945205478, + "grad_norm": 1.7007646560668945, + "learning_rate": 3.6393176279447608e-06, + "loss": 2.1606, + "step": 960 + }, + { + "epoch": 1.0520547945205478, + "eval_loss": 2.291552782058716, + "eval_runtime": 37.704, + "eval_samples_per_second": 26.522, + "eval_steps_per_second": 1.114, + "step": 960 + }, + { + "epoch": 1.0531506849315069, + "grad_norm": 1.632997989654541, + "learning_rate": 3.637286758732738e-06, + "loss": 2.1742, + "step": 961 + }, + { + "epoch": 1.0542465753424657, + "grad_norm": 1.6881494522094727, + "learning_rate": 3.6352558895207153e-06, + "loss": 2.1555, + "step": 962 + }, + { + "epoch": 1.0553424657534247, + "grad_norm": 1.6624116897583008, + "learning_rate": 3.633225020308692e-06, + "loss": 2.1278, + "step": 963 + }, + { + "epoch": 1.0564383561643835, + "grad_norm": 1.7851325273513794, + "learning_rate": 3.63119415109667e-06, + "loss": 2.1599, + "step": 964 + }, + { + "epoch": 1.0575342465753426, + "grad_norm": 1.7592188119888306, + "learning_rate": 3.6291632818846467e-06, + "loss": 2.0898, + "step": 965 + }, + { + "epoch": 1.0586301369863014, + "grad_norm": 1.9573521614074707, + "learning_rate": 3.6271324126726244e-06, + "loss": 2.0876, + "step": 966 + }, + { + "epoch": 1.0597260273972602, + "grad_norm": 1.6684988737106323, + "learning_rate": 3.6251015434606013e-06, + "loss": 2.1288, + "step": 967 + }, + { + "epoch": 1.0608219178082192, + "grad_norm": 2.0011937618255615, + "learning_rate": 3.623070674248579e-06, + "loss": 2.1166, + "step": 968 + }, + { + "epoch": 1.061917808219178, + "grad_norm": 1.7490673065185547, + "learning_rate": 3.621039805036556e-06, + "loss": 2.1532, + "step": 969 + }, + { + "epoch": 1.063013698630137, + "grad_norm": 1.756422758102417, + "learning_rate": 3.6190089358245335e-06, + "loss": 2.1486, + "step": 970 + }, + { + "epoch": 1.063013698630137, + "eval_loss": 2.29402494430542, + "eval_runtime": 37.7078, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 970 + }, + { + "epoch": 1.0641095890410959, + "grad_norm": 2.014153480529785, + "learning_rate": 3.6169780666125104e-06, + "loss": 2.1512, + "step": 971 + }, + { + "epoch": 1.0652054794520547, + "grad_norm": 1.6026027202606201, + "learning_rate": 3.614947197400488e-06, + "loss": 2.0787, + "step": 972 + }, + { + "epoch": 1.0663013698630137, + "grad_norm": 1.8149386644363403, + "learning_rate": 3.612916328188465e-06, + "loss": 2.1675, + "step": 973 + }, + { + "epoch": 1.0673972602739725, + "grad_norm": 1.697019338607788, + "learning_rate": 3.610885458976442e-06, + "loss": 2.1584, + "step": 974 + }, + { + "epoch": 1.0684931506849316, + "grad_norm": 1.7330286502838135, + "learning_rate": 3.6088545897644195e-06, + "loss": 2.1048, + "step": 975 + }, + { + "epoch": 1.0695890410958904, + "grad_norm": 1.650336742401123, + "learning_rate": 3.6068237205523963e-06, + "loss": 2.1078, + "step": 976 + }, + { + "epoch": 1.0706849315068494, + "grad_norm": 1.7817063331604004, + "learning_rate": 3.604792851340374e-06, + "loss": 2.1656, + "step": 977 + }, + { + "epoch": 1.0717808219178082, + "grad_norm": 1.818617820739746, + "learning_rate": 3.602761982128351e-06, + "loss": 2.1388, + "step": 978 + }, + { + "epoch": 1.072876712328767, + "grad_norm": 1.7448735237121582, + "learning_rate": 3.6007311129163286e-06, + "loss": 2.1022, + "step": 979 + }, + { + "epoch": 1.073972602739726, + "grad_norm": 1.8401628732681274, + "learning_rate": 3.5987002437043054e-06, + "loss": 2.1181, + "step": 980 + }, + { + "epoch": 1.073972602739726, + "eval_loss": 2.2948012351989746, + "eval_runtime": 37.6902, + "eval_samples_per_second": 26.532, + "eval_steps_per_second": 1.114, + "step": 980 + }, + { + "epoch": 1.0750684931506849, + "grad_norm": 1.979794979095459, + "learning_rate": 3.596669374492283e-06, + "loss": 2.1766, + "step": 981 + }, + { + "epoch": 1.076164383561644, + "grad_norm": 1.7657036781311035, + "learning_rate": 3.59463850528026e-06, + "loss": 2.0889, + "step": 982 + }, + { + "epoch": 1.0772602739726027, + "grad_norm": 1.8293733596801758, + "learning_rate": 3.5926076360682377e-06, + "loss": 2.159, + "step": 983 + }, + { + "epoch": 1.0783561643835617, + "grad_norm": 1.7979159355163574, + "learning_rate": 3.5905767668562145e-06, + "loss": 2.1391, + "step": 984 + }, + { + "epoch": 1.0794520547945206, + "grad_norm": 1.847598910331726, + "learning_rate": 3.5885458976441922e-06, + "loss": 2.1695, + "step": 985 + }, + { + "epoch": 1.0805479452054794, + "grad_norm": 1.608178734779358, + "learning_rate": 3.586515028432169e-06, + "loss": 2.1745, + "step": 986 + }, + { + "epoch": 1.0816438356164384, + "grad_norm": 1.8166873455047607, + "learning_rate": 3.5844841592201464e-06, + "loss": 2.0882, + "step": 987 + }, + { + "epoch": 1.0827397260273972, + "grad_norm": 1.847463607788086, + "learning_rate": 3.5824532900081237e-06, + "loss": 2.1244, + "step": 988 + }, + { + "epoch": 1.0838356164383562, + "grad_norm": 1.6194696426391602, + "learning_rate": 3.580422420796101e-06, + "loss": 2.1398, + "step": 989 + }, + { + "epoch": 1.084931506849315, + "grad_norm": 1.9055372476577759, + "learning_rate": 3.578391551584078e-06, + "loss": 2.0774, + "step": 990 + }, + { + "epoch": 1.084931506849315, + "eval_loss": 2.2899444103240967, + "eval_runtime": 37.7237, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 990 + }, + { + "epoch": 1.086027397260274, + "grad_norm": 1.6451077461242676, + "learning_rate": 3.5763606823720555e-06, + "loss": 2.1351, + "step": 991 + }, + { + "epoch": 1.087123287671233, + "grad_norm": 1.8144199848175049, + "learning_rate": 3.5743298131600328e-06, + "loss": 2.1536, + "step": 992 + }, + { + "epoch": 1.0882191780821917, + "grad_norm": 1.760130763053894, + "learning_rate": 3.57229894394801e-06, + "loss": 2.1485, + "step": 993 + }, + { + "epoch": 1.0893150684931507, + "grad_norm": 1.7544556856155396, + "learning_rate": 3.5702680747359873e-06, + "loss": 2.1136, + "step": 994 + }, + { + "epoch": 1.0904109589041096, + "grad_norm": 1.9884154796600342, + "learning_rate": 3.5682372055239646e-06, + "loss": 2.1366, + "step": 995 + }, + { + "epoch": 1.0915068493150686, + "grad_norm": 1.72066330909729, + "learning_rate": 3.566206336311942e-06, + "loss": 2.1103, + "step": 996 + }, + { + "epoch": 1.0926027397260274, + "grad_norm": 1.807808756828308, + "learning_rate": 3.564175467099919e-06, + "loss": 2.0905, + "step": 997 + }, + { + "epoch": 1.0936986301369862, + "grad_norm": 2.1075172424316406, + "learning_rate": 3.5621445978878964e-06, + "loss": 2.1427, + "step": 998 + }, + { + "epoch": 1.0947945205479452, + "grad_norm": 1.7037009000778198, + "learning_rate": 3.5601137286758737e-06, + "loss": 2.1124, + "step": 999 + }, + { + "epoch": 1.095890410958904, + "grad_norm": 1.842721700668335, + "learning_rate": 3.5580828594638505e-06, + "loss": 2.1379, + "step": 1000 + }, + { + "epoch": 1.095890410958904, + "eval_loss": 2.291067361831665, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 1000 + }, + { + "epoch": 1.096986301369863, + "grad_norm": 1.802979588508606, + "learning_rate": 3.5560519902518282e-06, + "loss": 2.156, + "step": 1001 + }, + { + "epoch": 1.098082191780822, + "grad_norm": 1.699763298034668, + "learning_rate": 3.554021121039805e-06, + "loss": 2.1353, + "step": 1002 + }, + { + "epoch": 1.099178082191781, + "grad_norm": 1.695544958114624, + "learning_rate": 3.551990251827783e-06, + "loss": 2.1039, + "step": 1003 + }, + { + "epoch": 1.1002739726027397, + "grad_norm": 1.9509766101837158, + "learning_rate": 3.5499593826157596e-06, + "loss": 2.1201, + "step": 1004 + }, + { + "epoch": 1.1013698630136985, + "grad_norm": 1.7657231092453003, + "learning_rate": 3.5479285134037373e-06, + "loss": 2.1149, + "step": 1005 + }, + { + "epoch": 1.1024657534246576, + "grad_norm": 1.7416653633117676, + "learning_rate": 3.545897644191714e-06, + "loss": 2.1116, + "step": 1006 + }, + { + "epoch": 1.1035616438356164, + "grad_norm": 1.804185390472412, + "learning_rate": 3.543866774979692e-06, + "loss": 2.1426, + "step": 1007 + }, + { + "epoch": 1.1046575342465754, + "grad_norm": 1.8011984825134277, + "learning_rate": 3.5418359057676687e-06, + "loss": 2.1132, + "step": 1008 + }, + { + "epoch": 1.1057534246575342, + "grad_norm": 1.9039791822433472, + "learning_rate": 3.5398050365556464e-06, + "loss": 2.1479, + "step": 1009 + }, + { + "epoch": 1.106849315068493, + "grad_norm": 1.9513524770736694, + "learning_rate": 3.5377741673436233e-06, + "loss": 2.1093, + "step": 1010 + }, + { + "epoch": 1.106849315068493, + "eval_loss": 2.290522575378418, + "eval_runtime": 37.4617, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 1010 + }, + { + "epoch": 1.107945205479452, + "grad_norm": 1.7262980937957764, + "learning_rate": 3.535743298131601e-06, + "loss": 2.138, + "step": 1011 + }, + { + "epoch": 1.1090410958904109, + "grad_norm": 1.9514356851577759, + "learning_rate": 3.533712428919578e-06, + "loss": 2.1669, + "step": 1012 + }, + { + "epoch": 1.11013698630137, + "grad_norm": 1.7947099208831787, + "learning_rate": 3.5316815597075547e-06, + "loss": 2.105, + "step": 1013 + }, + { + "epoch": 1.1112328767123287, + "grad_norm": 1.8680298328399658, + "learning_rate": 3.5296506904955324e-06, + "loss": 2.1306, + "step": 1014 + }, + { + "epoch": 1.1123287671232878, + "grad_norm": 2.121732711791992, + "learning_rate": 3.5276198212835093e-06, + "loss": 2.1681, + "step": 1015 + }, + { + "epoch": 1.1134246575342466, + "grad_norm": 1.662564754486084, + "learning_rate": 3.525588952071487e-06, + "loss": 2.1681, + "step": 1016 + }, + { + "epoch": 1.1145205479452054, + "grad_norm": 1.8358488082885742, + "learning_rate": 3.523558082859464e-06, + "loss": 2.1363, + "step": 1017 + }, + { + "epoch": 1.1156164383561644, + "grad_norm": 1.8742910623550415, + "learning_rate": 3.5215272136474415e-06, + "loss": 2.1179, + "step": 1018 + }, + { + "epoch": 1.1167123287671232, + "grad_norm": 1.8780633211135864, + "learning_rate": 3.5194963444354184e-06, + "loss": 2.103, + "step": 1019 + }, + { + "epoch": 1.1178082191780823, + "grad_norm": 1.6979540586471558, + "learning_rate": 3.517465475223396e-06, + "loss": 2.0885, + "step": 1020 + }, + { + "epoch": 1.1178082191780823, + "eval_loss": 2.2939751148223877, + "eval_runtime": 37.66, + "eval_samples_per_second": 26.553, + "eval_steps_per_second": 1.115, + "step": 1020 + }, + { + "epoch": 1.118904109589041, + "grad_norm": 2.494971752166748, + "learning_rate": 3.515434606011373e-06, + "loss": 2.1582, + "step": 1021 + }, + { + "epoch": 1.12, + "grad_norm": 1.9183858633041382, + "learning_rate": 3.5134037367993506e-06, + "loss": 2.189, + "step": 1022 + }, + { + "epoch": 1.121095890410959, + "grad_norm": 1.9055982828140259, + "learning_rate": 3.5113728675873275e-06, + "loss": 2.1503, + "step": 1023 + }, + { + "epoch": 1.1221917808219177, + "grad_norm": 2.118046522140503, + "learning_rate": 3.5093419983753047e-06, + "loss": 2.1414, + "step": 1024 + }, + { + "epoch": 1.1232876712328768, + "grad_norm": 1.8413280248641968, + "learning_rate": 3.507311129163282e-06, + "loss": 2.09, + "step": 1025 + }, + { + "epoch": 1.1243835616438356, + "grad_norm": 2.0261356830596924, + "learning_rate": 3.5052802599512593e-06, + "loss": 2.132, + "step": 1026 + }, + { + "epoch": 1.1254794520547946, + "grad_norm": 1.8393104076385498, + "learning_rate": 3.5032493907392366e-06, + "loss": 2.1366, + "step": 1027 + }, + { + "epoch": 1.1265753424657534, + "grad_norm": 2.0282111167907715, + "learning_rate": 3.501218521527214e-06, + "loss": 2.0758, + "step": 1028 + }, + { + "epoch": 1.1276712328767124, + "grad_norm": 1.8438650369644165, + "learning_rate": 3.499187652315191e-06, + "loss": 2.1294, + "step": 1029 + }, + { + "epoch": 1.1287671232876713, + "grad_norm": 1.890739917755127, + "learning_rate": 3.4971567831031684e-06, + "loss": 2.1591, + "step": 1030 + }, + { + "epoch": 1.1287671232876713, + "eval_loss": 2.2894623279571533, + "eval_runtime": 37.6941, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1030 + }, + { + "epoch": 1.12986301369863, + "grad_norm": 1.877964973449707, + "learning_rate": 3.4951259138911457e-06, + "loss": 2.0823, + "step": 1031 + }, + { + "epoch": 1.130958904109589, + "grad_norm": 1.7460334300994873, + "learning_rate": 3.493095044679123e-06, + "loss": 2.1758, + "step": 1032 + }, + { + "epoch": 1.132054794520548, + "grad_norm": 1.668459415435791, + "learning_rate": 3.4910641754671002e-06, + "loss": 2.0461, + "step": 1033 + }, + { + "epoch": 1.133150684931507, + "grad_norm": 1.9560495615005493, + "learning_rate": 3.4890333062550775e-06, + "loss": 2.1394, + "step": 1034 + }, + { + "epoch": 1.1342465753424658, + "grad_norm": 1.820432424545288, + "learning_rate": 3.4870024370430548e-06, + "loss": 2.0735, + "step": 1035 + }, + { + "epoch": 1.1353424657534246, + "grad_norm": 1.7064971923828125, + "learning_rate": 3.484971567831032e-06, + "loss": 2.1196, + "step": 1036 + }, + { + "epoch": 1.1364383561643836, + "grad_norm": 1.837594747543335, + "learning_rate": 3.482940698619009e-06, + "loss": 2.1562, + "step": 1037 + }, + { + "epoch": 1.1375342465753424, + "grad_norm": 1.7570106983184814, + "learning_rate": 3.4809098294069866e-06, + "loss": 2.1253, + "step": 1038 + }, + { + "epoch": 1.1386301369863014, + "grad_norm": 1.7931634187698364, + "learning_rate": 3.4788789601949634e-06, + "loss": 2.1369, + "step": 1039 + }, + { + "epoch": 1.1397260273972603, + "grad_norm": 1.957241177558899, + "learning_rate": 3.476848090982941e-06, + "loss": 2.1204, + "step": 1040 + }, + { + "epoch": 1.1397260273972603, + "eval_loss": 2.2888705730438232, + "eval_runtime": 37.7002, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 1040 + }, + { + "epoch": 1.140821917808219, + "grad_norm": 1.7801663875579834, + "learning_rate": 3.474817221770918e-06, + "loss": 2.2008, + "step": 1041 + }, + { + "epoch": 1.141917808219178, + "grad_norm": 1.654398798942566, + "learning_rate": 3.4727863525588957e-06, + "loss": 2.1451, + "step": 1042 + }, + { + "epoch": 1.143013698630137, + "grad_norm": 1.8888044357299805, + "learning_rate": 3.4707554833468725e-06, + "loss": 2.1318, + "step": 1043 + }, + { + "epoch": 1.144109589041096, + "grad_norm": 1.8398077487945557, + "learning_rate": 3.4687246141348502e-06, + "loss": 2.1051, + "step": 1044 + }, + { + "epoch": 1.1452054794520548, + "grad_norm": 1.6971104145050049, + "learning_rate": 3.466693744922827e-06, + "loss": 2.119, + "step": 1045 + }, + { + "epoch": 1.1463013698630138, + "grad_norm": 1.605700969696045, + "learning_rate": 3.464662875710805e-06, + "loss": 2.1343, + "step": 1046 + }, + { + "epoch": 1.1473972602739726, + "grad_norm": 1.8513860702514648, + "learning_rate": 3.4626320064987817e-06, + "loss": 2.1331, + "step": 1047 + }, + { + "epoch": 1.1484931506849314, + "grad_norm": 1.7160199880599976, + "learning_rate": 3.4606011372867594e-06, + "loss": 2.1367, + "step": 1048 + }, + { + "epoch": 1.1495890410958904, + "grad_norm": 1.5926575660705566, + "learning_rate": 3.458570268074736e-06, + "loss": 2.1183, + "step": 1049 + }, + { + "epoch": 1.1506849315068493, + "grad_norm": 1.663843035697937, + "learning_rate": 3.4565393988627135e-06, + "loss": 2.1107, + "step": 1050 + }, + { + "epoch": 1.1506849315068493, + "eval_loss": 2.286076545715332, + "eval_runtime": 37.6944, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1050 + }, + { + "epoch": 1.1517808219178083, + "grad_norm": 1.7697075605392456, + "learning_rate": 3.4545085296506908e-06, + "loss": 2.0912, + "step": 1051 + }, + { + "epoch": 1.152876712328767, + "grad_norm": 1.626094937324524, + "learning_rate": 3.452477660438668e-06, + "loss": 2.1508, + "step": 1052 + }, + { + "epoch": 1.1539726027397261, + "grad_norm": 1.9503792524337769, + "learning_rate": 3.4504467912266453e-06, + "loss": 2.0819, + "step": 1053 + }, + { + "epoch": 1.155068493150685, + "grad_norm": 1.7748280763626099, + "learning_rate": 3.4484159220146226e-06, + "loss": 2.1233, + "step": 1054 + }, + { + "epoch": 1.1561643835616437, + "grad_norm": 2.099339008331299, + "learning_rate": 3.4463850528026e-06, + "loss": 2.1245, + "step": 1055 + }, + { + "epoch": 1.1572602739726028, + "grad_norm": 1.6200031042099, + "learning_rate": 3.444354183590577e-06, + "loss": 2.1437, + "step": 1056 + }, + { + "epoch": 1.1583561643835616, + "grad_norm": 1.7703361511230469, + "learning_rate": 3.4423233143785544e-06, + "loss": 2.0874, + "step": 1057 + }, + { + "epoch": 1.1594520547945206, + "grad_norm": 1.6505805253982544, + "learning_rate": 3.4402924451665317e-06, + "loss": 2.0655, + "step": 1058 + }, + { + "epoch": 1.1605479452054794, + "grad_norm": 1.7199327945709229, + "learning_rate": 3.438261575954509e-06, + "loss": 2.1383, + "step": 1059 + }, + { + "epoch": 1.1616438356164385, + "grad_norm": 1.772292137145996, + "learning_rate": 3.4362307067424862e-06, + "loss": 2.1673, + "step": 1060 + }, + { + "epoch": 1.1616438356164385, + "eval_loss": 2.2869999408721924, + "eval_runtime": 37.7398, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1060 + }, + { + "epoch": 1.1627397260273973, + "grad_norm": 1.5849649906158447, + "learning_rate": 3.434199837530463e-06, + "loss": 2.1331, + "step": 1061 + }, + { + "epoch": 1.163835616438356, + "grad_norm": 1.8389405012130737, + "learning_rate": 3.432168968318441e-06, + "loss": 2.0832, + "step": 1062 + }, + { + "epoch": 1.1649315068493151, + "grad_norm": 1.7142471075057983, + "learning_rate": 3.4301380991064176e-06, + "loss": 2.1447, + "step": 1063 + }, + { + "epoch": 1.166027397260274, + "grad_norm": 1.7853225469589233, + "learning_rate": 3.4281072298943953e-06, + "loss": 2.1591, + "step": 1064 + }, + { + "epoch": 1.167123287671233, + "grad_norm": 1.8228695392608643, + "learning_rate": 3.426076360682372e-06, + "loss": 2.1684, + "step": 1065 + }, + { + "epoch": 1.1682191780821918, + "grad_norm": 1.7090480327606201, + "learning_rate": 3.42404549147035e-06, + "loss": 2.1339, + "step": 1066 + }, + { + "epoch": 1.1693150684931506, + "grad_norm": 1.642682433128357, + "learning_rate": 3.4220146222583267e-06, + "loss": 2.075, + "step": 1067 + }, + { + "epoch": 1.1704109589041096, + "grad_norm": 1.8230397701263428, + "learning_rate": 3.4199837530463044e-06, + "loss": 2.1128, + "step": 1068 + }, + { + "epoch": 1.1715068493150684, + "grad_norm": 1.761135458946228, + "learning_rate": 3.4179528838342813e-06, + "loss": 2.1885, + "step": 1069 + }, + { + "epoch": 1.1726027397260275, + "grad_norm": 1.7829216718673706, + "learning_rate": 3.415922014622259e-06, + "loss": 2.1584, + "step": 1070 + }, + { + "epoch": 1.1726027397260275, + "eval_loss": 2.2878382205963135, + "eval_runtime": 37.7433, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1070 + }, + { + "epoch": 1.1736986301369863, + "grad_norm": 1.5480386018753052, + "learning_rate": 3.413891145410236e-06, + "loss": 2.1602, + "step": 1071 + }, + { + "epoch": 1.174794520547945, + "grad_norm": 1.8957836627960205, + "learning_rate": 3.4118602761982135e-06, + "loss": 2.1706, + "step": 1072 + }, + { + "epoch": 1.1758904109589041, + "grad_norm": 1.924424648284912, + "learning_rate": 3.4098294069861904e-06, + "loss": 2.1339, + "step": 1073 + }, + { + "epoch": 1.176986301369863, + "grad_norm": 1.660370111465454, + "learning_rate": 3.4077985377741673e-06, + "loss": 2.1175, + "step": 1074 + }, + { + "epoch": 1.178082191780822, + "grad_norm": 1.7255185842514038, + "learning_rate": 3.405767668562145e-06, + "loss": 2.1672, + "step": 1075 + }, + { + "epoch": 1.1791780821917808, + "grad_norm": 1.8028507232666016, + "learning_rate": 3.403736799350122e-06, + "loss": 2.1527, + "step": 1076 + }, + { + "epoch": 1.1802739726027398, + "grad_norm": 1.7100828886032104, + "learning_rate": 3.4017059301380995e-06, + "loss": 2.1031, + "step": 1077 + }, + { + "epoch": 1.1813698630136986, + "grad_norm": 1.785955548286438, + "learning_rate": 3.3996750609260764e-06, + "loss": 2.1829, + "step": 1078 + }, + { + "epoch": 1.1824657534246574, + "grad_norm": 1.8583980798721313, + "learning_rate": 3.397644191714054e-06, + "loss": 2.1198, + "step": 1079 + }, + { + "epoch": 1.1835616438356165, + "grad_norm": 1.8848942518234253, + "learning_rate": 3.395613322502031e-06, + "loss": 2.123, + "step": 1080 + }, + { + "epoch": 1.1835616438356165, + "eval_loss": 2.288412094116211, + "eval_runtime": 37.7637, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 1080 + }, + { + "epoch": 1.1846575342465753, + "grad_norm": 1.6747279167175293, + "learning_rate": 3.3935824532900086e-06, + "loss": 2.1087, + "step": 1081 + }, + { + "epoch": 1.1857534246575343, + "grad_norm": 1.8993490934371948, + "learning_rate": 3.3915515840779855e-06, + "loss": 2.1316, + "step": 1082 + }, + { + "epoch": 1.1868493150684931, + "grad_norm": 1.7047377824783325, + "learning_rate": 3.389520714865963e-06, + "loss": 2.1242, + "step": 1083 + }, + { + "epoch": 1.1879452054794521, + "grad_norm": 1.767811894416809, + "learning_rate": 3.38748984565394e-06, + "loss": 2.1556, + "step": 1084 + }, + { + "epoch": 1.189041095890411, + "grad_norm": 1.665313720703125, + "learning_rate": 3.3854589764419177e-06, + "loss": 2.101, + "step": 1085 + }, + { + "epoch": 1.1901369863013698, + "grad_norm": 1.7752090692520142, + "learning_rate": 3.3834281072298946e-06, + "loss": 2.1307, + "step": 1086 + }, + { + "epoch": 1.1912328767123288, + "grad_norm": 1.9759247303009033, + "learning_rate": 3.381397238017872e-06, + "loss": 2.1093, + "step": 1087 + }, + { + "epoch": 1.1923287671232876, + "grad_norm": 1.8772896528244019, + "learning_rate": 3.379366368805849e-06, + "loss": 2.1851, + "step": 1088 + }, + { + "epoch": 1.1934246575342466, + "grad_norm": 1.7442713975906372, + "learning_rate": 3.3773354995938264e-06, + "loss": 2.1433, + "step": 1089 + }, + { + "epoch": 1.1945205479452055, + "grad_norm": 1.8106931447982788, + "learning_rate": 3.3753046303818037e-06, + "loss": 2.1246, + "step": 1090 + }, + { + "epoch": 1.1945205479452055, + "eval_loss": 2.2858333587646484, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1090 + }, + { + "epoch": 1.1956164383561645, + "grad_norm": 1.7938610315322876, + "learning_rate": 3.373273761169781e-06, + "loss": 2.1325, + "step": 1091 + }, + { + "epoch": 1.1967123287671233, + "grad_norm": 1.5760858058929443, + "learning_rate": 3.3712428919577582e-06, + "loss": 2.1036, + "step": 1092 + }, + { + "epoch": 1.1978082191780821, + "grad_norm": 1.8173094987869263, + "learning_rate": 3.3692120227457355e-06, + "loss": 2.1509, + "step": 1093 + }, + { + "epoch": 1.1989041095890411, + "grad_norm": 2.0316812992095947, + "learning_rate": 3.3671811535337128e-06, + "loss": 2.0883, + "step": 1094 + }, + { + "epoch": 1.2, + "grad_norm": 1.7735658884048462, + "learning_rate": 3.36515028432169e-06, + "loss": 2.1038, + "step": 1095 + }, + { + "epoch": 1.201095890410959, + "grad_norm": 1.752718210220337, + "learning_rate": 3.3631194151096673e-06, + "loss": 2.1393, + "step": 1096 + }, + { + "epoch": 1.2021917808219178, + "grad_norm": 1.778838872909546, + "learning_rate": 3.3610885458976446e-06, + "loss": 2.1641, + "step": 1097 + }, + { + "epoch": 1.2032876712328768, + "grad_norm": 1.7422083616256714, + "learning_rate": 3.359057676685622e-06, + "loss": 2.1624, + "step": 1098 + }, + { + "epoch": 1.2043835616438356, + "grad_norm": 1.7196860313415527, + "learning_rate": 3.357026807473599e-06, + "loss": 2.0845, + "step": 1099 + }, + { + "epoch": 1.2054794520547945, + "grad_norm": 1.8373215198516846, + "learning_rate": 3.354995938261576e-06, + "loss": 2.1092, + "step": 1100 + }, + { + "epoch": 1.2054794520547945, + "eval_loss": 2.285470485687256, + "eval_runtime": 37.728, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 1100 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 3.429040350298112e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/latest new file mode 100644 index 0000000000000000000000000000000000000000..873b08355ae55e303606c474ab5fba25f9e8262a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/latest @@ -0,0 +1 @@ +global_step1200 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..0099c3c310be709431f1e2425c4ec538e89ac967 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/trainer_state.json @@ -0,0 +1,9393 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.3150684931506849, + "eval_steps": 10, + "global_step": 1200, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9794018268585205, + "learning_rate": 1.551094890510949e-06, + "loss": 2.3047, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.396454334259033, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.3671, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 2.0826210975646973, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.4105, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.3590970039367676, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.3405, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 2.3945844173431396, + "learning_rate": 1.624087591240876e-06, + "loss": 2.2922, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 2.185126781463623, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.3399, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.2687461376190186, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.3547, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.920499324798584, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.28, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.127774715423584, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.3409, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 1.8677122592926025, + "learning_rate": 1.715328467153285e-06, + "loss": 2.3406, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.4506173133850098, + "eval_runtime": 37.5008, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9803308248519897, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.3642, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 1.7733794450759888, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.3666, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 1.7639747858047485, + "learning_rate": 1.77007299270073e-06, + "loss": 2.3926, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 1.9172815084457397, + "learning_rate": 1.788321167883212e-06, + "loss": 2.3431, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 1.580621361732483, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.2967, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.0889124870300293, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.331, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.6659903526306152, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.3181, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.028651237487793, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.3237, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.681212306022644, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.32, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 1.9986696243286133, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.3425, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.4408469200134277, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 1.832582950592041, + "learning_rate": 1.916058394160584e-06, + "loss": 2.2548, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 1.8609144687652588, + "learning_rate": 1.934306569343066e-06, + "loss": 2.3024, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 1.8699817657470703, + "learning_rate": 1.952554744525548e-06, + "loss": 2.3044, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 1.81354558467865, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.3098, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 1.7789452075958252, + "learning_rate": 1.989051094890511e-06, + "loss": 2.3623, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 1.6322330236434937, + "learning_rate": 2.007299270072993e-06, + "loss": 2.303, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 1.8742916584014893, + "learning_rate": 2.025547445255475e-06, + "loss": 2.3183, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.7790557146072388, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.3168, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 1.7624462842941284, + "learning_rate": 2.062043795620438e-06, + "loss": 2.3629, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.781114101409912, + "learning_rate": 2.08029197080292e-06, + "loss": 2.2813, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.426335334777832, + "eval_runtime": 37.7146, + "eval_samples_per_second": 26.515, + "eval_steps_per_second": 1.114, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.769988775253296, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.3396, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.0640275478363037, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.3688, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.7884094715118408, + "learning_rate": 2.135036496350365e-06, + "loss": 2.2775, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.831847071647644, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.2966, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9044525623321533, + "learning_rate": 2.171532846715329e-06, + "loss": 2.3359, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.9899096488952637, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.3099, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8570183515548706, + "learning_rate": 2.208029197080292e-06, + "loss": 2.3016, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 1.8255733251571655, + "learning_rate": 2.226277372262774e-06, + "loss": 2.335, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.8759729862213135, + "learning_rate": 2.244525547445256e-06, + "loss": 2.3357, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.9031389951705933, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.3014, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.4213266372680664, + "eval_runtime": 37.6757, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.1044118404388428, + "learning_rate": 2.281021897810219e-06, + "loss": 2.2946, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.005735158920288, + "learning_rate": 2.299270072992701e-06, + "loss": 2.3132, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.8821312189102173, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.2816, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7866829633712769, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.309, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.8452801704406738, + "learning_rate": 2.354014598540146e-06, + "loss": 2.3222, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.5781638622283936, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.2688, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 1.9746729135513306, + "learning_rate": 2.39051094890511e-06, + "loss": 2.2864, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 2.0312447547912598, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.2914, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.675674557685852, + "learning_rate": 2.427007299270073e-06, + "loss": 2.3441, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 1.8987798690795898, + "learning_rate": 2.445255474452555e-06, + "loss": 2.2589, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.4149818420410156, + "eval_runtime": 37.6627, + "eval_samples_per_second": 26.551, + "eval_steps_per_second": 1.115, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 2.041752338409424, + "learning_rate": 2.463503649635037e-06, + "loss": 2.2457, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.7845569849014282, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.3376, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 2.489330530166626, + "learning_rate": 2.5e-06, + "loss": 2.258, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 2.3793296813964844, + "learning_rate": 2.518248175182482e-06, + "loss": 2.2561, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 2.0474460124969482, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.2992, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.7181849479675293, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.3003, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 2.3854947090148926, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.2866, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.7228457927703857, + "learning_rate": 2.591240875912409e-06, + "loss": 2.3027, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.3928253650665283, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.3233, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.136934757232666, + "learning_rate": 2.627737226277373e-06, + "loss": 2.2949, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.410449266433716, + "eval_runtime": 37.6837, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.8906159400939941, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.2693, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 1.8668863773345947, + "learning_rate": 2.664233576642336e-06, + "loss": 2.2944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.319363594055176, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.2791, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0869688987731934, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.32, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 1.8910208940505981, + "learning_rate": 2.718978102189781e-06, + "loss": 2.3059, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.157984495162964, + "learning_rate": 2.737226277372263e-06, + "loss": 2.2756, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.387314558029175, + "learning_rate": 2.755474452554745e-06, + "loss": 2.3597, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.183333158493042, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.322, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.158015489578247, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.2663, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 1.9594391584396362, + "learning_rate": 2.81021897810219e-06, + "loss": 2.3, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.411637306213379, + "eval_runtime": 37.6428, + "eval_samples_per_second": 26.565, + "eval_steps_per_second": 1.116, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.28957462310791, + "learning_rate": 2.828467153284672e-06, + "loss": 2.2258, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.8085683584213257, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.2739, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 1.9662787914276123, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.2398, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 2.337609052658081, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.3216, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 1.6008479595184326, + "learning_rate": 2.901459854014599e-06, + "loss": 2.2997, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 2.177690029144287, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.2893, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8746743202209473, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.3303, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 1.7862998247146606, + "learning_rate": 2.956204379562044e-06, + "loss": 2.3007, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.7578285932540894, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.2591, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.1532702445983887, + "learning_rate": 2.992700729927008e-06, + "loss": 2.2544, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.3992533683776855, + "eval_runtime": 37.6571, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.0168659687042236, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.3108, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 1.8336135149002075, + "learning_rate": 3.029197080291971e-06, + "loss": 2.3044, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.6675384044647217, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.2978, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 1.8866558074951172, + "learning_rate": 3.065693430656935e-06, + "loss": 2.3169, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0588760375976562, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.3041, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.9300280809402466, + "learning_rate": 3.102189781021898e-06, + "loss": 2.2494, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 1.837496042251587, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.2339, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.155080795288086, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.324, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 1.9562510251998901, + "learning_rate": 3.156934306569343e-06, + "loss": 2.2937, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 2.522388219833374, + "learning_rate": 3.175182481751825e-06, + "loss": 2.2683, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.3936216831207275, + "eval_runtime": 37.6499, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 1.7409396171569824, + "learning_rate": 3.193430656934307e-06, + "loss": 2.2833, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.176687002182007, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.2748, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 1.8576542139053345, + "learning_rate": 3.2299270072992707e-06, + "loss": 2.2144, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.9564870595932007, + "learning_rate": 3.248175182481752e-06, + "loss": 2.2898, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.896238088607788, + "learning_rate": 3.266423357664234e-06, + "loss": 2.2793, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.735960841178894, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.2869, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.018303394317627, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.2775, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 1.7448145151138306, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.2736, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8245689868927002, + "learning_rate": 3.339416058394161e-06, + "loss": 2.2799, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 1.821118950843811, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.2823, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.3893656730651855, + "eval_runtime": 37.676, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 1.9112446308135986, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.287, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 1.6978867053985596, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.2865, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9376300573349, + "learning_rate": 3.412408759124088e-06, + "loss": 2.2326, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.8067563772201538, + "learning_rate": 3.43065693430657e-06, + "loss": 2.2739, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.8476370573043823, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.2328, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.927923321723938, + "learning_rate": 3.467153284671533e-06, + "loss": 2.2679, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7956453561782837, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.2607, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 1.857413649559021, + "learning_rate": 3.503649635036497e-06, + "loss": 2.2854, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 1.8384093046188354, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.3021, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.7346371412277222, + "learning_rate": 3.54014598540146e-06, + "loss": 2.3402, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.390045642852783, + "eval_runtime": 37.4358, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.7546677589416504, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.3214, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 1.7437881231307983, + "learning_rate": 3.576642335766424e-06, + "loss": 2.2004, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9509227275848389, + "learning_rate": 3.5948905109489054e-06, + "loss": 2.2044, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 1.6918139457702637, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.2369, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9664885997772217, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.254, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 1.7942848205566406, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.2296, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 1.8610197305679321, + "learning_rate": 3.667883211678833e-06, + "loss": 2.2273, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 1.8405956029891968, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.3034, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 1.7151877880096436, + "learning_rate": 3.704379562043796e-06, + "loss": 2.2689, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 1.8795795440673828, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.2363, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.384481430053711, + "eval_runtime": 37.6512, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 1.7494690418243408, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.257, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 1.7700639963150024, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.2457, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 1.846092700958252, + "learning_rate": 3.777372262773723e-06, + "loss": 2.2425, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 1.6699751615524292, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.2734, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 1.8978508710861206, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.2435, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.6387466192245483, + "learning_rate": 3.832116788321168e-06, + "loss": 2.2241, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 1.6999541521072388, + "learning_rate": 3.85036496350365e-06, + "loss": 2.2468, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 1.6322710514068604, + "learning_rate": 3.868613138686132e-06, + "loss": 2.2642, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.5633572340011597, + "learning_rate": 3.886861313868613e-06, + "loss": 2.2862, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 1.7642648220062256, + "learning_rate": 3.905109489051096e-06, + "loss": 2.2414, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.3775289058685303, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.76724374294281, + "learning_rate": 3.923357664233577e-06, + "loss": 2.2976, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.6049190759658813, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.3034, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.6591981649398804, + "learning_rate": 3.95985401459854e-06, + "loss": 2.1957, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.650676965713501, + "learning_rate": 3.978102189781022e-06, + "loss": 2.275, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.6665960550308228, + "learning_rate": 3.996350364963504e-06, + "loss": 2.1853, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.5785901546478271, + "learning_rate": 4.014598540145986e-06, + "loss": 2.2399, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.976162075996399, + "learning_rate": 4.032846715328467e-06, + "loss": 2.2499, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.642327904701233, + "learning_rate": 4.05109489051095e-06, + "loss": 2.235, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 2.0126261711120605, + "learning_rate": 4.069343065693431e-06, + "loss": 2.2466, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.6737161874771118, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.2984, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.375901222229004, + "eval_runtime": 37.7072, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 2.094257354736328, + "learning_rate": 4.105839416058395e-06, + "loss": 2.251, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.655393123626709, + "learning_rate": 4.124087591240876e-06, + "loss": 2.2366, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 2.5909082889556885, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.2986, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.8278532028198242, + "learning_rate": 4.16058394160584e-06, + "loss": 2.1977, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9404895305633545, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.2578, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.6525582075119019, + "learning_rate": 4.197080291970803e-06, + "loss": 2.2263, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 2.021822929382324, + "learning_rate": 4.215328467153285e-06, + "loss": 2.2326, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 1.9136033058166504, + "learning_rate": 4.233576642335767e-06, + "loss": 2.2457, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.389714002609253, + "learning_rate": 4.251824817518249e-06, + "loss": 2.2256, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.7886695861816406, + "learning_rate": 4.27007299270073e-06, + "loss": 2.2222, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.3729844093322754, + "eval_runtime": 37.7638, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.9499074220657349, + "learning_rate": 4.288321167883212e-06, + "loss": 2.2026, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 1.834797978401184, + "learning_rate": 4.306569343065693e-06, + "loss": 2.2564, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 2.0415048599243164, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.2462, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 1.6527886390686035, + "learning_rate": 4.343065693430658e-06, + "loss": 2.2493, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.0504889488220215, + "learning_rate": 4.361313868613139e-06, + "loss": 2.2734, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.004772424697876, + "learning_rate": 4.379562043795621e-06, + "loss": 2.2469, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.1973347663879395, + "learning_rate": 4.397810218978102e-06, + "loss": 2.2479, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.8222625255584717, + "learning_rate": 4.416058394160584e-06, + "loss": 2.2206, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 2.1759912967681885, + "learning_rate": 4.434306569343066e-06, + "loss": 2.261, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.8344624042510986, + "learning_rate": 4.452554744525548e-06, + "loss": 2.2605, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.3706893920898438, + "eval_runtime": 37.734, + "eval_samples_per_second": 26.501, + "eval_steps_per_second": 1.113, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.7783704996109009, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.284, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 2.101193428039551, + "learning_rate": 4.489051094890512e-06, + "loss": 2.2524, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 1.7923569679260254, + "learning_rate": 4.507299270072993e-06, + "loss": 2.2434, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 2.170883893966675, + "learning_rate": 4.525547445255475e-06, + "loss": 2.2173, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.6749192476272583, + "learning_rate": 4.543795620437956e-06, + "loss": 2.283, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.0484230518341064, + "learning_rate": 4.562043795620438e-06, + "loss": 2.2931, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.9212749004364014, + "learning_rate": 4.580291970802921e-06, + "loss": 2.2661, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.6729028224945068, + "learning_rate": 4.598540145985402e-06, + "loss": 2.2269, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 2.1190619468688965, + "learning_rate": 4.6167883211678835e-06, + "loss": 2.2278, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.7444697618484497, + "learning_rate": 4.635036496350365e-06, + "loss": 2.2634, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.3706092834472656, + "eval_runtime": 37.7391, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 1.8921526670455933, + "learning_rate": 4.653284671532847e-06, + "loss": 2.2697, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7347339391708374, + "learning_rate": 4.671532846715329e-06, + "loss": 2.2406, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 2.1617953777313232, + "learning_rate": 4.689781021897811e-06, + "loss": 2.2891, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.873384714126587, + "learning_rate": 4.708029197080292e-06, + "loss": 2.2554, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.801356554031372, + "learning_rate": 4.726277372262774e-06, + "loss": 2.2073, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.9250633716583252, + "learning_rate": 4.744525547445255e-06, + "loss": 2.2501, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 1.8360860347747803, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.2561, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 1.8711498975753784, + "learning_rate": 4.78102189781022e-06, + "loss": 2.2247, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.6604810953140259, + "learning_rate": 4.799270072992701e-06, + "loss": 2.2448, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 1.689201831817627, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.2494, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.3682427406311035, + "eval_runtime": 37.7233, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.720961093902588, + "learning_rate": 4.835766423357665e-06, + "loss": 2.2013, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.7863153219223022, + "learning_rate": 4.854014598540146e-06, + "loss": 2.2183, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 1.8583060503005981, + "learning_rate": 4.872262773722628e-06, + "loss": 2.2013, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 1.722428560256958, + "learning_rate": 4.89051094890511e-06, + "loss": 2.2399, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.8408993482589722, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.2489, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.887604832649231, + "learning_rate": 4.927007299270074e-06, + "loss": 2.2145, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 1.7112189531326294, + "learning_rate": 4.945255474452555e-06, + "loss": 2.2915, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 1.8314064741134644, + "learning_rate": 4.963503649635037e-06, + "loss": 2.2786, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.7001186609268188, + "learning_rate": 4.981751824817518e-06, + "loss": 2.249, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.5820505619049072, + "learning_rate": 5e-06, + "loss": 2.254, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.3624653816223145, + "eval_runtime": 37.7283, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 1.8516303300857544, + "learning_rate": 4.997969130787978e-06, + "loss": 2.1894, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 1.8591808080673218, + "learning_rate": 4.995938261575955e-06, + "loss": 2.2537, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.6822415590286255, + "learning_rate": 4.993907392363932e-06, + "loss": 2.1978, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 2.0779759883880615, + "learning_rate": 4.9918765231519095e-06, + "loss": 2.2286, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.6937508583068848, + "learning_rate": 4.989845653939887e-06, + "loss": 2.2339, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.9342437982559204, + "learning_rate": 4.987814784727864e-06, + "loss": 2.2145, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.773258924484253, + "learning_rate": 4.985783915515841e-06, + "loss": 2.2368, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.8196488618850708, + "learning_rate": 4.983753046303819e-06, + "loss": 2.2224, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.6786775588989258, + "learning_rate": 4.981722177091796e-06, + "loss": 2.2439, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.7761427164077759, + "learning_rate": 4.979691307879773e-06, + "loss": 2.236, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.3614282608032227, + "eval_runtime": 37.7314, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.8389159440994263, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.2745, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.7774744033813477, + "learning_rate": 4.975629569455728e-06, + "loss": 2.2122, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.7862017154693604, + "learning_rate": 4.973598700243705e-06, + "loss": 2.2097, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.8851722478866577, + "learning_rate": 4.971567831031681e-06, + "loss": 2.118, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.068608045578003, + "learning_rate": 4.9695369618196595e-06, + "loss": 2.1884, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.8885847330093384, + "learning_rate": 4.967506092607636e-06, + "loss": 2.2097, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.9441415071487427, + "learning_rate": 4.965475223395614e-06, + "loss": 2.2068, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 2.313577890396118, + "learning_rate": 4.9634443541835905e-06, + "loss": 2.2355, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 1.8542472124099731, + "learning_rate": 4.961413484971569e-06, + "loss": 2.2299, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9382838010787964, + "learning_rate": 4.959382615759545e-06, + "loss": 2.2072, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.361767053604126, + "eval_runtime": 37.4603, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.081242322921753, + "learning_rate": 4.957351746547523e-06, + "loss": 2.3167, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 1.9109065532684326, + "learning_rate": 4.9553208773355e-06, + "loss": 2.2336, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.124897003173828, + "learning_rate": 4.953290008123478e-06, + "loss": 2.184, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.8820433616638184, + "learning_rate": 4.951259138911454e-06, + "loss": 2.2836, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.016204595565796, + "learning_rate": 4.9492282696994315e-06, + "loss": 2.2221, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.0679776668548584, + "learning_rate": 4.947197400487409e-06, + "loss": 2.2113, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 1.8040891885757446, + "learning_rate": 4.945166531275386e-06, + "loss": 2.2292, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.0315871238708496, + "learning_rate": 4.943135662063363e-06, + "loss": 2.2657, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.71120023727417, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.2517, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0443193912506104, + "learning_rate": 4.939073923639318e-06, + "loss": 2.2105, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.3582265377044678, + "eval_runtime": 37.6749, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.9363837242126465, + "learning_rate": 4.937043054427295e-06, + "loss": 2.1944, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 2.0878167152404785, + "learning_rate": 4.935012185215272e-06, + "loss": 2.2477, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.9059010744094849, + "learning_rate": 4.93298131600325e-06, + "loss": 2.2278, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 2.220851421356201, + "learning_rate": 4.930950446791227e-06, + "loss": 2.2391, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.7555570602416992, + "learning_rate": 4.928919577579204e-06, + "loss": 2.2206, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 2.1546502113342285, + "learning_rate": 4.9268887083671815e-06, + "loss": 2.2216, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 1.9688911437988281, + "learning_rate": 4.924857839155159e-06, + "loss": 2.2423, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 1.983587622642517, + "learning_rate": 4.922826969943136e-06, + "loss": 2.1738, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 2.1990034580230713, + "learning_rate": 4.920796100731113e-06, + "loss": 2.2341, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 1.9690128564834595, + "learning_rate": 4.918765231519091e-06, + "loss": 2.1914, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.36171817779541, + "eval_runtime": 37.727, + "eval_samples_per_second": 26.506, + "eval_steps_per_second": 1.113, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 2.5950212478637695, + "learning_rate": 4.916734362307068e-06, + "loss": 2.1921, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 2.023130416870117, + "learning_rate": 4.914703493095045e-06, + "loss": 2.1932, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2486252784729004, + "learning_rate": 4.912672623883022e-06, + "loss": 2.2174, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 1.723677635192871, + "learning_rate": 4.910641754671e-06, + "loss": 2.2173, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.94509756565094, + "learning_rate": 4.908610885458977e-06, + "loss": 2.2539, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 1.6934322118759155, + "learning_rate": 4.906580016246954e-06, + "loss": 2.2345, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 1.6882652044296265, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.231, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.6730620861053467, + "learning_rate": 4.902518277822909e-06, + "loss": 2.2394, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.7261624336242676, + "learning_rate": 4.900487408610886e-06, + "loss": 2.2524, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.8310920000076294, + "learning_rate": 4.898456539398863e-06, + "loss": 2.3004, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.3536295890808105, + "eval_runtime": 37.7461, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.7025809288024902, + "learning_rate": 4.89642567018684e-06, + "loss": 2.1923, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.7285082340240479, + "learning_rate": 4.894394800974818e-06, + "loss": 2.241, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 1.7589807510375977, + "learning_rate": 4.892363931762794e-06, + "loss": 2.2019, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.6664931774139404, + "learning_rate": 4.8903330625507725e-06, + "loss": 2.1822, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 1.8308027982711792, + "learning_rate": 4.888302193338749e-06, + "loss": 2.2651, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.6426894664764404, + "learning_rate": 4.886271324126727e-06, + "loss": 2.2514, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.9765205383300781, + "learning_rate": 4.8842404549147034e-06, + "loss": 2.226, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.7872203588485718, + "learning_rate": 4.8822095857026816e-06, + "loss": 2.18, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 2.0645813941955566, + "learning_rate": 4.880178716490658e-06, + "loss": 2.2023, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 1.9823166131973267, + "learning_rate": 4.878147847278636e-06, + "loss": 2.1801, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.353776454925537, + "eval_runtime": 37.771, + "eval_samples_per_second": 26.475, + "eval_steps_per_second": 1.112, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.831152319908142, + "learning_rate": 4.8761169780666125e-06, + "loss": 2.1858, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 1.9321790933609009, + "learning_rate": 4.874086108854591e-06, + "loss": 2.1638, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.9097596406936646, + "learning_rate": 4.872055239642567e-06, + "loss": 2.2206, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8821815252304077, + "learning_rate": 4.870024370430544e-06, + "loss": 2.2338, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.7959749698638916, + "learning_rate": 4.867993501218522e-06, + "loss": 2.1882, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 1.625320315361023, + "learning_rate": 4.865962632006499e-06, + "loss": 2.2137, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 1.6914536952972412, + "learning_rate": 4.863931762794476e-06, + "loss": 2.2374, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0801584720611572, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.2517, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 1.6874804496765137, + "learning_rate": 4.859870024370431e-06, + "loss": 2.2621, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 1.9347832202911377, + "learning_rate": 4.857839155158408e-06, + "loss": 2.2326, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.3468236923217773, + "eval_runtime": 37.7184, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.6798334121704102, + "learning_rate": 4.855808285946385e-06, + "loss": 2.2222, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 1.8993884325027466, + "learning_rate": 4.8537774167343626e-06, + "loss": 2.2178, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 1.697892189025879, + "learning_rate": 4.85174654752234e-06, + "loss": 2.2217, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8484156131744385, + "learning_rate": 4.849715678310317e-06, + "loss": 2.2069, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 1.8125075101852417, + "learning_rate": 4.847684809098294e-06, + "loss": 2.2285, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 1.8215030431747437, + "learning_rate": 4.845653939886272e-06, + "loss": 2.2029, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 1.777129054069519, + "learning_rate": 4.843623070674249e-06, + "loss": 2.2512, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.7327351570129395, + "learning_rate": 4.841592201462226e-06, + "loss": 2.1767, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.7363202571868896, + "learning_rate": 4.8395613322502035e-06, + "loss": 2.2174, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9183942079544067, + "learning_rate": 4.837530463038181e-06, + "loss": 2.2383, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.345961332321167, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.7290767431259155, + "learning_rate": 4.835499593826158e-06, + "loss": 2.22, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 1.7892694473266602, + "learning_rate": 4.833468724614135e-06, + "loss": 2.1403, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.6671900749206543, + "learning_rate": 4.831437855402113e-06, + "loss": 2.1989, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 1.848751425743103, + "learning_rate": 4.82940698619009e-06, + "loss": 2.245, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.6957710981369019, + "learning_rate": 4.827376116978067e-06, + "loss": 2.2239, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.650583267211914, + "learning_rate": 4.8253452477660444e-06, + "loss": 2.2101, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.6510974168777466, + "learning_rate": 4.823314378554022e-06, + "loss": 2.204, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 1.9469026327133179, + "learning_rate": 4.821283509341998e-06, + "loss": 2.221, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 1.6031194925308228, + "learning_rate": 4.819252640129976e-06, + "loss": 2.1926, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.120325803756714, + "learning_rate": 4.817221770917953e-06, + "loss": 2.2166, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.3419017791748047, + "eval_runtime": 37.6854, + "eval_samples_per_second": 26.535, + "eval_steps_per_second": 1.114, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.5533722639083862, + "learning_rate": 4.815190901705931e-06, + "loss": 2.1981, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9650007486343384, + "learning_rate": 4.813160032493907e-06, + "loss": 2.2319, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 1.6874849796295166, + "learning_rate": 4.811129163281885e-06, + "loss": 2.2263, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.787996768951416, + "learning_rate": 4.809098294069862e-06, + "loss": 2.2546, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 1.7853440046310425, + "learning_rate": 4.80706742485784e-06, + "loss": 2.1994, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.7882641553878784, + "learning_rate": 4.805036555645816e-06, + "loss": 2.2449, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 1.8206003904342651, + "learning_rate": 4.8030056864337945e-06, + "loss": 2.2144, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.7012749910354614, + "learning_rate": 4.800974817221771e-06, + "loss": 2.2122, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 1.8711670637130737, + "learning_rate": 4.798943948009749e-06, + "loss": 2.2132, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 1.6880611181259155, + "learning_rate": 4.7969130787977254e-06, + "loss": 2.2336, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.340393543243408, + "eval_runtime": 37.6885, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 2.0543065071105957, + "learning_rate": 4.794882209585703e-06, + "loss": 2.1899, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 1.5774480104446411, + "learning_rate": 4.79285134037368e-06, + "loss": 2.1982, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 1.7154620885849, + "learning_rate": 4.790820471161657e-06, + "loss": 2.2096, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.8200151920318604, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.2391, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 1.6368298530578613, + "learning_rate": 4.786758732737612e-06, + "loss": 2.1855, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 1.6851340532302856, + "learning_rate": 4.784727863525589e-06, + "loss": 2.2114, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 1.8293818235397339, + "learning_rate": 4.782696994313566e-06, + "loss": 2.1636, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 1.6213513612747192, + "learning_rate": 4.780666125101544e-06, + "loss": 2.1884, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.7791825532913208, + "learning_rate": 4.778635255889521e-06, + "loss": 2.1977, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 1.6627994775772095, + "learning_rate": 4.776604386677498e-06, + "loss": 2.2444, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.3396637439727783, + "eval_runtime": 37.7069, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 1.7597277164459229, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.2369, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.7754322290420532, + "learning_rate": 4.772542648253453e-06, + "loss": 2.1859, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 1.7897601127624512, + "learning_rate": 4.77051177904143e-06, + "loss": 2.2245, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 1.7294282913208008, + "learning_rate": 4.768480909829407e-06, + "loss": 2.1998, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 1.8406974077224731, + "learning_rate": 4.766450040617385e-06, + "loss": 2.2353, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 1.849493384361267, + "learning_rate": 4.764419171405362e-06, + "loss": 2.2165, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.5568877458572388, + "learning_rate": 4.762388302193339e-06, + "loss": 2.1989, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 1.8816689252853394, + "learning_rate": 4.760357432981316e-06, + "loss": 2.2062, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 1.7648332118988037, + "learning_rate": 4.758326563769294e-06, + "loss": 2.2204, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.753944754600525, + "learning_rate": 4.756295694557271e-06, + "loss": 2.1748, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.339653253555298, + "eval_runtime": 37.4438, + "eval_samples_per_second": 26.707, + "eval_steps_per_second": 1.122, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.7343329191207886, + "learning_rate": 4.754264825345248e-06, + "loss": 2.1843, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 1.8125407695770264, + "learning_rate": 4.7522339561332255e-06, + "loss": 2.1492, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.5255184173583984, + "learning_rate": 4.750203086921203e-06, + "loss": 2.2178, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 1.6692545413970947, + "learning_rate": 4.74817221770918e-06, + "loss": 2.2516, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 1.7804819345474243, + "learning_rate": 4.7461413484971565e-06, + "loss": 2.1592, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8048933744430542, + "learning_rate": 4.744110479285135e-06, + "loss": 2.1869, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 1.6439207792282104, + "learning_rate": 4.742079610073111e-06, + "loss": 2.1995, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.7375173568725586, + "learning_rate": 4.740048740861089e-06, + "loss": 2.2067, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 1.9471659660339355, + "learning_rate": 4.738017871649066e-06, + "loss": 2.2513, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 1.6524564027786255, + "learning_rate": 4.735987002437044e-06, + "loss": 2.2643, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.337153434753418, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 1.8574563264846802, + "learning_rate": 4.73395613322502e-06, + "loss": 2.2373, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 1.836464524269104, + "learning_rate": 4.731925264012998e-06, + "loss": 2.1769, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 1.7012423276901245, + "learning_rate": 4.729894394800975e-06, + "loss": 2.1494, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 1.953180193901062, + "learning_rate": 4.727863525588953e-06, + "loss": 2.186, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 1.8468470573425293, + "learning_rate": 4.725832656376929e-06, + "loss": 2.2279, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 2.005077362060547, + "learning_rate": 4.723801787164907e-06, + "loss": 2.1798, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 1.8246952295303345, + "learning_rate": 4.721770917952885e-06, + "loss": 2.1916, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 1.8558839559555054, + "learning_rate": 4.719740048740861e-06, + "loss": 2.2237, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 1.7670999765396118, + "learning_rate": 4.717709179528839e-06, + "loss": 2.2179, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.8544219732284546, + "learning_rate": 4.715678310316816e-06, + "loss": 2.1811, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.333791732788086, + "eval_runtime": 37.7254, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 1.5792925357818604, + "learning_rate": 4.713647441104794e-06, + "loss": 2.1749, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 1.5678222179412842, + "learning_rate": 4.71161657189277e-06, + "loss": 2.2294, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 1.9467812776565552, + "learning_rate": 4.709585702680748e-06, + "loss": 2.2684, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 1.6647964715957642, + "learning_rate": 4.707554833468725e-06, + "loss": 2.1722, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 1.7037622928619385, + "learning_rate": 4.705523964256703e-06, + "loss": 2.1915, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.5857148170471191, + "learning_rate": 4.703493095044679e-06, + "loss": 2.1466, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 1.5792173147201538, + "learning_rate": 4.701462225832657e-06, + "loss": 2.2472, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 1.5998480319976807, + "learning_rate": 4.699431356620634e-06, + "loss": 2.2129, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.6728194952011108, + "learning_rate": 4.697400487408612e-06, + "loss": 2.2057, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 1.5402852296829224, + "learning_rate": 4.695369618196588e-06, + "loss": 2.1903, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.3325934410095215, + "eval_runtime": 37.7463, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 1.5890164375305176, + "learning_rate": 4.693338748984566e-06, + "loss": 2.1584, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 1.805923581123352, + "learning_rate": 4.691307879772543e-06, + "loss": 2.2263, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.5932700634002686, + "learning_rate": 4.68927701056052e-06, + "loss": 2.2109, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 1.5611685514450073, + "learning_rate": 4.6872461413484975e-06, + "loss": 2.192, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 1.6976194381713867, + "learning_rate": 4.685215272136475e-06, + "loss": 2.1998, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.7149630784988403, + "learning_rate": 4.683184402924452e-06, + "loss": 2.2599, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 1.7597423791885376, + "learning_rate": 4.681153533712429e-06, + "loss": 2.1765, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 1.539652943611145, + "learning_rate": 4.679122664500407e-06, + "loss": 2.1635, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.690402865409851, + "learning_rate": 4.677091795288384e-06, + "loss": 2.2417, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.7267673015594482, + "learning_rate": 4.675060926076361e-06, + "loss": 2.166, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.331728935241699, + "eval_runtime": 37.7133, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.6383073329925537, + "learning_rate": 4.673030056864338e-06, + "loss": 2.1624, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.5364140272140503, + "learning_rate": 4.670999187652316e-06, + "loss": 2.1696, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.591477870941162, + "learning_rate": 4.668968318440293e-06, + "loss": 2.2306, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 1.8121449947357178, + "learning_rate": 4.66693744922827e-06, + "loss": 2.1449, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.515539288520813, + "learning_rate": 4.6649065800162475e-06, + "loss": 2.2108, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 1.6521453857421875, + "learning_rate": 4.662875710804225e-06, + "loss": 2.1398, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 1.7074251174926758, + "learning_rate": 4.660844841592202e-06, + "loss": 2.2342, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.7688119411468506, + "learning_rate": 4.658813972380179e-06, + "loss": 2.1287, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.7170679569244385, + "learning_rate": 4.656783103168157e-06, + "loss": 2.2143, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 1.7401350736618042, + "learning_rate": 4.654752233956134e-06, + "loss": 2.1855, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.335416555404663, + "eval_runtime": 37.7115, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.6344956159591675, + "learning_rate": 4.652721364744111e-06, + "loss": 2.2205, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.629184603691101, + "learning_rate": 4.6506904955320885e-06, + "loss": 2.2271, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.769491195678711, + "learning_rate": 4.648659626320066e-06, + "loss": 2.1945, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.6513251066207886, + "learning_rate": 4.646628757108043e-06, + "loss": 2.1705, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9357609748840332, + "learning_rate": 4.6445978878960194e-06, + "loss": 2.1968, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 1.75972318649292, + "learning_rate": 4.6425670186839976e-06, + "loss": 2.2229, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 1.686574935913086, + "learning_rate": 4.640536149471974e-06, + "loss": 2.2361, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 1.655661702156067, + "learning_rate": 4.638505280259952e-06, + "loss": 2.2064, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.7583094835281372, + "learning_rate": 4.6364744110479285e-06, + "loss": 2.2352, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.5906383991241455, + "learning_rate": 4.634443541835907e-06, + "loss": 2.187, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.332923412322998, + "eval_runtime": 37.7496, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 1.708019733428955, + "learning_rate": 4.632412672623883e-06, + "loss": 2.1032, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 1.6350252628326416, + "learning_rate": 4.630381803411861e-06, + "loss": 2.1976, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.9037389755249023, + "learning_rate": 4.628350934199838e-06, + "loss": 2.224, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 1.7126816511154175, + "learning_rate": 4.626320064987816e-06, + "loss": 2.1925, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 1.5828406810760498, + "learning_rate": 4.624289195775792e-06, + "loss": 2.1928, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.6797102689743042, + "learning_rate": 4.62225832656377e-06, + "loss": 2.1902, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 1.6871609687805176, + "learning_rate": 4.620227457351747e-06, + "loss": 2.2024, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.8785804510116577, + "learning_rate": 4.618196588139724e-06, + "loss": 2.1778, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.6175488233566284, + "learning_rate": 4.616165718927701e-06, + "loss": 2.2438, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.7724019289016724, + "learning_rate": 4.6141348497156786e-06, + "loss": 2.2272, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.328923225402832, + "eval_runtime": 37.7443, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.7230926752090454, + "learning_rate": 4.612103980503656e-06, + "loss": 2.2021, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.7899917364120483, + "learning_rate": 4.610073111291633e-06, + "loss": 2.1854, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.5864417552947998, + "learning_rate": 4.60804224207961e-06, + "loss": 2.2042, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.875375509262085, + "learning_rate": 4.606011372867588e-06, + "loss": 2.1844, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 1.830597162246704, + "learning_rate": 4.603980503655565e-06, + "loss": 2.1474, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 1.8624461889266968, + "learning_rate": 4.601949634443542e-06, + "loss": 2.158, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.7161756753921509, + "learning_rate": 4.5999187652315195e-06, + "loss": 2.1838, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.6212890148162842, + "learning_rate": 4.597887896019497e-06, + "loss": 2.1911, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.8586416244506836, + "learning_rate": 4.595857026807474e-06, + "loss": 2.1794, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.653976559638977, + "learning_rate": 4.593826157595451e-06, + "loss": 2.2428, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.3279106616973877, + "eval_runtime": 37.739, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 2.054537057876587, + "learning_rate": 4.591795288383429e-06, + "loss": 2.2003, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.730942964553833, + "learning_rate": 4.589764419171406e-06, + "loss": 2.1816, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.7264728546142578, + "learning_rate": 4.587733549959383e-06, + "loss": 2.1615, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 1.8276281356811523, + "learning_rate": 4.5857026807473604e-06, + "loss": 2.1442, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 1.9947556257247925, + "learning_rate": 4.583671811535338e-06, + "loss": 2.2192, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 1.770601511001587, + "learning_rate": 4.581640942323315e-06, + "loss": 2.233, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 1.8946491479873657, + "learning_rate": 4.579610073111292e-06, + "loss": 2.2249, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 1.7839634418487549, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.2469, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.6298807859420776, + "learning_rate": 4.575548334687247e-06, + "loss": 2.1723, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 1.9310041666030884, + "learning_rate": 4.573517465475224e-06, + "loss": 2.2156, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.3249237537384033, + "eval_runtime": 37.7505, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 1.7901647090911865, + "learning_rate": 4.571486596263201e-06, + "loss": 2.2169, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.7739574909210205, + "learning_rate": 4.569455727051178e-06, + "loss": 2.2005, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 1.7310121059417725, + "learning_rate": 4.567424857839156e-06, + "loss": 2.2369, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.004469156265259, + "learning_rate": 4.565393988627132e-06, + "loss": 2.2051, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 1.7906666994094849, + "learning_rate": 4.5633631194151105e-06, + "loss": 2.1908, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 1.7224774360656738, + "learning_rate": 4.561332250203087e-06, + "loss": 2.1529, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 1.9950156211853027, + "learning_rate": 4.559301380991065e-06, + "loss": 2.1776, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 1.8257590532302856, + "learning_rate": 4.5572705117790414e-06, + "loss": 2.2267, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.038490056991577, + "learning_rate": 4.5552396425670196e-06, + "loss": 2.1703, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 1.7314448356628418, + "learning_rate": 4.553208773354996e-06, + "loss": 2.2409, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.327078104019165, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 2.0013606548309326, + "learning_rate": 4.551177904142974e-06, + "loss": 2.1732, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.7352690696716309, + "learning_rate": 4.5491470349309505e-06, + "loss": 2.2027, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9024865627288818, + "learning_rate": 4.547116165718929e-06, + "loss": 2.1806, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8870717287063599, + "learning_rate": 4.545085296506905e-06, + "loss": 2.1664, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.781226634979248, + "learning_rate": 4.543054427294882e-06, + "loss": 2.166, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.8994221687316895, + "learning_rate": 4.54102355808286e-06, + "loss": 2.1951, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.6999517679214478, + "learning_rate": 4.538992688870837e-06, + "loss": 2.1786, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.899590253829956, + "learning_rate": 4.536961819658814e-06, + "loss": 2.2051, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.8054970502853394, + "learning_rate": 4.5349309504467915e-06, + "loss": 2.1675, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.6732861995697021, + "learning_rate": 4.532900081234769e-06, + "loss": 2.1997, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.3242335319519043, + "eval_runtime": 37.6748, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 2.0063157081604004, + "learning_rate": 4.530869212022746e-06, + "loss": 2.1909, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.7750036716461182, + "learning_rate": 4.528838342810723e-06, + "loss": 2.1935, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9993029832839966, + "learning_rate": 4.526807473598701e-06, + "loss": 2.1753, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.6742994785308838, + "learning_rate": 4.524776604386678e-06, + "loss": 2.2205, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 1.9300428628921509, + "learning_rate": 4.522745735174655e-06, + "loss": 2.2081, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.8770017623901367, + "learning_rate": 4.520714865962632e-06, + "loss": 2.2074, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.847765326499939, + "learning_rate": 4.51868399675061e-06, + "loss": 2.1775, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.905117392539978, + "learning_rate": 4.516653127538587e-06, + "loss": 2.1511, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.7623670101165771, + "learning_rate": 4.514622258326564e-06, + "loss": 2.188, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.9232176542282104, + "learning_rate": 4.5125913891145415e-06, + "loss": 2.2109, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.3252475261688232, + "eval_runtime": 37.6923, + "eval_samples_per_second": 26.531, + "eval_steps_per_second": 1.114, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.6523535251617432, + "learning_rate": 4.510560519902519e-06, + "loss": 2.1499, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9934184551239014, + "learning_rate": 4.508529650690496e-06, + "loss": 2.1383, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.704522728919983, + "learning_rate": 4.506498781478473e-06, + "loss": 2.1892, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.9909989833831787, + "learning_rate": 4.504467912266451e-06, + "loss": 2.2422, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.7490800619125366, + "learning_rate": 4.502437043054428e-06, + "loss": 2.2008, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8620145320892334, + "learning_rate": 4.500406173842405e-06, + "loss": 2.1892, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.7069931030273438, + "learning_rate": 4.4983753046303824e-06, + "loss": 2.1816, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 2.152491807937622, + "learning_rate": 4.49634443541836e-06, + "loss": 2.2331, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.8526151180267334, + "learning_rate": 4.494313566206336e-06, + "loss": 2.1811, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.6865259408950806, + "learning_rate": 4.492282696994314e-06, + "loss": 2.1599, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.3265392780303955, + "eval_runtime": 37.7255, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.7710354328155518, + "learning_rate": 4.490251827782291e-06, + "loss": 2.1636, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.8444210290908813, + "learning_rate": 4.488220958570269e-06, + "loss": 2.1345, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.9095014333724976, + "learning_rate": 4.486190089358245e-06, + "loss": 2.2236, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.7406108379364014, + "learning_rate": 4.484159220146223e-06, + "loss": 2.2078, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.712534785270691, + "learning_rate": 4.4821283509342e-06, + "loss": 2.1809, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8921459913253784, + "learning_rate": 4.480097481722178e-06, + "loss": 2.158, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.805146336555481, + "learning_rate": 4.478066612510154e-06, + "loss": 2.1765, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 2.1620516777038574, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.2832, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.661618947982788, + "learning_rate": 4.474004874086109e-06, + "loss": 2.1397, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.732483983039856, + "learning_rate": 4.471974004874087e-06, + "loss": 2.225, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.3276846408843994, + "eval_runtime": 37.7156, + "eval_samples_per_second": 26.514, + "eval_steps_per_second": 1.114, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.8693495988845825, + "learning_rate": 4.4699431356620635e-06, + "loss": 2.1959, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8201690912246704, + "learning_rate": 4.467912266450041e-06, + "loss": 2.1882, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.9229968786239624, + "learning_rate": 4.465881397238018e-06, + "loss": 2.2006, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.6330770254135132, + "learning_rate": 4.463850528025995e-06, + "loss": 2.209, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9654744863510132, + "learning_rate": 4.4618196588139726e-06, + "loss": 2.1963, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.8808507919311523, + "learning_rate": 4.45978878960195e-06, + "loss": 2.1538, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 1.8481040000915527, + "learning_rate": 4.457757920389927e-06, + "loss": 2.1769, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 2.114439010620117, + "learning_rate": 4.455727051177904e-06, + "loss": 2.2241, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.5492331981658936, + "learning_rate": 4.453696181965882e-06, + "loss": 2.206, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 2.11787486076355, + "learning_rate": 4.451665312753859e-06, + "loss": 2.2328, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.3215034008026123, + "eval_runtime": 37.7471, + "eval_samples_per_second": 26.492, + "eval_steps_per_second": 1.113, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.6561832427978516, + "learning_rate": 4.449634443541836e-06, + "loss": 2.1607, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.7510966062545776, + "learning_rate": 4.4476035743298135e-06, + "loss": 2.1699, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.9424657821655273, + "learning_rate": 4.445572705117791e-06, + "loss": 2.2298, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 1.5836161375045776, + "learning_rate": 4.443541835905768e-06, + "loss": 2.1711, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.895216464996338, + "learning_rate": 4.441510966693745e-06, + "loss": 2.2252, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.3365776538848877, + "learning_rate": 4.439480097481723e-06, + "loss": 2.1309, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.6988815069198608, + "learning_rate": 4.4374492282697e-06, + "loss": 2.2023, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 2.228900671005249, + "learning_rate": 4.435418359057677e-06, + "loss": 2.2021, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.5536632537841797, + "learning_rate": 4.433387489845654e-06, + "loss": 2.1706, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.8463878631591797, + "learning_rate": 4.431356620633632e-06, + "loss": 2.1715, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.320664167404175, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.924741506576538, + "learning_rate": 4.429325751421609e-06, + "loss": 2.1894, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 1.5921828746795654, + "learning_rate": 4.427294882209586e-06, + "loss": 2.1358, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 2.4617693424224854, + "learning_rate": 4.4252640129975635e-06, + "loss": 2.1918, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.6239196062088013, + "learning_rate": 4.423233143785541e-06, + "loss": 2.1996, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.952112078666687, + "learning_rate": 4.421202274573518e-06, + "loss": 2.112, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.663409948348999, + "learning_rate": 4.4191714053614945e-06, + "loss": 2.149, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.831711769104004, + "learning_rate": 4.417140536149473e-06, + "loss": 2.209, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 1.9281738996505737, + "learning_rate": 4.415109666937449e-06, + "loss": 2.1637, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.5661600828170776, + "learning_rate": 4.413078797725427e-06, + "loss": 2.1832, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.6930522918701172, + "learning_rate": 4.411047928513404e-06, + "loss": 2.196, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.318861246109009, + "eval_runtime": 37.6994, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.74558424949646, + "learning_rate": 4.409017059301382e-06, + "loss": 2.2097, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.664880633354187, + "learning_rate": 4.406986190089358e-06, + "loss": 2.2076, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.5981497764587402, + "learning_rate": 4.404955320877336e-06, + "loss": 2.1706, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.6035112142562866, + "learning_rate": 4.402924451665313e-06, + "loss": 2.1538, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 1.6819255352020264, + "learning_rate": 4.400893582453291e-06, + "loss": 2.1816, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.7082457542419434, + "learning_rate": 4.398862713241267e-06, + "loss": 2.1804, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.6893151998519897, + "learning_rate": 4.396831844029245e-06, + "loss": 2.1783, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 1.8639923334121704, + "learning_rate": 4.394800974817222e-06, + "loss": 2.1989, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.6407119035720825, + "learning_rate": 4.392770105605199e-06, + "loss": 2.2013, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.6435848474502563, + "learning_rate": 4.390739236393176e-06, + "loss": 2.2049, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.318084478378296, + "eval_runtime": 37.6999, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.7974674701690674, + "learning_rate": 4.388708367181154e-06, + "loss": 2.188, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 1.8340431451797485, + "learning_rate": 4.386677497969131e-06, + "loss": 2.2305, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.55428946018219, + "learning_rate": 4.384646628757108e-06, + "loss": 2.1837, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.7258585691452026, + "learning_rate": 4.3826157595450855e-06, + "loss": 2.1426, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.7347694635391235, + "learning_rate": 4.380584890333063e-06, + "loss": 2.1161, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.6082992553710938, + "learning_rate": 4.37855402112104e-06, + "loss": 2.2284, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.9446425437927246, + "learning_rate": 4.376523151909017e-06, + "loss": 2.1571, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 1.8597826957702637, + "learning_rate": 4.3744922826969946e-06, + "loss": 2.1982, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.8589327335357666, + "learning_rate": 4.372461413484972e-06, + "loss": 2.2097, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.6825448274612427, + "learning_rate": 4.370430544272949e-06, + "loss": 2.2077, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.315889596939087, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 600 + }, + { + "epoch": 0.6586301369863014, + "grad_norm": 1.8007057905197144, + "learning_rate": 4.368399675060926e-06, + "loss": 2.1843, + "step": 601 + }, + { + "epoch": 0.6597260273972603, + "grad_norm": 2.057697057723999, + "learning_rate": 4.366368805848904e-06, + "loss": 2.13, + "step": 602 + }, + { + "epoch": 0.6608219178082192, + "grad_norm": 1.5714081525802612, + "learning_rate": 4.364337936636881e-06, + "loss": 2.1879, + "step": 603 + }, + { + "epoch": 0.6619178082191781, + "grad_norm": 1.855185866355896, + "learning_rate": 4.362307067424858e-06, + "loss": 2.1659, + "step": 604 + }, + { + "epoch": 0.663013698630137, + "grad_norm": 1.826479196548462, + "learning_rate": 4.3602761982128355e-06, + "loss": 2.1716, + "step": 605 + }, + { + "epoch": 0.6641095890410958, + "grad_norm": 1.6905351877212524, + "learning_rate": 4.358245329000813e-06, + "loss": 2.1825, + "step": 606 + }, + { + "epoch": 0.6652054794520548, + "grad_norm": 1.8157447576522827, + "learning_rate": 4.35621445978879e-06, + "loss": 2.1811, + "step": 607 + }, + { + "epoch": 0.6663013698630137, + "grad_norm": 1.6311850547790527, + "learning_rate": 4.354183590576767e-06, + "loss": 2.1507, + "step": 608 + }, + { + "epoch": 0.6673972602739726, + "grad_norm": 1.579039454460144, + "learning_rate": 4.352152721364745e-06, + "loss": 2.2031, + "step": 609 + }, + { + "epoch": 0.6684931506849315, + "grad_norm": 1.7773518562316895, + "learning_rate": 4.350121852152722e-06, + "loss": 2.1978, + "step": 610 + }, + { + "epoch": 0.6684931506849315, + "eval_loss": 2.314854383468628, + "eval_runtime": 37.4501, + "eval_samples_per_second": 26.702, + "eval_steps_per_second": 1.121, + "step": 610 + }, + { + "epoch": 0.6695890410958905, + "grad_norm": 1.7606587409973145, + "learning_rate": 4.348090982940699e-06, + "loss": 2.1891, + "step": 611 + }, + { + "epoch": 0.6706849315068493, + "grad_norm": 1.6145644187927246, + "learning_rate": 4.3460601137286764e-06, + "loss": 2.1799, + "step": 612 + }, + { + "epoch": 0.6717808219178082, + "grad_norm": 1.8070226907730103, + "learning_rate": 4.344029244516654e-06, + "loss": 2.241, + "step": 613 + }, + { + "epoch": 0.6728767123287671, + "grad_norm": 1.6185903549194336, + "learning_rate": 4.341998375304631e-06, + "loss": 2.1908, + "step": 614 + }, + { + "epoch": 0.673972602739726, + "grad_norm": 1.6690049171447754, + "learning_rate": 4.339967506092607e-06, + "loss": 2.1804, + "step": 615 + }, + { + "epoch": 0.675068493150685, + "grad_norm": 1.740003228187561, + "learning_rate": 4.3379366368805855e-06, + "loss": 2.1734, + "step": 616 + }, + { + "epoch": 0.6761643835616439, + "grad_norm": 1.7311010360717773, + "learning_rate": 4.335905767668562e-06, + "loss": 2.2015, + "step": 617 + }, + { + "epoch": 0.6772602739726027, + "grad_norm": 1.7491734027862549, + "learning_rate": 4.33387489845654e-06, + "loss": 2.1894, + "step": 618 + }, + { + "epoch": 0.6783561643835616, + "grad_norm": 1.6343051195144653, + "learning_rate": 4.3318440292445165e-06, + "loss": 2.1583, + "step": 619 + }, + { + "epoch": 0.6794520547945205, + "grad_norm": 1.584243893623352, + "learning_rate": 4.329813160032495e-06, + "loss": 2.1396, + "step": 620 + }, + { + "epoch": 0.6794520547945205, + "eval_loss": 2.3157904148101807, + "eval_runtime": 37.6734, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 620 + }, + { + "epoch": 0.6805479452054795, + "grad_norm": 1.6275745630264282, + "learning_rate": 4.327782290820471e-06, + "loss": 2.2064, + "step": 621 + }, + { + "epoch": 0.6816438356164384, + "grad_norm": 1.5353672504425049, + "learning_rate": 4.325751421608449e-06, + "loss": 2.1686, + "step": 622 + }, + { + "epoch": 0.6827397260273973, + "grad_norm": 1.6155133247375488, + "learning_rate": 4.323720552396426e-06, + "loss": 2.1652, + "step": 623 + }, + { + "epoch": 0.6838356164383562, + "grad_norm": 1.638383150100708, + "learning_rate": 4.321689683184404e-06, + "loss": 2.1747, + "step": 624 + }, + { + "epoch": 0.684931506849315, + "grad_norm": 1.6153279542922974, + "learning_rate": 4.31965881397238e-06, + "loss": 2.1975, + "step": 625 + }, + { + "epoch": 0.686027397260274, + "grad_norm": 2.023699998855591, + "learning_rate": 4.3176279447603574e-06, + "loss": 2.1904, + "step": 626 + }, + { + "epoch": 0.6871232876712329, + "grad_norm": 1.5969340801239014, + "learning_rate": 4.315597075548335e-06, + "loss": 2.2133, + "step": 627 + }, + { + "epoch": 0.6882191780821918, + "grad_norm": 2.138498067855835, + "learning_rate": 4.313566206336312e-06, + "loss": 2.1685, + "step": 628 + }, + { + "epoch": 0.6893150684931507, + "grad_norm": 1.7141013145446777, + "learning_rate": 4.311535337124289e-06, + "loss": 2.1818, + "step": 629 + }, + { + "epoch": 0.6904109589041096, + "grad_norm": 1.9214462041854858, + "learning_rate": 4.3095044679122665e-06, + "loss": 2.191, + "step": 630 + }, + { + "epoch": 0.6904109589041096, + "eval_loss": 2.3150644302368164, + "eval_runtime": 37.723, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 630 + }, + { + "epoch": 0.6915068493150685, + "grad_norm": 2.03241229057312, + "learning_rate": 4.307473598700244e-06, + "loss": 2.1398, + "step": 631 + }, + { + "epoch": 0.6926027397260274, + "grad_norm": 1.608492374420166, + "learning_rate": 4.305442729488221e-06, + "loss": 2.2113, + "step": 632 + }, + { + "epoch": 0.6936986301369863, + "grad_norm": 2.1223549842834473, + "learning_rate": 4.303411860276198e-06, + "loss": 2.2012, + "step": 633 + }, + { + "epoch": 0.6947945205479452, + "grad_norm": 1.5212339162826538, + "learning_rate": 4.301380991064176e-06, + "loss": 2.2201, + "step": 634 + }, + { + "epoch": 0.6958904109589041, + "grad_norm": 1.9565715789794922, + "learning_rate": 4.299350121852153e-06, + "loss": 2.2155, + "step": 635 + }, + { + "epoch": 0.6969863013698631, + "grad_norm": 1.700822114944458, + "learning_rate": 4.29731925264013e-06, + "loss": 2.1764, + "step": 636 + }, + { + "epoch": 0.6980821917808219, + "grad_norm": 1.5625321865081787, + "learning_rate": 4.2952883834281075e-06, + "loss": 2.1407, + "step": 637 + }, + { + "epoch": 0.6991780821917808, + "grad_norm": 2.3319756984710693, + "learning_rate": 4.293257514216085e-06, + "loss": 2.2058, + "step": 638 + }, + { + "epoch": 0.7002739726027397, + "grad_norm": 1.617644190788269, + "learning_rate": 4.291226645004062e-06, + "loss": 2.2157, + "step": 639 + }, + { + "epoch": 0.7013698630136986, + "grad_norm": 1.9357377290725708, + "learning_rate": 4.289195775792039e-06, + "loss": 2.1736, + "step": 640 + }, + { + "epoch": 0.7013698630136986, + "eval_loss": 2.310267925262451, + "eval_runtime": 37.731, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 640 + }, + { + "epoch": 0.7024657534246576, + "grad_norm": 1.6908962726593018, + "learning_rate": 4.287164906580017e-06, + "loss": 2.1523, + "step": 641 + }, + { + "epoch": 0.7035616438356165, + "grad_norm": 1.6051634550094604, + "learning_rate": 4.285134037367994e-06, + "loss": 2.1499, + "step": 642 + }, + { + "epoch": 0.7046575342465753, + "grad_norm": 1.760164499282837, + "learning_rate": 4.283103168155971e-06, + "loss": 2.1493, + "step": 643 + }, + { + "epoch": 0.7057534246575342, + "grad_norm": 1.6225790977478027, + "learning_rate": 4.281072298943948e-06, + "loss": 2.1737, + "step": 644 + }, + { + "epoch": 0.7068493150684931, + "grad_norm": 1.8128079175949097, + "learning_rate": 4.279041429731926e-06, + "loss": 2.1492, + "step": 645 + }, + { + "epoch": 0.707945205479452, + "grad_norm": 1.993295431137085, + "learning_rate": 4.277010560519903e-06, + "loss": 2.1702, + "step": 646 + }, + { + "epoch": 0.709041095890411, + "grad_norm": 1.7715190649032593, + "learning_rate": 4.27497969130788e-06, + "loss": 2.1939, + "step": 647 + }, + { + "epoch": 0.7101369863013699, + "grad_norm": 1.921738624572754, + "learning_rate": 4.2729488220958575e-06, + "loss": 2.201, + "step": 648 + }, + { + "epoch": 0.7112328767123288, + "grad_norm": 1.5670042037963867, + "learning_rate": 4.270917952883835e-06, + "loss": 2.152, + "step": 649 + }, + { + "epoch": 0.7123287671232876, + "grad_norm": 1.6405668258666992, + "learning_rate": 4.268887083671812e-06, + "loss": 2.2488, + "step": 650 + }, + { + "epoch": 0.7123287671232876, + "eval_loss": 2.3100948333740234, + "eval_runtime": 37.7393, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 650 + }, + { + "epoch": 0.7134246575342466, + "grad_norm": 1.79966402053833, + "learning_rate": 4.266856214459789e-06, + "loss": 2.1506, + "step": 651 + }, + { + "epoch": 0.7145205479452055, + "grad_norm": 1.6062074899673462, + "learning_rate": 4.264825345247766e-06, + "loss": 2.1818, + "step": 652 + }, + { + "epoch": 0.7156164383561644, + "grad_norm": 1.61439049243927, + "learning_rate": 4.262794476035744e-06, + "loss": 2.1269, + "step": 653 + }, + { + "epoch": 0.7167123287671233, + "grad_norm": 1.7649872303009033, + "learning_rate": 4.26076360682372e-06, + "loss": 2.1703, + "step": 654 + }, + { + "epoch": 0.7178082191780822, + "grad_norm": 1.6238528490066528, + "learning_rate": 4.2587327376116984e-06, + "loss": 2.1764, + "step": 655 + }, + { + "epoch": 0.718904109589041, + "grad_norm": 1.6847354173660278, + "learning_rate": 4.256701868399675e-06, + "loss": 2.1796, + "step": 656 + }, + { + "epoch": 0.72, + "grad_norm": 1.7208436727523804, + "learning_rate": 4.254670999187653e-06, + "loss": 2.1903, + "step": 657 + }, + { + "epoch": 0.7210958904109589, + "grad_norm": 1.5888099670410156, + "learning_rate": 4.252640129975629e-06, + "loss": 2.1686, + "step": 658 + }, + { + "epoch": 0.7221917808219178, + "grad_norm": 1.742118000984192, + "learning_rate": 4.2506092607636075e-06, + "loss": 2.1287, + "step": 659 + }, + { + "epoch": 0.7232876712328767, + "grad_norm": 1.8450883626937866, + "learning_rate": 4.248578391551584e-06, + "loss": 2.1761, + "step": 660 + }, + { + "epoch": 0.7232876712328767, + "eval_loss": 2.3134732246398926, + "eval_runtime": 37.7568, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 660 + }, + { + "epoch": 0.7243835616438357, + "grad_norm": 1.630985140800476, + "learning_rate": 4.246547522339562e-06, + "loss": 2.1743, + "step": 661 + }, + { + "epoch": 0.7254794520547945, + "grad_norm": 1.6984914541244507, + "learning_rate": 4.2445166531275385e-06, + "loss": 2.1356, + "step": 662 + }, + { + "epoch": 0.7265753424657534, + "grad_norm": 1.6732920408248901, + "learning_rate": 4.242485783915517e-06, + "loss": 2.1377, + "step": 663 + }, + { + "epoch": 0.7276712328767123, + "grad_norm": 1.6210153102874756, + "learning_rate": 4.240454914703493e-06, + "loss": 2.1455, + "step": 664 + }, + { + "epoch": 0.7287671232876712, + "grad_norm": 1.7643382549285889, + "learning_rate": 4.23842404549147e-06, + "loss": 2.1373, + "step": 665 + }, + { + "epoch": 0.7298630136986302, + "grad_norm": 1.6491565704345703, + "learning_rate": 4.236393176279448e-06, + "loss": 2.1897, + "step": 666 + }, + { + "epoch": 0.7309589041095891, + "grad_norm": 1.670323133468628, + "learning_rate": 4.234362307067425e-06, + "loss": 2.2081, + "step": 667 + }, + { + "epoch": 0.7320547945205479, + "grad_norm": 1.6350057125091553, + "learning_rate": 4.232331437855402e-06, + "loss": 2.2002, + "step": 668 + }, + { + "epoch": 0.7331506849315068, + "grad_norm": 1.6191266775131226, + "learning_rate": 4.2303005686433795e-06, + "loss": 2.1744, + "step": 669 + }, + { + "epoch": 0.7342465753424657, + "grad_norm": 1.5511432886123657, + "learning_rate": 4.228269699431357e-06, + "loss": 2.1797, + "step": 670 + }, + { + "epoch": 0.7342465753424657, + "eval_loss": 2.313626527786255, + "eval_runtime": 37.7763, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 670 + }, + { + "epoch": 0.7353424657534247, + "grad_norm": 1.5128833055496216, + "learning_rate": 4.226238830219334e-06, + "loss": 2.1729, + "step": 671 + }, + { + "epoch": 0.7364383561643836, + "grad_norm": 1.6752032041549683, + "learning_rate": 4.224207961007311e-06, + "loss": 2.1747, + "step": 672 + }, + { + "epoch": 0.7375342465753425, + "grad_norm": 1.8643760681152344, + "learning_rate": 4.2221770917952886e-06, + "loss": 2.174, + "step": 673 + }, + { + "epoch": 0.7386301369863014, + "grad_norm": 1.5659266710281372, + "learning_rate": 4.220146222583266e-06, + "loss": 2.1706, + "step": 674 + }, + { + "epoch": 0.7397260273972602, + "grad_norm": 1.9675986766815186, + "learning_rate": 4.218115353371243e-06, + "loss": 2.1558, + "step": 675 + }, + { + "epoch": 0.7408219178082192, + "grad_norm": 1.6091173887252808, + "learning_rate": 4.21608448415922e-06, + "loss": 2.2126, + "step": 676 + }, + { + "epoch": 0.7419178082191781, + "grad_norm": 1.6801071166992188, + "learning_rate": 4.214053614947198e-06, + "loss": 2.1352, + "step": 677 + }, + { + "epoch": 0.743013698630137, + "grad_norm": 1.6964598894119263, + "learning_rate": 4.212022745735175e-06, + "loss": 2.1863, + "step": 678 + }, + { + "epoch": 0.7441095890410959, + "grad_norm": 1.670597791671753, + "learning_rate": 4.209991876523152e-06, + "loss": 2.1478, + "step": 679 + }, + { + "epoch": 0.7452054794520548, + "grad_norm": 1.6591588258743286, + "learning_rate": 4.2079610073111295e-06, + "loss": 2.2198, + "step": 680 + }, + { + "epoch": 0.7452054794520548, + "eval_loss": 2.3091039657592773, + "eval_runtime": 37.7551, + "eval_samples_per_second": 26.486, + "eval_steps_per_second": 1.112, + "step": 680 + }, + { + "epoch": 0.7463013698630137, + "grad_norm": 1.5147838592529297, + "learning_rate": 4.205930138099107e-06, + "loss": 2.1836, + "step": 681 + }, + { + "epoch": 0.7473972602739726, + "grad_norm": 1.6278952360153198, + "learning_rate": 4.203899268887084e-06, + "loss": 2.196, + "step": 682 + }, + { + "epoch": 0.7484931506849315, + "grad_norm": 1.5964363813400269, + "learning_rate": 4.201868399675061e-06, + "loss": 2.1162, + "step": 683 + }, + { + "epoch": 0.7495890410958904, + "grad_norm": 1.6993104219436646, + "learning_rate": 4.199837530463039e-06, + "loss": 2.1731, + "step": 684 + }, + { + "epoch": 0.7506849315068493, + "grad_norm": 1.6196811199188232, + "learning_rate": 4.197806661251016e-06, + "loss": 2.16, + "step": 685 + }, + { + "epoch": 0.7517808219178083, + "grad_norm": 1.6826317310333252, + "learning_rate": 4.195775792038993e-06, + "loss": 2.1123, + "step": 686 + }, + { + "epoch": 0.7528767123287671, + "grad_norm": 1.7604600191116333, + "learning_rate": 4.19374492282697e-06, + "loss": 2.1962, + "step": 687 + }, + { + "epoch": 0.753972602739726, + "grad_norm": 1.5533616542816162, + "learning_rate": 4.191714053614948e-06, + "loss": 2.1593, + "step": 688 + }, + { + "epoch": 0.7550684931506849, + "grad_norm": 1.8652232885360718, + "learning_rate": 4.189683184402924e-06, + "loss": 2.2067, + "step": 689 + }, + { + "epoch": 0.7561643835616438, + "grad_norm": 1.6330097913742065, + "learning_rate": 4.187652315190902e-06, + "loss": 2.1483, + "step": 690 + }, + { + "epoch": 0.7561643835616438, + "eval_loss": 2.309835195541382, + "eval_runtime": 37.746, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 690 + }, + { + "epoch": 0.7572602739726028, + "grad_norm": 1.5989009141921997, + "learning_rate": 4.185621445978879e-06, + "loss": 2.1686, + "step": 691 + }, + { + "epoch": 0.7583561643835617, + "grad_norm": 1.7416167259216309, + "learning_rate": 4.183590576766857e-06, + "loss": 2.2111, + "step": 692 + }, + { + "epoch": 0.7594520547945205, + "grad_norm": 1.5840280055999756, + "learning_rate": 4.181559707554833e-06, + "loss": 2.1499, + "step": 693 + }, + { + "epoch": 0.7605479452054794, + "grad_norm": 1.839306354522705, + "learning_rate": 4.179528838342811e-06, + "loss": 2.1492, + "step": 694 + }, + { + "epoch": 0.7616438356164383, + "grad_norm": 1.54163658618927, + "learning_rate": 4.177497969130788e-06, + "loss": 2.2029, + "step": 695 + }, + { + "epoch": 0.7627397260273973, + "grad_norm": 1.8050168752670288, + "learning_rate": 4.175467099918766e-06, + "loss": 2.1615, + "step": 696 + }, + { + "epoch": 0.7638356164383562, + "grad_norm": 1.6971197128295898, + "learning_rate": 4.173436230706742e-06, + "loss": 2.1476, + "step": 697 + }, + { + "epoch": 0.7649315068493151, + "grad_norm": 1.6200546026229858, + "learning_rate": 4.1714053614947205e-06, + "loss": 2.1602, + "step": 698 + }, + { + "epoch": 0.766027397260274, + "grad_norm": 1.7440853118896484, + "learning_rate": 4.169374492282697e-06, + "loss": 2.1949, + "step": 699 + }, + { + "epoch": 0.7671232876712328, + "grad_norm": 1.5590331554412842, + "learning_rate": 4.167343623070675e-06, + "loss": 2.1656, + "step": 700 + }, + { + "epoch": 0.7671232876712328, + "eval_loss": 2.309752941131592, + "eval_runtime": 37.7663, + "eval_samples_per_second": 26.479, + "eval_steps_per_second": 1.112, + "step": 700 + }, + { + "epoch": 0.7682191780821918, + "grad_norm": 1.7314256429672241, + "learning_rate": 4.165312753858652e-06, + "loss": 2.1516, + "step": 701 + }, + { + "epoch": 0.7693150684931507, + "grad_norm": 1.5145938396453857, + "learning_rate": 4.163281884646629e-06, + "loss": 2.1795, + "step": 702 + }, + { + "epoch": 0.7704109589041096, + "grad_norm": 1.5300369262695312, + "learning_rate": 4.161251015434607e-06, + "loss": 2.1356, + "step": 703 + }, + { + "epoch": 0.7715068493150685, + "grad_norm": 1.6113280057907104, + "learning_rate": 4.159220146222583e-06, + "loss": 2.157, + "step": 704 + }, + { + "epoch": 0.7726027397260274, + "grad_norm": 1.5458502769470215, + "learning_rate": 4.157189277010561e-06, + "loss": 2.2017, + "step": 705 + }, + { + "epoch": 0.7736986301369863, + "grad_norm": 1.5456187725067139, + "learning_rate": 4.155158407798538e-06, + "loss": 2.1758, + "step": 706 + }, + { + "epoch": 0.7747945205479452, + "grad_norm": 1.6051454544067383, + "learning_rate": 4.153127538586516e-06, + "loss": 2.183, + "step": 707 + }, + { + "epoch": 0.7758904109589041, + "grad_norm": 1.565788745880127, + "learning_rate": 4.151096669374492e-06, + "loss": 2.1538, + "step": 708 + }, + { + "epoch": 0.776986301369863, + "grad_norm": 1.7484439611434937, + "learning_rate": 4.1490658001624705e-06, + "loss": 2.169, + "step": 709 + }, + { + "epoch": 0.7780821917808219, + "grad_norm": 1.7206509113311768, + "learning_rate": 4.147034930950447e-06, + "loss": 2.1661, + "step": 710 + }, + { + "epoch": 0.7780821917808219, + "eval_loss": 2.307927131652832, + "eval_runtime": 37.5145, + "eval_samples_per_second": 26.656, + "eval_steps_per_second": 1.12, + "step": 710 + }, + { + "epoch": 0.7791780821917809, + "grad_norm": 1.8865350484848022, + "learning_rate": 4.145004061738425e-06, + "loss": 2.093, + "step": 711 + }, + { + "epoch": 0.7802739726027397, + "grad_norm": 1.935110092163086, + "learning_rate": 4.1429731925264015e-06, + "loss": 2.1298, + "step": 712 + }, + { + "epoch": 0.7813698630136986, + "grad_norm": 1.6938892602920532, + "learning_rate": 4.140942323314379e-06, + "loss": 2.1508, + "step": 713 + }, + { + "epoch": 0.7824657534246575, + "grad_norm": 1.5774824619293213, + "learning_rate": 4.138911454102356e-06, + "loss": 2.1808, + "step": 714 + }, + { + "epoch": 0.7835616438356164, + "grad_norm": 1.8637802600860596, + "learning_rate": 4.136880584890333e-06, + "loss": 2.1822, + "step": 715 + }, + { + "epoch": 0.7846575342465754, + "grad_norm": 1.6216785907745361, + "learning_rate": 4.1348497156783106e-06, + "loss": 2.1807, + "step": 716 + }, + { + "epoch": 0.7857534246575343, + "grad_norm": 1.7688151597976685, + "learning_rate": 4.132818846466288e-06, + "loss": 2.1967, + "step": 717 + }, + { + "epoch": 0.7868493150684932, + "grad_norm": 1.6824712753295898, + "learning_rate": 4.130787977254265e-06, + "loss": 2.1952, + "step": 718 + }, + { + "epoch": 0.787945205479452, + "grad_norm": 1.911048412322998, + "learning_rate": 4.128757108042242e-06, + "loss": 2.2408, + "step": 719 + }, + { + "epoch": 0.7890410958904109, + "grad_norm": 1.62872314453125, + "learning_rate": 4.12672623883022e-06, + "loss": 2.1647, + "step": 720 + }, + { + "epoch": 0.7890410958904109, + "eval_loss": 2.3071532249450684, + "eval_runtime": 37.6838, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 720 + }, + { + "epoch": 0.7901369863013699, + "grad_norm": 2.080674171447754, + "learning_rate": 4.124695369618197e-06, + "loss": 2.1614, + "step": 721 + }, + { + "epoch": 0.7912328767123288, + "grad_norm": 1.7225316762924194, + "learning_rate": 4.122664500406174e-06, + "loss": 2.1784, + "step": 722 + }, + { + "epoch": 0.7923287671232877, + "grad_norm": 1.838957667350769, + "learning_rate": 4.1206336311941515e-06, + "loss": 2.1281, + "step": 723 + }, + { + "epoch": 0.7934246575342466, + "grad_norm": 1.9360777139663696, + "learning_rate": 4.118602761982129e-06, + "loss": 2.1168, + "step": 724 + }, + { + "epoch": 0.7945205479452054, + "grad_norm": 1.7712862491607666, + "learning_rate": 4.116571892770106e-06, + "loss": 2.2004, + "step": 725 + }, + { + "epoch": 0.7956164383561644, + "grad_norm": 1.6759192943572998, + "learning_rate": 4.114541023558083e-06, + "loss": 2.1588, + "step": 726 + }, + { + "epoch": 0.7967123287671233, + "grad_norm": 1.8396296501159668, + "learning_rate": 4.112510154346061e-06, + "loss": 2.1777, + "step": 727 + }, + { + "epoch": 0.7978082191780822, + "grad_norm": 1.6775752305984497, + "learning_rate": 4.110479285134038e-06, + "loss": 2.1923, + "step": 728 + }, + { + "epoch": 0.7989041095890411, + "grad_norm": 1.7111576795578003, + "learning_rate": 4.108448415922015e-06, + "loss": 2.1854, + "step": 729 + }, + { + "epoch": 0.8, + "grad_norm": 1.7400974035263062, + "learning_rate": 4.1064175467099924e-06, + "loss": 2.0897, + "step": 730 + }, + { + "epoch": 0.8, + "eval_loss": 2.3070170879364014, + "eval_runtime": 37.7602, + "eval_samples_per_second": 26.483, + "eval_steps_per_second": 1.112, + "step": 730 + }, + { + "epoch": 0.8010958904109589, + "grad_norm": 1.9166353940963745, + "learning_rate": 4.10438667749797e-06, + "loss": 2.2417, + "step": 731 + }, + { + "epoch": 0.8021917808219178, + "grad_norm": 1.7914725542068481, + "learning_rate": 4.102355808285947e-06, + "loss": 2.1677, + "step": 732 + }, + { + "epoch": 0.8032876712328767, + "grad_norm": 1.7602828741073608, + "learning_rate": 4.100324939073924e-06, + "loss": 2.1623, + "step": 733 + }, + { + "epoch": 0.8043835616438356, + "grad_norm": 1.6864640712738037, + "learning_rate": 4.0982940698619015e-06, + "loss": 2.1748, + "step": 734 + }, + { + "epoch": 0.8054794520547945, + "grad_norm": 1.564070463180542, + "learning_rate": 4.096263200649879e-06, + "loss": 2.1458, + "step": 735 + }, + { + "epoch": 0.8065753424657535, + "grad_norm": 1.9004014730453491, + "learning_rate": 4.094232331437856e-06, + "loss": 2.1536, + "step": 736 + }, + { + "epoch": 0.8076712328767123, + "grad_norm": 1.6670420169830322, + "learning_rate": 4.092201462225833e-06, + "loss": 2.1809, + "step": 737 + }, + { + "epoch": 0.8087671232876712, + "grad_norm": 1.6553101539611816, + "learning_rate": 4.090170593013811e-06, + "loss": 2.1343, + "step": 738 + }, + { + "epoch": 0.8098630136986301, + "grad_norm": 1.9083986282348633, + "learning_rate": 4.088139723801787e-06, + "loss": 2.1455, + "step": 739 + }, + { + "epoch": 0.810958904109589, + "grad_norm": 1.6069375276565552, + "learning_rate": 4.086108854589765e-06, + "loss": 2.1621, + "step": 740 + }, + { + "epoch": 0.810958904109589, + "eval_loss": 2.3045499324798584, + "eval_runtime": 37.7804, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 740 + }, + { + "epoch": 0.812054794520548, + "grad_norm": 1.6774693727493286, + "learning_rate": 4.084077985377742e-06, + "loss": 2.2079, + "step": 741 + }, + { + "epoch": 0.8131506849315069, + "grad_norm": 1.772355556488037, + "learning_rate": 4.08204711616572e-06, + "loss": 2.1276, + "step": 742 + }, + { + "epoch": 0.8142465753424658, + "grad_norm": 1.6998035907745361, + "learning_rate": 4.080016246953696e-06, + "loss": 2.1903, + "step": 743 + }, + { + "epoch": 0.8153424657534246, + "grad_norm": 1.850844383239746, + "learning_rate": 4.077985377741674e-06, + "loss": 2.1965, + "step": 744 + }, + { + "epoch": 0.8164383561643835, + "grad_norm": 1.5916879177093506, + "learning_rate": 4.075954508529651e-06, + "loss": 2.1982, + "step": 745 + }, + { + "epoch": 0.8175342465753425, + "grad_norm": 1.8195812702178955, + "learning_rate": 4.073923639317629e-06, + "loss": 2.1507, + "step": 746 + }, + { + "epoch": 0.8186301369863014, + "grad_norm": 1.5145621299743652, + "learning_rate": 4.071892770105605e-06, + "loss": 2.1236, + "step": 747 + }, + { + "epoch": 0.8197260273972603, + "grad_norm": 1.5750877857208252, + "learning_rate": 4.069861900893583e-06, + "loss": 2.142, + "step": 748 + }, + { + "epoch": 0.8208219178082192, + "grad_norm": 1.708666443824768, + "learning_rate": 4.06783103168156e-06, + "loss": 2.1577, + "step": 749 + }, + { + "epoch": 0.821917808219178, + "grad_norm": 1.5988762378692627, + "learning_rate": 4.065800162469538e-06, + "loss": 2.1795, + "step": 750 + }, + { + "epoch": 0.821917808219178, + "eval_loss": 2.3052196502685547, + "eval_runtime": 37.7983, + "eval_samples_per_second": 26.456, + "eval_steps_per_second": 1.111, + "step": 750 + }, + { + "epoch": 0.823013698630137, + "grad_norm": 1.608738899230957, + "learning_rate": 4.063769293257514e-06, + "loss": 2.1578, + "step": 751 + }, + { + "epoch": 0.8241095890410959, + "grad_norm": 1.5578334331512451, + "learning_rate": 4.061738424045492e-06, + "loss": 2.1724, + "step": 752 + }, + { + "epoch": 0.8252054794520548, + "grad_norm": 1.6957333087921143, + "learning_rate": 4.059707554833469e-06, + "loss": 2.1975, + "step": 753 + }, + { + "epoch": 0.8263013698630137, + "grad_norm": 1.691512107849121, + "learning_rate": 4.057676685621446e-06, + "loss": 2.1705, + "step": 754 + }, + { + "epoch": 0.8273972602739726, + "grad_norm": 1.6061747074127197, + "learning_rate": 4.0556458164094235e-06, + "loss": 2.1031, + "step": 755 + }, + { + "epoch": 0.8284931506849315, + "grad_norm": 1.6074447631835938, + "learning_rate": 4.053614947197401e-06, + "loss": 2.1738, + "step": 756 + }, + { + "epoch": 0.8295890410958904, + "grad_norm": 1.9722917079925537, + "learning_rate": 4.051584077985378e-06, + "loss": 2.2043, + "step": 757 + }, + { + "epoch": 0.8306849315068493, + "grad_norm": 1.597692847251892, + "learning_rate": 4.049553208773355e-06, + "loss": 2.1621, + "step": 758 + }, + { + "epoch": 0.8317808219178082, + "grad_norm": 1.85458242893219, + "learning_rate": 4.047522339561333e-06, + "loss": 2.1504, + "step": 759 + }, + { + "epoch": 0.8328767123287671, + "grad_norm": 1.8328344821929932, + "learning_rate": 4.04549147034931e-06, + "loss": 2.2541, + "step": 760 + }, + { + "epoch": 0.8328767123287671, + "eval_loss": 2.3033809661865234, + "eval_runtime": 37.8088, + "eval_samples_per_second": 26.449, + "eval_steps_per_second": 1.111, + "step": 760 + }, + { + "epoch": 0.8339726027397261, + "grad_norm": 1.6544418334960938, + "learning_rate": 4.043460601137287e-06, + "loss": 2.1605, + "step": 761 + }, + { + "epoch": 0.8350684931506849, + "grad_norm": 1.9443641901016235, + "learning_rate": 4.041429731925264e-06, + "loss": 2.1621, + "step": 762 + }, + { + "epoch": 0.8361643835616438, + "grad_norm": 1.731277346611023, + "learning_rate": 4.039398862713242e-06, + "loss": 2.2168, + "step": 763 + }, + { + "epoch": 0.8372602739726027, + "grad_norm": 1.671072244644165, + "learning_rate": 4.037367993501219e-06, + "loss": 2.1612, + "step": 764 + }, + { + "epoch": 0.8383561643835616, + "grad_norm": 1.8777449131011963, + "learning_rate": 4.035337124289196e-06, + "loss": 2.1339, + "step": 765 + }, + { + "epoch": 0.8394520547945206, + "grad_norm": 1.5941543579101562, + "learning_rate": 4.0333062550771735e-06, + "loss": 2.1545, + "step": 766 + }, + { + "epoch": 0.8405479452054795, + "grad_norm": 1.9245301485061646, + "learning_rate": 4.031275385865151e-06, + "loss": 2.187, + "step": 767 + }, + { + "epoch": 0.8416438356164384, + "grad_norm": 1.69217848777771, + "learning_rate": 4.029244516653128e-06, + "loss": 2.1168, + "step": 768 + }, + { + "epoch": 0.8427397260273972, + "grad_norm": 1.7042242288589478, + "learning_rate": 4.027213647441105e-06, + "loss": 2.1353, + "step": 769 + }, + { + "epoch": 0.8438356164383561, + "grad_norm": 1.6391927003860474, + "learning_rate": 4.025182778229083e-06, + "loss": 2.1568, + "step": 770 + }, + { + "epoch": 0.8438356164383561, + "eval_loss": 2.310056686401367, + "eval_runtime": 37.7573, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 770 + }, + { + "epoch": 0.8449315068493151, + "grad_norm": 1.7833209037780762, + "learning_rate": 4.02315190901706e-06, + "loss": 2.1664, + "step": 771 + }, + { + "epoch": 0.846027397260274, + "grad_norm": 1.8203891515731812, + "learning_rate": 4.021121039805037e-06, + "loss": 2.2342, + "step": 772 + }, + { + "epoch": 0.8471232876712329, + "grad_norm": 1.572940707206726, + "learning_rate": 4.0190901705930144e-06, + "loss": 2.1418, + "step": 773 + }, + { + "epoch": 0.8482191780821918, + "grad_norm": 1.6871955394744873, + "learning_rate": 4.017059301380992e-06, + "loss": 2.2023, + "step": 774 + }, + { + "epoch": 0.8493150684931506, + "grad_norm": 1.6343133449554443, + "learning_rate": 4.015028432168969e-06, + "loss": 2.1302, + "step": 775 + }, + { + "epoch": 0.8504109589041096, + "grad_norm": 1.4492065906524658, + "learning_rate": 4.012997562956945e-06, + "loss": 2.1427, + "step": 776 + }, + { + "epoch": 0.8515068493150685, + "grad_norm": 1.6599146127700806, + "learning_rate": 4.0109666937449235e-06, + "loss": 2.1479, + "step": 777 + }, + { + "epoch": 0.8526027397260274, + "grad_norm": 1.6836353540420532, + "learning_rate": 4.0089358245329e-06, + "loss": 2.148, + "step": 778 + }, + { + "epoch": 0.8536986301369863, + "grad_norm": 1.5977990627288818, + "learning_rate": 4.006904955320878e-06, + "loss": 2.1946, + "step": 779 + }, + { + "epoch": 0.8547945205479452, + "grad_norm": 1.6259957551956177, + "learning_rate": 4.0048740861088545e-06, + "loss": 2.1761, + "step": 780 + }, + { + "epoch": 0.8547945205479452, + "eval_loss": 2.303910970687866, + "eval_runtime": 37.7728, + "eval_samples_per_second": 26.474, + "eval_steps_per_second": 1.112, + "step": 780 + }, + { + "epoch": 0.8558904109589041, + "grad_norm": 1.6990300416946411, + "learning_rate": 4.002843216896833e-06, + "loss": 2.1309, + "step": 781 + }, + { + "epoch": 0.856986301369863, + "grad_norm": 1.6556029319763184, + "learning_rate": 4.000812347684809e-06, + "loss": 2.151, + "step": 782 + }, + { + "epoch": 0.8580821917808219, + "grad_norm": 1.5659290552139282, + "learning_rate": 3.998781478472787e-06, + "loss": 2.1317, + "step": 783 + }, + { + "epoch": 0.8591780821917808, + "grad_norm": 1.6553704738616943, + "learning_rate": 3.996750609260764e-06, + "loss": 2.2344, + "step": 784 + }, + { + "epoch": 0.8602739726027397, + "grad_norm": 1.6608418226242065, + "learning_rate": 3.994719740048742e-06, + "loss": 2.1807, + "step": 785 + }, + { + "epoch": 0.8613698630136987, + "grad_norm": 1.5862030982971191, + "learning_rate": 3.992688870836718e-06, + "loss": 2.1351, + "step": 786 + }, + { + "epoch": 0.8624657534246576, + "grad_norm": 1.767526388168335, + "learning_rate": 3.990658001624696e-06, + "loss": 2.1856, + "step": 787 + }, + { + "epoch": 0.8635616438356164, + "grad_norm": 1.6902577877044678, + "learning_rate": 3.988627132412673e-06, + "loss": 2.1592, + "step": 788 + }, + { + "epoch": 0.8646575342465753, + "grad_norm": 1.7760623693466187, + "learning_rate": 3.98659626320065e-06, + "loss": 2.1645, + "step": 789 + }, + { + "epoch": 0.8657534246575342, + "grad_norm": 1.6791038513183594, + "learning_rate": 3.984565393988627e-06, + "loss": 2.1611, + "step": 790 + }, + { + "epoch": 0.8657534246575342, + "eval_loss": 2.3027942180633545, + "eval_runtime": 37.736, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 790 + }, + { + "epoch": 0.8668493150684932, + "grad_norm": 1.6695773601531982, + "learning_rate": 3.9825345247766046e-06, + "loss": 2.1747, + "step": 791 + }, + { + "epoch": 0.8679452054794521, + "grad_norm": 1.8736445903778076, + "learning_rate": 3.980503655564582e-06, + "loss": 2.1806, + "step": 792 + }, + { + "epoch": 0.869041095890411, + "grad_norm": 1.6695518493652344, + "learning_rate": 3.978472786352559e-06, + "loss": 2.1631, + "step": 793 + }, + { + "epoch": 0.8701369863013698, + "grad_norm": 1.6782366037368774, + "learning_rate": 3.976441917140536e-06, + "loss": 2.2175, + "step": 794 + }, + { + "epoch": 0.8712328767123287, + "grad_norm": 1.678619623184204, + "learning_rate": 3.974411047928514e-06, + "loss": 2.1883, + "step": 795 + }, + { + "epoch": 0.8723287671232877, + "grad_norm": 1.6274909973144531, + "learning_rate": 3.972380178716491e-06, + "loss": 2.1866, + "step": 796 + }, + { + "epoch": 0.8734246575342466, + "grad_norm": 1.5609477758407593, + "learning_rate": 3.970349309504468e-06, + "loss": 2.1257, + "step": 797 + }, + { + "epoch": 0.8745205479452055, + "grad_norm": 1.7419973611831665, + "learning_rate": 3.9683184402924455e-06, + "loss": 2.1253, + "step": 798 + }, + { + "epoch": 0.8756164383561644, + "grad_norm": 1.6296330690383911, + "learning_rate": 3.966287571080423e-06, + "loss": 2.1755, + "step": 799 + }, + { + "epoch": 0.8767123287671232, + "grad_norm": 1.5995864868164062, + "learning_rate": 3.9642567018684e-06, + "loss": 2.169, + "step": 800 + }, + { + "epoch": 0.8767123287671232, + "eval_loss": 2.301098108291626, + "eval_runtime": 37.7893, + "eval_samples_per_second": 26.463, + "eval_steps_per_second": 1.111, + "step": 800 + }, + { + "epoch": 0.8778082191780822, + "grad_norm": 1.653707504272461, + "learning_rate": 3.962225832656377e-06, + "loss": 2.115, + "step": 801 + }, + { + "epoch": 0.8789041095890411, + "grad_norm": 1.5908581018447876, + "learning_rate": 3.960194963444355e-06, + "loss": 2.1659, + "step": 802 + }, + { + "epoch": 0.88, + "grad_norm": 1.6865545511245728, + "learning_rate": 3.958164094232332e-06, + "loss": 2.1666, + "step": 803 + }, + { + "epoch": 0.8810958904109589, + "grad_norm": 1.6234707832336426, + "learning_rate": 3.956133225020309e-06, + "loss": 2.1594, + "step": 804 + }, + { + "epoch": 0.8821917808219178, + "grad_norm": 1.7375986576080322, + "learning_rate": 3.954102355808286e-06, + "loss": 2.1596, + "step": 805 + }, + { + "epoch": 0.8832876712328767, + "grad_norm": 1.6529247760772705, + "learning_rate": 3.952071486596264e-06, + "loss": 2.1495, + "step": 806 + }, + { + "epoch": 0.8843835616438356, + "grad_norm": 1.517357349395752, + "learning_rate": 3.950040617384241e-06, + "loss": 2.1394, + "step": 807 + }, + { + "epoch": 0.8854794520547945, + "grad_norm": 1.7367429733276367, + "learning_rate": 3.948009748172218e-06, + "loss": 2.206, + "step": 808 + }, + { + "epoch": 0.8865753424657534, + "grad_norm": 1.5467783212661743, + "learning_rate": 3.9459788789601955e-06, + "loss": 2.1672, + "step": 809 + }, + { + "epoch": 0.8876712328767123, + "grad_norm": 1.5647791624069214, + "learning_rate": 3.943948009748173e-06, + "loss": 2.1695, + "step": 810 + }, + { + "epoch": 0.8876712328767123, + "eval_loss": 2.2975995540618896, + "eval_runtime": 37.4912, + "eval_samples_per_second": 26.673, + "eval_steps_per_second": 1.12, + "step": 810 + }, + { + "epoch": 0.8887671232876713, + "grad_norm": 1.597876787185669, + "learning_rate": 3.94191714053615e-06, + "loss": 2.1808, + "step": 811 + }, + { + "epoch": 0.8898630136986302, + "grad_norm": 1.7645189762115479, + "learning_rate": 3.939886271324127e-06, + "loss": 2.1291, + "step": 812 + }, + { + "epoch": 0.890958904109589, + "grad_norm": 1.5209035873413086, + "learning_rate": 3.937855402112104e-06, + "loss": 2.1269, + "step": 813 + }, + { + "epoch": 0.8920547945205479, + "grad_norm": 1.5804905891418457, + "learning_rate": 3.935824532900082e-06, + "loss": 2.1751, + "step": 814 + }, + { + "epoch": 0.8931506849315068, + "grad_norm": 1.687865972518921, + "learning_rate": 3.933793663688058e-06, + "loss": 2.1917, + "step": 815 + }, + { + "epoch": 0.8942465753424658, + "grad_norm": 1.6767957210540771, + "learning_rate": 3.9317627944760365e-06, + "loss": 2.118, + "step": 816 + }, + { + "epoch": 0.8953424657534247, + "grad_norm": 1.657046914100647, + "learning_rate": 3.929731925264013e-06, + "loss": 2.1479, + "step": 817 + }, + { + "epoch": 0.8964383561643836, + "grad_norm": 1.5638110637664795, + "learning_rate": 3.927701056051991e-06, + "loss": 2.1565, + "step": 818 + }, + { + "epoch": 0.8975342465753424, + "grad_norm": 1.6865980625152588, + "learning_rate": 3.9256701868399674e-06, + "loss": 2.1382, + "step": 819 + }, + { + "epoch": 0.8986301369863013, + "grad_norm": 1.6738817691802979, + "learning_rate": 3.9236393176279456e-06, + "loss": 2.1257, + "step": 820 + }, + { + "epoch": 0.8986301369863013, + "eval_loss": 2.2953474521636963, + "eval_runtime": 37.7016, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 820 + }, + { + "epoch": 0.8997260273972603, + "grad_norm": 1.568678617477417, + "learning_rate": 3.921608448415922e-06, + "loss": 2.169, + "step": 821 + }, + { + "epoch": 0.9008219178082192, + "grad_norm": 1.6803818941116333, + "learning_rate": 3.9195775792039e-06, + "loss": 2.1489, + "step": 822 + }, + { + "epoch": 0.9019178082191781, + "grad_norm": 1.7010858058929443, + "learning_rate": 3.9175467099918765e-06, + "loss": 2.1846, + "step": 823 + }, + { + "epoch": 0.903013698630137, + "grad_norm": 1.5782372951507568, + "learning_rate": 3.915515840779855e-06, + "loss": 2.1788, + "step": 824 + }, + { + "epoch": 0.9041095890410958, + "grad_norm": 1.7804685831069946, + "learning_rate": 3.913484971567831e-06, + "loss": 2.1058, + "step": 825 + }, + { + "epoch": 0.9052054794520548, + "grad_norm": 1.5909292697906494, + "learning_rate": 3.911454102355808e-06, + "loss": 2.1716, + "step": 826 + }, + { + "epoch": 0.9063013698630137, + "grad_norm": 1.7001307010650635, + "learning_rate": 3.909423233143786e-06, + "loss": 2.1647, + "step": 827 + }, + { + "epoch": 0.9073972602739726, + "grad_norm": 1.5576646327972412, + "learning_rate": 3.907392363931763e-06, + "loss": 2.2174, + "step": 828 + }, + { + "epoch": 0.9084931506849315, + "grad_norm": 1.6621390581130981, + "learning_rate": 3.90536149471974e-06, + "loss": 2.2344, + "step": 829 + }, + { + "epoch": 0.9095890410958904, + "grad_norm": 1.592091679573059, + "learning_rate": 3.9033306255077175e-06, + "loss": 2.2501, + "step": 830 + }, + { + "epoch": 0.9095890410958904, + "eval_loss": 2.294229507446289, + "eval_runtime": 37.7195, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.113, + "step": 830 + }, + { + "epoch": 0.9106849315068493, + "grad_norm": 1.5698535442352295, + "learning_rate": 3.901299756295695e-06, + "loss": 2.1534, + "step": 831 + }, + { + "epoch": 0.9117808219178082, + "grad_norm": 1.7159922122955322, + "learning_rate": 3.899268887083672e-06, + "loss": 2.1601, + "step": 832 + }, + { + "epoch": 0.9128767123287671, + "grad_norm": 1.5601210594177246, + "learning_rate": 3.897238017871649e-06, + "loss": 2.1598, + "step": 833 + }, + { + "epoch": 0.913972602739726, + "grad_norm": 1.5987435579299927, + "learning_rate": 3.8952071486596266e-06, + "loss": 2.1307, + "step": 834 + }, + { + "epoch": 0.915068493150685, + "grad_norm": 1.7568409442901611, + "learning_rate": 3.893176279447604e-06, + "loss": 2.2209, + "step": 835 + }, + { + "epoch": 0.9161643835616439, + "grad_norm": 1.5044634342193604, + "learning_rate": 3.891145410235581e-06, + "loss": 2.1699, + "step": 836 + }, + { + "epoch": 0.9172602739726028, + "grad_norm": 1.7287421226501465, + "learning_rate": 3.889114541023558e-06, + "loss": 2.1489, + "step": 837 + }, + { + "epoch": 0.9183561643835616, + "grad_norm": 1.6595203876495361, + "learning_rate": 3.887083671811536e-06, + "loss": 2.1167, + "step": 838 + }, + { + "epoch": 0.9194520547945205, + "grad_norm": 1.5620248317718506, + "learning_rate": 3.885052802599513e-06, + "loss": 2.2119, + "step": 839 + }, + { + "epoch": 0.9205479452054794, + "grad_norm": 1.6492353677749634, + "learning_rate": 3.88302193338749e-06, + "loss": 2.1725, + "step": 840 + }, + { + "epoch": 0.9205479452054794, + "eval_loss": 2.2956349849700928, + "eval_runtime": 37.779, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 840 + }, + { + "epoch": 0.9216438356164384, + "grad_norm": 1.6601520776748657, + "learning_rate": 3.8809910641754675e-06, + "loss": 2.1532, + "step": 841 + }, + { + "epoch": 0.9227397260273973, + "grad_norm": 1.6723994016647339, + "learning_rate": 3.878960194963445e-06, + "loss": 2.1931, + "step": 842 + }, + { + "epoch": 0.9238356164383562, + "grad_norm": 1.7594473361968994, + "learning_rate": 3.876929325751422e-06, + "loss": 2.1219, + "step": 843 + }, + { + "epoch": 0.924931506849315, + "grad_norm": 1.546398401260376, + "learning_rate": 3.874898456539399e-06, + "loss": 2.1679, + "step": 844 + }, + { + "epoch": 0.9260273972602739, + "grad_norm": 1.8171741962432861, + "learning_rate": 3.872867587327377e-06, + "loss": 2.1434, + "step": 845 + }, + { + "epoch": 0.9271232876712329, + "grad_norm": 1.703196406364441, + "learning_rate": 3.870836718115354e-06, + "loss": 2.1574, + "step": 846 + }, + { + "epoch": 0.9282191780821918, + "grad_norm": 1.6290628910064697, + "learning_rate": 3.868805848903331e-06, + "loss": 2.1657, + "step": 847 + }, + { + "epoch": 0.9293150684931507, + "grad_norm": 1.5868806838989258, + "learning_rate": 3.8667749796913084e-06, + "loss": 2.1659, + "step": 848 + }, + { + "epoch": 0.9304109589041096, + "grad_norm": 1.551581621170044, + "learning_rate": 3.864744110479286e-06, + "loss": 2.1916, + "step": 849 + }, + { + "epoch": 0.9315068493150684, + "grad_norm": 1.6634963750839233, + "learning_rate": 3.862713241267262e-06, + "loss": 2.2027, + "step": 850 + }, + { + "epoch": 0.9315068493150684, + "eval_loss": 2.2988946437835693, + "eval_runtime": 37.7841, + "eval_samples_per_second": 26.466, + "eval_steps_per_second": 1.112, + "step": 850 + }, + { + "epoch": 0.9326027397260274, + "grad_norm": 1.6138510704040527, + "learning_rate": 3.86068237205524e-06, + "loss": 2.1634, + "step": 851 + }, + { + "epoch": 0.9336986301369863, + "grad_norm": 1.6294615268707275, + "learning_rate": 3.858651502843217e-06, + "loss": 2.1175, + "step": 852 + }, + { + "epoch": 0.9347945205479452, + "grad_norm": 1.7515549659729004, + "learning_rate": 3.856620633631195e-06, + "loss": 2.1562, + "step": 853 + }, + { + "epoch": 0.9358904109589041, + "grad_norm": 1.6304186582565308, + "learning_rate": 3.854589764419171e-06, + "loss": 2.2256, + "step": 854 + }, + { + "epoch": 0.936986301369863, + "grad_norm": 1.6189597845077515, + "learning_rate": 3.852558895207149e-06, + "loss": 2.1742, + "step": 855 + }, + { + "epoch": 0.938082191780822, + "grad_norm": 1.665515661239624, + "learning_rate": 3.850528025995126e-06, + "loss": 2.1481, + "step": 856 + }, + { + "epoch": 0.9391780821917808, + "grad_norm": 1.7066419124603271, + "learning_rate": 3.848497156783104e-06, + "loss": 2.1522, + "step": 857 + }, + { + "epoch": 0.9402739726027397, + "grad_norm": 1.884599208831787, + "learning_rate": 3.84646628757108e-06, + "loss": 2.1493, + "step": 858 + }, + { + "epoch": 0.9413698630136986, + "grad_norm": 1.6224873065948486, + "learning_rate": 3.8444354183590585e-06, + "loss": 2.1461, + "step": 859 + }, + { + "epoch": 0.9424657534246575, + "grad_norm": 1.916416049003601, + "learning_rate": 3.842404549147035e-06, + "loss": 2.1265, + "step": 860 + }, + { + "epoch": 0.9424657534246575, + "eval_loss": 2.297154426574707, + "eval_runtime": 37.77, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 860 + }, + { + "epoch": 0.9435616438356165, + "grad_norm": 1.5281857252120972, + "learning_rate": 3.840373679935013e-06, + "loss": 2.1476, + "step": 861 + }, + { + "epoch": 0.9446575342465754, + "grad_norm": 1.7905207872390747, + "learning_rate": 3.8383428107229894e-06, + "loss": 2.1421, + "step": 862 + }, + { + "epoch": 0.9457534246575342, + "grad_norm": 1.8282849788665771, + "learning_rate": 3.836311941510967e-06, + "loss": 2.1807, + "step": 863 + }, + { + "epoch": 0.9468493150684931, + "grad_norm": 1.7861522436141968, + "learning_rate": 3.834281072298944e-06, + "loss": 2.1337, + "step": 864 + }, + { + "epoch": 0.947945205479452, + "grad_norm": 1.6107250452041626, + "learning_rate": 3.832250203086921e-06, + "loss": 2.1911, + "step": 865 + }, + { + "epoch": 0.949041095890411, + "grad_norm": 1.657909631729126, + "learning_rate": 3.8302193338748985e-06, + "loss": 2.2257, + "step": 866 + }, + { + "epoch": 0.9501369863013699, + "grad_norm": 1.8969722986221313, + "learning_rate": 3.828188464662876e-06, + "loss": 2.1778, + "step": 867 + }, + { + "epoch": 0.9512328767123288, + "grad_norm": 1.6498360633850098, + "learning_rate": 3.826157595450853e-06, + "loss": 2.127, + "step": 868 + }, + { + "epoch": 0.9523287671232876, + "grad_norm": 1.5750585794448853, + "learning_rate": 3.82412672623883e-06, + "loss": 2.1716, + "step": 869 + }, + { + "epoch": 0.9534246575342465, + "grad_norm": 1.6047886610031128, + "learning_rate": 3.822095857026808e-06, + "loss": 2.1354, + "step": 870 + }, + { + "epoch": 0.9534246575342465, + "eval_loss": 2.295785665512085, + "eval_runtime": 37.7419, + "eval_samples_per_second": 26.496, + "eval_steps_per_second": 1.113, + "step": 870 + }, + { + "epoch": 0.9545205479452055, + "grad_norm": 1.6426187753677368, + "learning_rate": 3.820064987814785e-06, + "loss": 2.1375, + "step": 871 + }, + { + "epoch": 0.9556164383561644, + "grad_norm": 1.5118632316589355, + "learning_rate": 3.818034118602762e-06, + "loss": 2.1491, + "step": 872 + }, + { + "epoch": 0.9567123287671233, + "grad_norm": 1.630476474761963, + "learning_rate": 3.8160032493907395e-06, + "loss": 2.1497, + "step": 873 + }, + { + "epoch": 0.9578082191780822, + "grad_norm": 1.6820838451385498, + "learning_rate": 3.813972380178717e-06, + "loss": 2.1746, + "step": 874 + }, + { + "epoch": 0.958904109589041, + "grad_norm": 1.5480482578277588, + "learning_rate": 3.811941510966694e-06, + "loss": 2.1226, + "step": 875 + }, + { + "epoch": 0.96, + "grad_norm": 1.601493239402771, + "learning_rate": 3.809910641754671e-06, + "loss": 2.0934, + "step": 876 + }, + { + "epoch": 0.9610958904109589, + "grad_norm": 1.5756622552871704, + "learning_rate": 3.8078797725426486e-06, + "loss": 2.1679, + "step": 877 + }, + { + "epoch": 0.9621917808219178, + "grad_norm": 1.6811070442199707, + "learning_rate": 3.8058489033306254e-06, + "loss": 2.1407, + "step": 878 + }, + { + "epoch": 0.9632876712328767, + "grad_norm": 1.5748940706253052, + "learning_rate": 3.803818034118603e-06, + "loss": 2.1985, + "step": 879 + }, + { + "epoch": 0.9643835616438357, + "grad_norm": 1.5412342548370361, + "learning_rate": 3.80178716490658e-06, + "loss": 2.1876, + "step": 880 + }, + { + "epoch": 0.9643835616438357, + "eval_loss": 2.291630268096924, + "eval_runtime": 37.7702, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 880 + }, + { + "epoch": 0.9654794520547946, + "grad_norm": 1.7080684900283813, + "learning_rate": 3.7997562956945577e-06, + "loss": 2.1453, + "step": 881 + }, + { + "epoch": 0.9665753424657534, + "grad_norm": 1.8705110549926758, + "learning_rate": 3.7977254264825345e-06, + "loss": 2.1298, + "step": 882 + }, + { + "epoch": 0.9676712328767123, + "grad_norm": 1.7332537174224854, + "learning_rate": 3.7956945572705122e-06, + "loss": 2.2117, + "step": 883 + }, + { + "epoch": 0.9687671232876712, + "grad_norm": 1.88144052028656, + "learning_rate": 3.793663688058489e-06, + "loss": 2.2005, + "step": 884 + }, + { + "epoch": 0.9698630136986301, + "grad_norm": 1.6263943910598755, + "learning_rate": 3.791632818846467e-06, + "loss": 2.1853, + "step": 885 + }, + { + "epoch": 0.9709589041095891, + "grad_norm": 1.5968067646026611, + "learning_rate": 3.7896019496344436e-06, + "loss": 2.1998, + "step": 886 + }, + { + "epoch": 0.972054794520548, + "grad_norm": 1.7149827480316162, + "learning_rate": 3.787571080422421e-06, + "loss": 2.1855, + "step": 887 + }, + { + "epoch": 0.9731506849315068, + "grad_norm": 1.7375026941299438, + "learning_rate": 3.785540211210398e-06, + "loss": 2.1664, + "step": 888 + }, + { + "epoch": 0.9742465753424657, + "grad_norm": 1.6411423683166504, + "learning_rate": 3.7835093419983755e-06, + "loss": 2.1492, + "step": 889 + }, + { + "epoch": 0.9753424657534246, + "grad_norm": 1.6208266019821167, + "learning_rate": 3.7814784727863527e-06, + "loss": 2.2341, + "step": 890 + }, + { + "epoch": 0.9753424657534246, + "eval_loss": 2.2929985523223877, + "eval_runtime": 37.6693, + "eval_samples_per_second": 26.547, + "eval_steps_per_second": 1.115, + "step": 890 + }, + { + "epoch": 0.9764383561643836, + "grad_norm": 1.6871188879013062, + "learning_rate": 3.77944760357433e-06, + "loss": 2.1985, + "step": 891 + }, + { + "epoch": 0.9775342465753425, + "grad_norm": 1.8475587368011475, + "learning_rate": 3.7774167343623073e-06, + "loss": 2.1423, + "step": 892 + }, + { + "epoch": 0.9786301369863014, + "grad_norm": 1.6115065813064575, + "learning_rate": 3.7753858651502846e-06, + "loss": 2.1944, + "step": 893 + }, + { + "epoch": 0.9797260273972602, + "grad_norm": 1.6232231855392456, + "learning_rate": 3.773354995938262e-06, + "loss": 2.1751, + "step": 894 + }, + { + "epoch": 0.9808219178082191, + "grad_norm": 1.8564099073410034, + "learning_rate": 3.771324126726239e-06, + "loss": 2.2021, + "step": 895 + }, + { + "epoch": 0.9819178082191781, + "grad_norm": 1.7270197868347168, + "learning_rate": 3.7692932575142164e-06, + "loss": 2.1351, + "step": 896 + }, + { + "epoch": 0.983013698630137, + "grad_norm": 1.7187668085098267, + "learning_rate": 3.7672623883021937e-06, + "loss": 2.0856, + "step": 897 + }, + { + "epoch": 0.9841095890410959, + "grad_norm": 1.7766387462615967, + "learning_rate": 3.765231519090171e-06, + "loss": 2.168, + "step": 898 + }, + { + "epoch": 0.9852054794520548, + "grad_norm": 1.8063374757766724, + "learning_rate": 3.7632006498781482e-06, + "loss": 2.1317, + "step": 899 + }, + { + "epoch": 0.9863013698630136, + "grad_norm": 1.7601871490478516, + "learning_rate": 3.761169780666125e-06, + "loss": 2.1742, + "step": 900 + }, + { + "epoch": 0.9863013698630136, + "eval_loss": 2.292574167251587, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 900 + }, + { + "epoch": 0.9873972602739726, + "grad_norm": 1.8501280546188354, + "learning_rate": 3.7591389114541028e-06, + "loss": 2.1856, + "step": 901 + }, + { + "epoch": 0.9884931506849315, + "grad_norm": 1.8786728382110596, + "learning_rate": 3.7571080422420796e-06, + "loss": 2.1432, + "step": 902 + }, + { + "epoch": 0.9895890410958904, + "grad_norm": 1.7662389278411865, + "learning_rate": 3.7550771730300573e-06, + "loss": 2.2011, + "step": 903 + }, + { + "epoch": 0.9906849315068493, + "grad_norm": 1.6810210943222046, + "learning_rate": 3.753046303818034e-06, + "loss": 2.1719, + "step": 904 + }, + { + "epoch": 0.9917808219178083, + "grad_norm": 2.034982681274414, + "learning_rate": 3.751015434606012e-06, + "loss": 2.2206, + "step": 905 + }, + { + "epoch": 0.9928767123287672, + "grad_norm": 1.804049015045166, + "learning_rate": 3.7489845653939887e-06, + "loss": 2.1248, + "step": 906 + }, + { + "epoch": 0.993972602739726, + "grad_norm": 1.6039106845855713, + "learning_rate": 3.7469536961819664e-06, + "loss": 2.1545, + "step": 907 + }, + { + "epoch": 0.9950684931506849, + "grad_norm": 1.9352415800094604, + "learning_rate": 3.7449228269699433e-06, + "loss": 2.1842, + "step": 908 + }, + { + "epoch": 0.9961643835616438, + "grad_norm": 1.5759649276733398, + "learning_rate": 3.742891957757921e-06, + "loss": 2.1588, + "step": 909 + }, + { + "epoch": 0.9972602739726028, + "grad_norm": 1.6169486045837402, + "learning_rate": 3.740861088545898e-06, + "loss": 2.1871, + "step": 910 + }, + { + "epoch": 0.9972602739726028, + "eval_loss": 2.2923426628112793, + "eval_runtime": 37.4584, + "eval_samples_per_second": 26.696, + "eval_steps_per_second": 1.121, + "step": 910 + }, + { + "epoch": 0.9983561643835617, + "grad_norm": 1.748138427734375, + "learning_rate": 3.7388302193338755e-06, + "loss": 2.144, + "step": 911 + }, + { + "epoch": 0.9994520547945206, + "grad_norm": 1.8182991743087769, + "learning_rate": 3.7367993501218524e-06, + "loss": 2.1474, + "step": 912 + }, + { + "epoch": 1.0005479452054795, + "grad_norm": 1.6532196998596191, + "learning_rate": 3.7347684809098297e-06, + "loss": 2.1146, + "step": 913 + }, + { + "epoch": 1.0016438356164383, + "grad_norm": 2.0338973999023438, + "learning_rate": 3.732737611697807e-06, + "loss": 2.0851, + "step": 914 + }, + { + "epoch": 1.0027397260273974, + "grad_norm": 1.71133553981781, + "learning_rate": 3.7307067424857842e-06, + "loss": 2.1197, + "step": 915 + }, + { + "epoch": 1.0038356164383562, + "grad_norm": 1.6081736087799072, + "learning_rate": 3.7286758732737615e-06, + "loss": 2.1106, + "step": 916 + }, + { + "epoch": 1.004931506849315, + "grad_norm": 2.100734233856201, + "learning_rate": 3.7266450040617388e-06, + "loss": 2.2125, + "step": 917 + }, + { + "epoch": 1.006027397260274, + "grad_norm": 1.758898377418518, + "learning_rate": 3.724614134849716e-06, + "loss": 2.0985, + "step": 918 + }, + { + "epoch": 1.0071232876712328, + "grad_norm": 1.7335586547851562, + "learning_rate": 3.7225832656376933e-06, + "loss": 2.1911, + "step": 919 + }, + { + "epoch": 1.0082191780821919, + "grad_norm": 1.7879023551940918, + "learning_rate": 3.7205523964256706e-06, + "loss": 2.1632, + "step": 920 + }, + { + "epoch": 1.0082191780821919, + "eval_loss": 2.294053554534912, + "eval_runtime": 37.6132, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 920 + }, + { + "epoch": 1.0093150684931507, + "grad_norm": 1.7201987504959106, + "learning_rate": 3.718521527213648e-06, + "loss": 2.1501, + "step": 921 + }, + { + "epoch": 1.0104109589041095, + "grad_norm": 1.7638986110687256, + "learning_rate": 3.716490658001625e-06, + "loss": 2.0661, + "step": 922 + }, + { + "epoch": 1.0115068493150685, + "grad_norm": 1.7376288175582886, + "learning_rate": 3.7144597887896024e-06, + "loss": 2.1613, + "step": 923 + }, + { + "epoch": 1.0126027397260273, + "grad_norm": 1.9208155870437622, + "learning_rate": 3.7124289195775797e-06, + "loss": 2.1562, + "step": 924 + }, + { + "epoch": 1.0136986301369864, + "grad_norm": 1.6709166765213013, + "learning_rate": 3.710398050365557e-06, + "loss": 2.1309, + "step": 925 + }, + { + "epoch": 1.0147945205479452, + "grad_norm": 1.7070904970169067, + "learning_rate": 3.708367181153534e-06, + "loss": 2.1538, + "step": 926 + }, + { + "epoch": 1.0158904109589042, + "grad_norm": 1.7832821607589722, + "learning_rate": 3.7063363119415115e-06, + "loss": 2.1262, + "step": 927 + }, + { + "epoch": 1.016986301369863, + "grad_norm": 1.8703486919403076, + "learning_rate": 3.7043054427294884e-06, + "loss": 2.1142, + "step": 928 + }, + { + "epoch": 1.0180821917808218, + "grad_norm": 1.692858338356018, + "learning_rate": 3.702274573517466e-06, + "loss": 2.1182, + "step": 929 + }, + { + "epoch": 1.0191780821917809, + "grad_norm": 1.7156164646148682, + "learning_rate": 3.700243704305443e-06, + "loss": 2.1189, + "step": 930 + }, + { + "epoch": 1.0191780821917809, + "eval_loss": 2.2969048023223877, + "eval_runtime": 37.6673, + "eval_samples_per_second": 26.548, + "eval_steps_per_second": 1.115, + "step": 930 + }, + { + "epoch": 1.0202739726027397, + "grad_norm": 1.7443044185638428, + "learning_rate": 3.6982128350934206e-06, + "loss": 2.1058, + "step": 931 + }, + { + "epoch": 1.0213698630136987, + "grad_norm": 2.00412917137146, + "learning_rate": 3.6961819658813975e-06, + "loss": 2.2071, + "step": 932 + }, + { + "epoch": 1.0224657534246575, + "grad_norm": 1.7170531749725342, + "learning_rate": 3.694151096669375e-06, + "loss": 2.1254, + "step": 933 + }, + { + "epoch": 1.0235616438356165, + "grad_norm": 1.679018497467041, + "learning_rate": 3.692120227457352e-06, + "loss": 2.1487, + "step": 934 + }, + { + "epoch": 1.0246575342465754, + "grad_norm": 1.8477305173873901, + "learning_rate": 3.6900893582453297e-06, + "loss": 2.062, + "step": 935 + }, + { + "epoch": 1.0257534246575342, + "grad_norm": 1.664095401763916, + "learning_rate": 3.6880584890333066e-06, + "loss": 2.1418, + "step": 936 + }, + { + "epoch": 1.0268493150684932, + "grad_norm": 1.8831602334976196, + "learning_rate": 3.6860276198212834e-06, + "loss": 2.1483, + "step": 937 + }, + { + "epoch": 1.027945205479452, + "grad_norm": 1.7625365257263184, + "learning_rate": 3.683996750609261e-06, + "loss": 2.146, + "step": 938 + }, + { + "epoch": 1.029041095890411, + "grad_norm": 1.8168137073516846, + "learning_rate": 3.681965881397238e-06, + "loss": 2.1594, + "step": 939 + }, + { + "epoch": 1.0301369863013699, + "grad_norm": 1.6803343296051025, + "learning_rate": 3.6799350121852157e-06, + "loss": 2.1168, + "step": 940 + }, + { + "epoch": 1.0301369863013699, + "eval_loss": 2.2941982746124268, + "eval_runtime": 37.7112, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 940 + }, + { + "epoch": 1.0312328767123287, + "grad_norm": 2.0829122066497803, + "learning_rate": 3.6779041429731925e-06, + "loss": 2.1399, + "step": 941 + }, + { + "epoch": 1.0323287671232877, + "grad_norm": 1.6813585758209229, + "learning_rate": 3.6758732737611702e-06, + "loss": 2.1312, + "step": 942 + }, + { + "epoch": 1.0334246575342465, + "grad_norm": 1.7511613368988037, + "learning_rate": 3.673842404549147e-06, + "loss": 2.1273, + "step": 943 + }, + { + "epoch": 1.0345205479452055, + "grad_norm": 1.8264555931091309, + "learning_rate": 3.671811535337125e-06, + "loss": 2.1343, + "step": 944 + }, + { + "epoch": 1.0356164383561643, + "grad_norm": 1.7675082683563232, + "learning_rate": 3.6697806661251016e-06, + "loss": 2.2153, + "step": 945 + }, + { + "epoch": 1.0367123287671234, + "grad_norm": 1.6589823961257935, + "learning_rate": 3.6677497969130793e-06, + "loss": 2.1575, + "step": 946 + }, + { + "epoch": 1.0378082191780822, + "grad_norm": 1.6730010509490967, + "learning_rate": 3.665718927701056e-06, + "loss": 2.1773, + "step": 947 + }, + { + "epoch": 1.038904109589041, + "grad_norm": 1.627983570098877, + "learning_rate": 3.663688058489034e-06, + "loss": 2.1394, + "step": 948 + }, + { + "epoch": 1.04, + "grad_norm": 1.6363394260406494, + "learning_rate": 3.6616571892770107e-06, + "loss": 2.0828, + "step": 949 + }, + { + "epoch": 1.0410958904109588, + "grad_norm": 1.6270133256912231, + "learning_rate": 3.659626320064988e-06, + "loss": 2.1308, + "step": 950 + }, + { + "epoch": 1.0410958904109588, + "eval_loss": 2.293911933898926, + "eval_runtime": 37.7095, + "eval_samples_per_second": 26.519, + "eval_steps_per_second": 1.114, + "step": 950 + }, + { + "epoch": 1.0421917808219179, + "grad_norm": 1.784549593925476, + "learning_rate": 3.6575954508529653e-06, + "loss": 2.1239, + "step": 951 + }, + { + "epoch": 1.0432876712328767, + "grad_norm": 1.7265311479568481, + "learning_rate": 3.6555645816409426e-06, + "loss": 2.1573, + "step": 952 + }, + { + "epoch": 1.0443835616438357, + "grad_norm": 1.8133327960968018, + "learning_rate": 3.65353371242892e-06, + "loss": 2.046, + "step": 953 + }, + { + "epoch": 1.0454794520547945, + "grad_norm": 1.7085450887680054, + "learning_rate": 3.651502843216897e-06, + "loss": 2.1471, + "step": 954 + }, + { + "epoch": 1.0465753424657533, + "grad_norm": 2.0099172592163086, + "learning_rate": 3.6494719740048744e-06, + "loss": 2.1193, + "step": 955 + }, + { + "epoch": 1.0476712328767124, + "grad_norm": 1.582085132598877, + "learning_rate": 3.6474411047928517e-06, + "loss": 2.1522, + "step": 956 + }, + { + "epoch": 1.0487671232876712, + "grad_norm": 1.7923028469085693, + "learning_rate": 3.645410235580829e-06, + "loss": 2.1358, + "step": 957 + }, + { + "epoch": 1.0498630136986302, + "grad_norm": 1.6636080741882324, + "learning_rate": 3.6433793663688062e-06, + "loss": 2.1263, + "step": 958 + }, + { + "epoch": 1.050958904109589, + "grad_norm": 1.681675672531128, + "learning_rate": 3.6413484971567835e-06, + "loss": 2.1369, + "step": 959 + }, + { + "epoch": 1.0520547945205478, + "grad_norm": 1.7007646560668945, + "learning_rate": 3.6393176279447608e-06, + "loss": 2.1606, + "step": 960 + }, + { + "epoch": 1.0520547945205478, + "eval_loss": 2.291552782058716, + "eval_runtime": 37.704, + "eval_samples_per_second": 26.522, + "eval_steps_per_second": 1.114, + "step": 960 + }, + { + "epoch": 1.0531506849315069, + "grad_norm": 1.632997989654541, + "learning_rate": 3.637286758732738e-06, + "loss": 2.1742, + "step": 961 + }, + { + "epoch": 1.0542465753424657, + "grad_norm": 1.6881494522094727, + "learning_rate": 3.6352558895207153e-06, + "loss": 2.1555, + "step": 962 + }, + { + "epoch": 1.0553424657534247, + "grad_norm": 1.6624116897583008, + "learning_rate": 3.633225020308692e-06, + "loss": 2.1278, + "step": 963 + }, + { + "epoch": 1.0564383561643835, + "grad_norm": 1.7851325273513794, + "learning_rate": 3.63119415109667e-06, + "loss": 2.1599, + "step": 964 + }, + { + "epoch": 1.0575342465753426, + "grad_norm": 1.7592188119888306, + "learning_rate": 3.6291632818846467e-06, + "loss": 2.0898, + "step": 965 + }, + { + "epoch": 1.0586301369863014, + "grad_norm": 1.9573521614074707, + "learning_rate": 3.6271324126726244e-06, + "loss": 2.0876, + "step": 966 + }, + { + "epoch": 1.0597260273972602, + "grad_norm": 1.6684988737106323, + "learning_rate": 3.6251015434606013e-06, + "loss": 2.1288, + "step": 967 + }, + { + "epoch": 1.0608219178082192, + "grad_norm": 2.0011937618255615, + "learning_rate": 3.623070674248579e-06, + "loss": 2.1166, + "step": 968 + }, + { + "epoch": 1.061917808219178, + "grad_norm": 1.7490673065185547, + "learning_rate": 3.621039805036556e-06, + "loss": 2.1532, + "step": 969 + }, + { + "epoch": 1.063013698630137, + "grad_norm": 1.756422758102417, + "learning_rate": 3.6190089358245335e-06, + "loss": 2.1486, + "step": 970 + }, + { + "epoch": 1.063013698630137, + "eval_loss": 2.29402494430542, + "eval_runtime": 37.7078, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 970 + }, + { + "epoch": 1.0641095890410959, + "grad_norm": 2.014153480529785, + "learning_rate": 3.6169780666125104e-06, + "loss": 2.1512, + "step": 971 + }, + { + "epoch": 1.0652054794520547, + "grad_norm": 1.6026027202606201, + "learning_rate": 3.614947197400488e-06, + "loss": 2.0787, + "step": 972 + }, + { + "epoch": 1.0663013698630137, + "grad_norm": 1.8149386644363403, + "learning_rate": 3.612916328188465e-06, + "loss": 2.1675, + "step": 973 + }, + { + "epoch": 1.0673972602739725, + "grad_norm": 1.697019338607788, + "learning_rate": 3.610885458976442e-06, + "loss": 2.1584, + "step": 974 + }, + { + "epoch": 1.0684931506849316, + "grad_norm": 1.7330286502838135, + "learning_rate": 3.6088545897644195e-06, + "loss": 2.1048, + "step": 975 + }, + { + "epoch": 1.0695890410958904, + "grad_norm": 1.650336742401123, + "learning_rate": 3.6068237205523963e-06, + "loss": 2.1078, + "step": 976 + }, + { + "epoch": 1.0706849315068494, + "grad_norm": 1.7817063331604004, + "learning_rate": 3.604792851340374e-06, + "loss": 2.1656, + "step": 977 + }, + { + "epoch": 1.0717808219178082, + "grad_norm": 1.818617820739746, + "learning_rate": 3.602761982128351e-06, + "loss": 2.1388, + "step": 978 + }, + { + "epoch": 1.072876712328767, + "grad_norm": 1.7448735237121582, + "learning_rate": 3.6007311129163286e-06, + "loss": 2.1022, + "step": 979 + }, + { + "epoch": 1.073972602739726, + "grad_norm": 1.8401628732681274, + "learning_rate": 3.5987002437043054e-06, + "loss": 2.1181, + "step": 980 + }, + { + "epoch": 1.073972602739726, + "eval_loss": 2.2948012351989746, + "eval_runtime": 37.6902, + "eval_samples_per_second": 26.532, + "eval_steps_per_second": 1.114, + "step": 980 + }, + { + "epoch": 1.0750684931506849, + "grad_norm": 1.979794979095459, + "learning_rate": 3.596669374492283e-06, + "loss": 2.1766, + "step": 981 + }, + { + "epoch": 1.076164383561644, + "grad_norm": 1.7657036781311035, + "learning_rate": 3.59463850528026e-06, + "loss": 2.0889, + "step": 982 + }, + { + "epoch": 1.0772602739726027, + "grad_norm": 1.8293733596801758, + "learning_rate": 3.5926076360682377e-06, + "loss": 2.159, + "step": 983 + }, + { + "epoch": 1.0783561643835617, + "grad_norm": 1.7979159355163574, + "learning_rate": 3.5905767668562145e-06, + "loss": 2.1391, + "step": 984 + }, + { + "epoch": 1.0794520547945206, + "grad_norm": 1.847598910331726, + "learning_rate": 3.5885458976441922e-06, + "loss": 2.1695, + "step": 985 + }, + { + "epoch": 1.0805479452054794, + "grad_norm": 1.608178734779358, + "learning_rate": 3.586515028432169e-06, + "loss": 2.1745, + "step": 986 + }, + { + "epoch": 1.0816438356164384, + "grad_norm": 1.8166873455047607, + "learning_rate": 3.5844841592201464e-06, + "loss": 2.0882, + "step": 987 + }, + { + "epoch": 1.0827397260273972, + "grad_norm": 1.847463607788086, + "learning_rate": 3.5824532900081237e-06, + "loss": 2.1244, + "step": 988 + }, + { + "epoch": 1.0838356164383562, + "grad_norm": 1.6194696426391602, + "learning_rate": 3.580422420796101e-06, + "loss": 2.1398, + "step": 989 + }, + { + "epoch": 1.084931506849315, + "grad_norm": 1.9055372476577759, + "learning_rate": 3.578391551584078e-06, + "loss": 2.0774, + "step": 990 + }, + { + "epoch": 1.084931506849315, + "eval_loss": 2.2899444103240967, + "eval_runtime": 37.7237, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 990 + }, + { + "epoch": 1.086027397260274, + "grad_norm": 1.6451077461242676, + "learning_rate": 3.5763606823720555e-06, + "loss": 2.1351, + "step": 991 + }, + { + "epoch": 1.087123287671233, + "grad_norm": 1.8144199848175049, + "learning_rate": 3.5743298131600328e-06, + "loss": 2.1536, + "step": 992 + }, + { + "epoch": 1.0882191780821917, + "grad_norm": 1.760130763053894, + "learning_rate": 3.57229894394801e-06, + "loss": 2.1485, + "step": 993 + }, + { + "epoch": 1.0893150684931507, + "grad_norm": 1.7544556856155396, + "learning_rate": 3.5702680747359873e-06, + "loss": 2.1136, + "step": 994 + }, + { + "epoch": 1.0904109589041096, + "grad_norm": 1.9884154796600342, + "learning_rate": 3.5682372055239646e-06, + "loss": 2.1366, + "step": 995 + }, + { + "epoch": 1.0915068493150686, + "grad_norm": 1.72066330909729, + "learning_rate": 3.566206336311942e-06, + "loss": 2.1103, + "step": 996 + }, + { + "epoch": 1.0926027397260274, + "grad_norm": 1.807808756828308, + "learning_rate": 3.564175467099919e-06, + "loss": 2.0905, + "step": 997 + }, + { + "epoch": 1.0936986301369862, + "grad_norm": 2.1075172424316406, + "learning_rate": 3.5621445978878964e-06, + "loss": 2.1427, + "step": 998 + }, + { + "epoch": 1.0947945205479452, + "grad_norm": 1.7037009000778198, + "learning_rate": 3.5601137286758737e-06, + "loss": 2.1124, + "step": 999 + }, + { + "epoch": 1.095890410958904, + "grad_norm": 1.842721700668335, + "learning_rate": 3.5580828594638505e-06, + "loss": 2.1379, + "step": 1000 + }, + { + "epoch": 1.095890410958904, + "eval_loss": 2.291067361831665, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 1000 + }, + { + "epoch": 1.096986301369863, + "grad_norm": 1.802979588508606, + "learning_rate": 3.5560519902518282e-06, + "loss": 2.156, + "step": 1001 + }, + { + "epoch": 1.098082191780822, + "grad_norm": 1.699763298034668, + "learning_rate": 3.554021121039805e-06, + "loss": 2.1353, + "step": 1002 + }, + { + "epoch": 1.099178082191781, + "grad_norm": 1.695544958114624, + "learning_rate": 3.551990251827783e-06, + "loss": 2.1039, + "step": 1003 + }, + { + "epoch": 1.1002739726027397, + "grad_norm": 1.9509766101837158, + "learning_rate": 3.5499593826157596e-06, + "loss": 2.1201, + "step": 1004 + }, + { + "epoch": 1.1013698630136985, + "grad_norm": 1.7657231092453003, + "learning_rate": 3.5479285134037373e-06, + "loss": 2.1149, + "step": 1005 + }, + { + "epoch": 1.1024657534246576, + "grad_norm": 1.7416653633117676, + "learning_rate": 3.545897644191714e-06, + "loss": 2.1116, + "step": 1006 + }, + { + "epoch": 1.1035616438356164, + "grad_norm": 1.804185390472412, + "learning_rate": 3.543866774979692e-06, + "loss": 2.1426, + "step": 1007 + }, + { + "epoch": 1.1046575342465754, + "grad_norm": 1.8011984825134277, + "learning_rate": 3.5418359057676687e-06, + "loss": 2.1132, + "step": 1008 + }, + { + "epoch": 1.1057534246575342, + "grad_norm": 1.9039791822433472, + "learning_rate": 3.5398050365556464e-06, + "loss": 2.1479, + "step": 1009 + }, + { + "epoch": 1.106849315068493, + "grad_norm": 1.9513524770736694, + "learning_rate": 3.5377741673436233e-06, + "loss": 2.1093, + "step": 1010 + }, + { + "epoch": 1.106849315068493, + "eval_loss": 2.290522575378418, + "eval_runtime": 37.4617, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 1010 + }, + { + "epoch": 1.107945205479452, + "grad_norm": 1.7262980937957764, + "learning_rate": 3.535743298131601e-06, + "loss": 2.138, + "step": 1011 + }, + { + "epoch": 1.1090410958904109, + "grad_norm": 1.9514356851577759, + "learning_rate": 3.533712428919578e-06, + "loss": 2.1669, + "step": 1012 + }, + { + "epoch": 1.11013698630137, + "grad_norm": 1.7947099208831787, + "learning_rate": 3.5316815597075547e-06, + "loss": 2.105, + "step": 1013 + }, + { + "epoch": 1.1112328767123287, + "grad_norm": 1.8680298328399658, + "learning_rate": 3.5296506904955324e-06, + "loss": 2.1306, + "step": 1014 + }, + { + "epoch": 1.1123287671232878, + "grad_norm": 2.121732711791992, + "learning_rate": 3.5276198212835093e-06, + "loss": 2.1681, + "step": 1015 + }, + { + "epoch": 1.1134246575342466, + "grad_norm": 1.662564754486084, + "learning_rate": 3.525588952071487e-06, + "loss": 2.1681, + "step": 1016 + }, + { + "epoch": 1.1145205479452054, + "grad_norm": 1.8358488082885742, + "learning_rate": 3.523558082859464e-06, + "loss": 2.1363, + "step": 1017 + }, + { + "epoch": 1.1156164383561644, + "grad_norm": 1.8742910623550415, + "learning_rate": 3.5215272136474415e-06, + "loss": 2.1179, + "step": 1018 + }, + { + "epoch": 1.1167123287671232, + "grad_norm": 1.8780633211135864, + "learning_rate": 3.5194963444354184e-06, + "loss": 2.103, + "step": 1019 + }, + { + "epoch": 1.1178082191780823, + "grad_norm": 1.6979540586471558, + "learning_rate": 3.517465475223396e-06, + "loss": 2.0885, + "step": 1020 + }, + { + "epoch": 1.1178082191780823, + "eval_loss": 2.2939751148223877, + "eval_runtime": 37.66, + "eval_samples_per_second": 26.553, + "eval_steps_per_second": 1.115, + "step": 1020 + }, + { + "epoch": 1.118904109589041, + "grad_norm": 2.494971752166748, + "learning_rate": 3.515434606011373e-06, + "loss": 2.1582, + "step": 1021 + }, + { + "epoch": 1.12, + "grad_norm": 1.9183858633041382, + "learning_rate": 3.5134037367993506e-06, + "loss": 2.189, + "step": 1022 + }, + { + "epoch": 1.121095890410959, + "grad_norm": 1.9055982828140259, + "learning_rate": 3.5113728675873275e-06, + "loss": 2.1503, + "step": 1023 + }, + { + "epoch": 1.1221917808219177, + "grad_norm": 2.118046522140503, + "learning_rate": 3.5093419983753047e-06, + "loss": 2.1414, + "step": 1024 + }, + { + "epoch": 1.1232876712328768, + "grad_norm": 1.8413280248641968, + "learning_rate": 3.507311129163282e-06, + "loss": 2.09, + "step": 1025 + }, + { + "epoch": 1.1243835616438356, + "grad_norm": 2.0261356830596924, + "learning_rate": 3.5052802599512593e-06, + "loss": 2.132, + "step": 1026 + }, + { + "epoch": 1.1254794520547946, + "grad_norm": 1.8393104076385498, + "learning_rate": 3.5032493907392366e-06, + "loss": 2.1366, + "step": 1027 + }, + { + "epoch": 1.1265753424657534, + "grad_norm": 2.0282111167907715, + "learning_rate": 3.501218521527214e-06, + "loss": 2.0758, + "step": 1028 + }, + { + "epoch": 1.1276712328767124, + "grad_norm": 1.8438650369644165, + "learning_rate": 3.499187652315191e-06, + "loss": 2.1294, + "step": 1029 + }, + { + "epoch": 1.1287671232876713, + "grad_norm": 1.890739917755127, + "learning_rate": 3.4971567831031684e-06, + "loss": 2.1591, + "step": 1030 + }, + { + "epoch": 1.1287671232876713, + "eval_loss": 2.2894623279571533, + "eval_runtime": 37.6941, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1030 + }, + { + "epoch": 1.12986301369863, + "grad_norm": 1.877964973449707, + "learning_rate": 3.4951259138911457e-06, + "loss": 2.0823, + "step": 1031 + }, + { + "epoch": 1.130958904109589, + "grad_norm": 1.7460334300994873, + "learning_rate": 3.493095044679123e-06, + "loss": 2.1758, + "step": 1032 + }, + { + "epoch": 1.132054794520548, + "grad_norm": 1.668459415435791, + "learning_rate": 3.4910641754671002e-06, + "loss": 2.0461, + "step": 1033 + }, + { + "epoch": 1.133150684931507, + "grad_norm": 1.9560495615005493, + "learning_rate": 3.4890333062550775e-06, + "loss": 2.1394, + "step": 1034 + }, + { + "epoch": 1.1342465753424658, + "grad_norm": 1.820432424545288, + "learning_rate": 3.4870024370430548e-06, + "loss": 2.0735, + "step": 1035 + }, + { + "epoch": 1.1353424657534246, + "grad_norm": 1.7064971923828125, + "learning_rate": 3.484971567831032e-06, + "loss": 2.1196, + "step": 1036 + }, + { + "epoch": 1.1364383561643836, + "grad_norm": 1.837594747543335, + "learning_rate": 3.482940698619009e-06, + "loss": 2.1562, + "step": 1037 + }, + { + "epoch": 1.1375342465753424, + "grad_norm": 1.7570106983184814, + "learning_rate": 3.4809098294069866e-06, + "loss": 2.1253, + "step": 1038 + }, + { + "epoch": 1.1386301369863014, + "grad_norm": 1.7931634187698364, + "learning_rate": 3.4788789601949634e-06, + "loss": 2.1369, + "step": 1039 + }, + { + "epoch": 1.1397260273972603, + "grad_norm": 1.957241177558899, + "learning_rate": 3.476848090982941e-06, + "loss": 2.1204, + "step": 1040 + }, + { + "epoch": 1.1397260273972603, + "eval_loss": 2.2888705730438232, + "eval_runtime": 37.7002, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 1040 + }, + { + "epoch": 1.140821917808219, + "grad_norm": 1.7801663875579834, + "learning_rate": 3.474817221770918e-06, + "loss": 2.2008, + "step": 1041 + }, + { + "epoch": 1.141917808219178, + "grad_norm": 1.654398798942566, + "learning_rate": 3.4727863525588957e-06, + "loss": 2.1451, + "step": 1042 + }, + { + "epoch": 1.143013698630137, + "grad_norm": 1.8888044357299805, + "learning_rate": 3.4707554833468725e-06, + "loss": 2.1318, + "step": 1043 + }, + { + "epoch": 1.144109589041096, + "grad_norm": 1.8398077487945557, + "learning_rate": 3.4687246141348502e-06, + "loss": 2.1051, + "step": 1044 + }, + { + "epoch": 1.1452054794520548, + "grad_norm": 1.6971104145050049, + "learning_rate": 3.466693744922827e-06, + "loss": 2.119, + "step": 1045 + }, + { + "epoch": 1.1463013698630138, + "grad_norm": 1.605700969696045, + "learning_rate": 3.464662875710805e-06, + "loss": 2.1343, + "step": 1046 + }, + { + "epoch": 1.1473972602739726, + "grad_norm": 1.8513860702514648, + "learning_rate": 3.4626320064987817e-06, + "loss": 2.1331, + "step": 1047 + }, + { + "epoch": 1.1484931506849314, + "grad_norm": 1.7160199880599976, + "learning_rate": 3.4606011372867594e-06, + "loss": 2.1367, + "step": 1048 + }, + { + "epoch": 1.1495890410958904, + "grad_norm": 1.5926575660705566, + "learning_rate": 3.458570268074736e-06, + "loss": 2.1183, + "step": 1049 + }, + { + "epoch": 1.1506849315068493, + "grad_norm": 1.663843035697937, + "learning_rate": 3.4565393988627135e-06, + "loss": 2.1107, + "step": 1050 + }, + { + "epoch": 1.1506849315068493, + "eval_loss": 2.286076545715332, + "eval_runtime": 37.6944, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1050 + }, + { + "epoch": 1.1517808219178083, + "grad_norm": 1.7697075605392456, + "learning_rate": 3.4545085296506908e-06, + "loss": 2.0912, + "step": 1051 + }, + { + "epoch": 1.152876712328767, + "grad_norm": 1.626094937324524, + "learning_rate": 3.452477660438668e-06, + "loss": 2.1508, + "step": 1052 + }, + { + "epoch": 1.1539726027397261, + "grad_norm": 1.9503792524337769, + "learning_rate": 3.4504467912266453e-06, + "loss": 2.0819, + "step": 1053 + }, + { + "epoch": 1.155068493150685, + "grad_norm": 1.7748280763626099, + "learning_rate": 3.4484159220146226e-06, + "loss": 2.1233, + "step": 1054 + }, + { + "epoch": 1.1561643835616437, + "grad_norm": 2.099339008331299, + "learning_rate": 3.4463850528026e-06, + "loss": 2.1245, + "step": 1055 + }, + { + "epoch": 1.1572602739726028, + "grad_norm": 1.6200031042099, + "learning_rate": 3.444354183590577e-06, + "loss": 2.1437, + "step": 1056 + }, + { + "epoch": 1.1583561643835616, + "grad_norm": 1.7703361511230469, + "learning_rate": 3.4423233143785544e-06, + "loss": 2.0874, + "step": 1057 + }, + { + "epoch": 1.1594520547945206, + "grad_norm": 1.6505805253982544, + "learning_rate": 3.4402924451665317e-06, + "loss": 2.0655, + "step": 1058 + }, + { + "epoch": 1.1605479452054794, + "grad_norm": 1.7199327945709229, + "learning_rate": 3.438261575954509e-06, + "loss": 2.1383, + "step": 1059 + }, + { + "epoch": 1.1616438356164385, + "grad_norm": 1.772292137145996, + "learning_rate": 3.4362307067424862e-06, + "loss": 2.1673, + "step": 1060 + }, + { + "epoch": 1.1616438356164385, + "eval_loss": 2.2869999408721924, + "eval_runtime": 37.7398, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1060 + }, + { + "epoch": 1.1627397260273973, + "grad_norm": 1.5849649906158447, + "learning_rate": 3.434199837530463e-06, + "loss": 2.1331, + "step": 1061 + }, + { + "epoch": 1.163835616438356, + "grad_norm": 1.8389405012130737, + "learning_rate": 3.432168968318441e-06, + "loss": 2.0832, + "step": 1062 + }, + { + "epoch": 1.1649315068493151, + "grad_norm": 1.7142471075057983, + "learning_rate": 3.4301380991064176e-06, + "loss": 2.1447, + "step": 1063 + }, + { + "epoch": 1.166027397260274, + "grad_norm": 1.7853225469589233, + "learning_rate": 3.4281072298943953e-06, + "loss": 2.1591, + "step": 1064 + }, + { + "epoch": 1.167123287671233, + "grad_norm": 1.8228695392608643, + "learning_rate": 3.426076360682372e-06, + "loss": 2.1684, + "step": 1065 + }, + { + "epoch": 1.1682191780821918, + "grad_norm": 1.7090480327606201, + "learning_rate": 3.42404549147035e-06, + "loss": 2.1339, + "step": 1066 + }, + { + "epoch": 1.1693150684931506, + "grad_norm": 1.642682433128357, + "learning_rate": 3.4220146222583267e-06, + "loss": 2.075, + "step": 1067 + }, + { + "epoch": 1.1704109589041096, + "grad_norm": 1.8230397701263428, + "learning_rate": 3.4199837530463044e-06, + "loss": 2.1128, + "step": 1068 + }, + { + "epoch": 1.1715068493150684, + "grad_norm": 1.761135458946228, + "learning_rate": 3.4179528838342813e-06, + "loss": 2.1885, + "step": 1069 + }, + { + "epoch": 1.1726027397260275, + "grad_norm": 1.7829216718673706, + "learning_rate": 3.415922014622259e-06, + "loss": 2.1584, + "step": 1070 + }, + { + "epoch": 1.1726027397260275, + "eval_loss": 2.2878382205963135, + "eval_runtime": 37.7433, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1070 + }, + { + "epoch": 1.1736986301369863, + "grad_norm": 1.5480386018753052, + "learning_rate": 3.413891145410236e-06, + "loss": 2.1602, + "step": 1071 + }, + { + "epoch": 1.174794520547945, + "grad_norm": 1.8957836627960205, + "learning_rate": 3.4118602761982135e-06, + "loss": 2.1706, + "step": 1072 + }, + { + "epoch": 1.1758904109589041, + "grad_norm": 1.924424648284912, + "learning_rate": 3.4098294069861904e-06, + "loss": 2.1339, + "step": 1073 + }, + { + "epoch": 1.176986301369863, + "grad_norm": 1.660370111465454, + "learning_rate": 3.4077985377741673e-06, + "loss": 2.1175, + "step": 1074 + }, + { + "epoch": 1.178082191780822, + "grad_norm": 1.7255185842514038, + "learning_rate": 3.405767668562145e-06, + "loss": 2.1672, + "step": 1075 + }, + { + "epoch": 1.1791780821917808, + "grad_norm": 1.8028507232666016, + "learning_rate": 3.403736799350122e-06, + "loss": 2.1527, + "step": 1076 + }, + { + "epoch": 1.1802739726027398, + "grad_norm": 1.7100828886032104, + "learning_rate": 3.4017059301380995e-06, + "loss": 2.1031, + "step": 1077 + }, + { + "epoch": 1.1813698630136986, + "grad_norm": 1.785955548286438, + "learning_rate": 3.3996750609260764e-06, + "loss": 2.1829, + "step": 1078 + }, + { + "epoch": 1.1824657534246574, + "grad_norm": 1.8583980798721313, + "learning_rate": 3.397644191714054e-06, + "loss": 2.1198, + "step": 1079 + }, + { + "epoch": 1.1835616438356165, + "grad_norm": 1.8848942518234253, + "learning_rate": 3.395613322502031e-06, + "loss": 2.123, + "step": 1080 + }, + { + "epoch": 1.1835616438356165, + "eval_loss": 2.288412094116211, + "eval_runtime": 37.7637, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 1080 + }, + { + "epoch": 1.1846575342465753, + "grad_norm": 1.6747279167175293, + "learning_rate": 3.3935824532900086e-06, + "loss": 2.1087, + "step": 1081 + }, + { + "epoch": 1.1857534246575343, + "grad_norm": 1.8993490934371948, + "learning_rate": 3.3915515840779855e-06, + "loss": 2.1316, + "step": 1082 + }, + { + "epoch": 1.1868493150684931, + "grad_norm": 1.7047377824783325, + "learning_rate": 3.389520714865963e-06, + "loss": 2.1242, + "step": 1083 + }, + { + "epoch": 1.1879452054794521, + "grad_norm": 1.767811894416809, + "learning_rate": 3.38748984565394e-06, + "loss": 2.1556, + "step": 1084 + }, + { + "epoch": 1.189041095890411, + "grad_norm": 1.665313720703125, + "learning_rate": 3.3854589764419177e-06, + "loss": 2.101, + "step": 1085 + }, + { + "epoch": 1.1901369863013698, + "grad_norm": 1.7752090692520142, + "learning_rate": 3.3834281072298946e-06, + "loss": 2.1307, + "step": 1086 + }, + { + "epoch": 1.1912328767123288, + "grad_norm": 1.9759247303009033, + "learning_rate": 3.381397238017872e-06, + "loss": 2.1093, + "step": 1087 + }, + { + "epoch": 1.1923287671232876, + "grad_norm": 1.8772896528244019, + "learning_rate": 3.379366368805849e-06, + "loss": 2.1851, + "step": 1088 + }, + { + "epoch": 1.1934246575342466, + "grad_norm": 1.7442713975906372, + "learning_rate": 3.3773354995938264e-06, + "loss": 2.1433, + "step": 1089 + }, + { + "epoch": 1.1945205479452055, + "grad_norm": 1.8106931447982788, + "learning_rate": 3.3753046303818037e-06, + "loss": 2.1246, + "step": 1090 + }, + { + "epoch": 1.1945205479452055, + "eval_loss": 2.2858333587646484, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1090 + }, + { + "epoch": 1.1956164383561645, + "grad_norm": 1.7938610315322876, + "learning_rate": 3.373273761169781e-06, + "loss": 2.1325, + "step": 1091 + }, + { + "epoch": 1.1967123287671233, + "grad_norm": 1.5760858058929443, + "learning_rate": 3.3712428919577582e-06, + "loss": 2.1036, + "step": 1092 + }, + { + "epoch": 1.1978082191780821, + "grad_norm": 1.8173094987869263, + "learning_rate": 3.3692120227457355e-06, + "loss": 2.1509, + "step": 1093 + }, + { + "epoch": 1.1989041095890411, + "grad_norm": 2.0316812992095947, + "learning_rate": 3.3671811535337128e-06, + "loss": 2.0883, + "step": 1094 + }, + { + "epoch": 1.2, + "grad_norm": 1.7735658884048462, + "learning_rate": 3.36515028432169e-06, + "loss": 2.1038, + "step": 1095 + }, + { + "epoch": 1.201095890410959, + "grad_norm": 1.752718210220337, + "learning_rate": 3.3631194151096673e-06, + "loss": 2.1393, + "step": 1096 + }, + { + "epoch": 1.2021917808219178, + "grad_norm": 1.778838872909546, + "learning_rate": 3.3610885458976446e-06, + "loss": 2.1641, + "step": 1097 + }, + { + "epoch": 1.2032876712328768, + "grad_norm": 1.7422083616256714, + "learning_rate": 3.359057676685622e-06, + "loss": 2.1624, + "step": 1098 + }, + { + "epoch": 1.2043835616438356, + "grad_norm": 1.7196860313415527, + "learning_rate": 3.357026807473599e-06, + "loss": 2.0845, + "step": 1099 + }, + { + "epoch": 1.2054794520547945, + "grad_norm": 1.8373215198516846, + "learning_rate": 3.354995938261576e-06, + "loss": 2.1092, + "step": 1100 + }, + { + "epoch": 1.2054794520547945, + "eval_loss": 2.285470485687256, + "eval_runtime": 37.728, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 1100 + }, + { + "epoch": 1.2065753424657535, + "grad_norm": 1.626450538635254, + "learning_rate": 3.3529650690495537e-06, + "loss": 2.0789, + "step": 1101 + }, + { + "epoch": 1.2076712328767123, + "grad_norm": 1.736264705657959, + "learning_rate": 3.3509341998375305e-06, + "loss": 2.1227, + "step": 1102 + }, + { + "epoch": 1.208767123287671, + "grad_norm": 1.6195942163467407, + "learning_rate": 3.3489033306255082e-06, + "loss": 2.1301, + "step": 1103 + }, + { + "epoch": 1.2098630136986301, + "grad_norm": 1.6958918571472168, + "learning_rate": 3.346872461413485e-06, + "loss": 2.1059, + "step": 1104 + }, + { + "epoch": 1.210958904109589, + "grad_norm": 1.6385109424591064, + "learning_rate": 3.344841592201463e-06, + "loss": 2.0917, + "step": 1105 + }, + { + "epoch": 1.212054794520548, + "grad_norm": 1.6833516359329224, + "learning_rate": 3.3428107229894397e-06, + "loss": 2.1204, + "step": 1106 + }, + { + "epoch": 1.2131506849315068, + "grad_norm": 1.724780559539795, + "learning_rate": 3.3407798537774174e-06, + "loss": 2.138, + "step": 1107 + }, + { + "epoch": 1.2142465753424658, + "grad_norm": 1.7483834028244019, + "learning_rate": 3.338748984565394e-06, + "loss": 2.1628, + "step": 1108 + }, + { + "epoch": 1.2153424657534246, + "grad_norm": 1.6398268938064575, + "learning_rate": 3.336718115353372e-06, + "loss": 2.1738, + "step": 1109 + }, + { + "epoch": 1.2164383561643834, + "grad_norm": 1.6916258335113525, + "learning_rate": 3.3346872461413488e-06, + "loss": 2.1182, + "step": 1110 + }, + { + "epoch": 1.2164383561643834, + "eval_loss": 2.285295009613037, + "eval_runtime": 37.508, + "eval_samples_per_second": 26.661, + "eval_steps_per_second": 1.12, + "step": 1110 + }, + { + "epoch": 1.2175342465753425, + "grad_norm": 1.764147400856018, + "learning_rate": 3.3326563769293256e-06, + "loss": 2.1085, + "step": 1111 + }, + { + "epoch": 1.2186301369863013, + "grad_norm": 1.5984162092208862, + "learning_rate": 3.3306255077173033e-06, + "loss": 2.1208, + "step": 1112 + }, + { + "epoch": 1.2197260273972603, + "grad_norm": 1.6280403137207031, + "learning_rate": 3.32859463850528e-06, + "loss": 2.1226, + "step": 1113 + }, + { + "epoch": 1.2208219178082191, + "grad_norm": 1.5922425985336304, + "learning_rate": 3.326563769293258e-06, + "loss": 2.071, + "step": 1114 + }, + { + "epoch": 1.2219178082191782, + "grad_norm": 1.5941506624221802, + "learning_rate": 3.3245329000812347e-06, + "loss": 2.1059, + "step": 1115 + }, + { + "epoch": 1.223013698630137, + "grad_norm": 1.7812453508377075, + "learning_rate": 3.3225020308692124e-06, + "loss": 2.1192, + "step": 1116 + }, + { + "epoch": 1.2241095890410958, + "grad_norm": 1.6033459901809692, + "learning_rate": 3.3204711616571893e-06, + "loss": 2.0898, + "step": 1117 + }, + { + "epoch": 1.2252054794520548, + "grad_norm": 1.7053911685943604, + "learning_rate": 3.318440292445167e-06, + "loss": 2.1237, + "step": 1118 + }, + { + "epoch": 1.2263013698630136, + "grad_norm": 1.699777364730835, + "learning_rate": 3.316409423233144e-06, + "loss": 2.1014, + "step": 1119 + }, + { + "epoch": 1.2273972602739727, + "grad_norm": 1.7440428733825684, + "learning_rate": 3.3143785540211215e-06, + "loss": 2.1777, + "step": 1120 + }, + { + "epoch": 1.2273972602739727, + "eval_loss": 2.2857329845428467, + "eval_runtime": 37.7298, + "eval_samples_per_second": 26.504, + "eval_steps_per_second": 1.113, + "step": 1120 + }, + { + "epoch": 1.2284931506849315, + "grad_norm": 1.6901912689208984, + "learning_rate": 3.3123476848090984e-06, + "loss": 2.1373, + "step": 1121 + }, + { + "epoch": 1.2295890410958905, + "grad_norm": 1.5694825649261475, + "learning_rate": 3.310316815597076e-06, + "loss": 2.1285, + "step": 1122 + }, + { + "epoch": 1.2306849315068493, + "grad_norm": 1.6858313083648682, + "learning_rate": 3.308285946385053e-06, + "loss": 2.1858, + "step": 1123 + }, + { + "epoch": 1.2317808219178081, + "grad_norm": 1.6308691501617432, + "learning_rate": 3.30625507717303e-06, + "loss": 2.0793, + "step": 1124 + }, + { + "epoch": 1.2328767123287672, + "grad_norm": 1.6708226203918457, + "learning_rate": 3.3042242079610075e-06, + "loss": 2.1445, + "step": 1125 + }, + { + "epoch": 1.233972602739726, + "grad_norm": 1.7215708494186401, + "learning_rate": 3.3021933387489847e-06, + "loss": 2.159, + "step": 1126 + }, + { + "epoch": 1.235068493150685, + "grad_norm": 1.5764784812927246, + "learning_rate": 3.300162469536962e-06, + "loss": 2.0993, + "step": 1127 + }, + { + "epoch": 1.2361643835616438, + "grad_norm": 1.6322427988052368, + "learning_rate": 3.2981316003249393e-06, + "loss": 2.1643, + "step": 1128 + }, + { + "epoch": 1.2372602739726029, + "grad_norm": 1.6528997421264648, + "learning_rate": 3.2961007311129166e-06, + "loss": 2.1332, + "step": 1129 + }, + { + "epoch": 1.2383561643835617, + "grad_norm": 1.6434602737426758, + "learning_rate": 3.294069861900894e-06, + "loss": 2.1875, + "step": 1130 + }, + { + "epoch": 1.2383561643835617, + "eval_loss": 2.2840054035186768, + "eval_runtime": 37.7227, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 1130 + }, + { + "epoch": 1.2394520547945205, + "grad_norm": 1.6974412202835083, + "learning_rate": 3.292038992688871e-06, + "loss": 2.1396, + "step": 1131 + }, + { + "epoch": 1.2405479452054795, + "grad_norm": 1.7754731178283691, + "learning_rate": 3.2900081234768484e-06, + "loss": 2.1448, + "step": 1132 + }, + { + "epoch": 1.2416438356164383, + "grad_norm": 1.5928994417190552, + "learning_rate": 3.2879772542648257e-06, + "loss": 2.1665, + "step": 1133 + }, + { + "epoch": 1.2427397260273974, + "grad_norm": 1.609265685081482, + "learning_rate": 3.285946385052803e-06, + "loss": 2.0787, + "step": 1134 + }, + { + "epoch": 1.2438356164383562, + "grad_norm": 1.7366653680801392, + "learning_rate": 3.2839155158407802e-06, + "loss": 2.1453, + "step": 1135 + }, + { + "epoch": 1.244931506849315, + "grad_norm": 1.5945746898651123, + "learning_rate": 3.2818846466287575e-06, + "loss": 2.0705, + "step": 1136 + }, + { + "epoch": 1.246027397260274, + "grad_norm": 1.55001962184906, + "learning_rate": 3.2798537774167344e-06, + "loss": 2.1802, + "step": 1137 + }, + { + "epoch": 1.2471232876712328, + "grad_norm": 1.6575016975402832, + "learning_rate": 3.277822908204712e-06, + "loss": 2.1205, + "step": 1138 + }, + { + "epoch": 1.2482191780821918, + "grad_norm": 1.7479991912841797, + "learning_rate": 3.275792038992689e-06, + "loss": 2.0747, + "step": 1139 + }, + { + "epoch": 1.2493150684931507, + "grad_norm": 1.5351295471191406, + "learning_rate": 3.2737611697806666e-06, + "loss": 2.1263, + "step": 1140 + }, + { + "epoch": 1.2493150684931507, + "eval_loss": 2.286621570587158, + "eval_runtime": 37.744, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1140 + }, + { + "epoch": 1.2504109589041095, + "grad_norm": 1.6440327167510986, + "learning_rate": 3.2717303005686435e-06, + "loss": 2.0967, + "step": 1141 + }, + { + "epoch": 1.2515068493150685, + "grad_norm": 1.6527445316314697, + "learning_rate": 3.269699431356621e-06, + "loss": 2.1375, + "step": 1142 + }, + { + "epoch": 1.2526027397260273, + "grad_norm": 1.6786249876022339, + "learning_rate": 3.267668562144598e-06, + "loss": 2.1292, + "step": 1143 + }, + { + "epoch": 1.2536986301369863, + "grad_norm": 1.693476676940918, + "learning_rate": 3.2656376929325757e-06, + "loss": 2.1008, + "step": 1144 + }, + { + "epoch": 1.2547945205479452, + "grad_norm": 1.5634204149246216, + "learning_rate": 3.2636068237205526e-06, + "loss": 2.1504, + "step": 1145 + }, + { + "epoch": 1.2558904109589042, + "grad_norm": 1.6024060249328613, + "learning_rate": 3.2615759545085303e-06, + "loss": 2.1347, + "step": 1146 + }, + { + "epoch": 1.256986301369863, + "grad_norm": 1.7107738256454468, + "learning_rate": 3.259545085296507e-06, + "loss": 2.1722, + "step": 1147 + }, + { + "epoch": 1.2580821917808218, + "grad_norm": 1.6810829639434814, + "learning_rate": 3.257514216084484e-06, + "loss": 2.0948, + "step": 1148 + }, + { + "epoch": 1.2591780821917808, + "grad_norm": 1.6437987089157104, + "learning_rate": 3.2554833468724617e-06, + "loss": 2.1677, + "step": 1149 + }, + { + "epoch": 1.2602739726027397, + "grad_norm": 1.6707998514175415, + "learning_rate": 3.2534524776604385e-06, + "loss": 2.136, + "step": 1150 + }, + { + "epoch": 1.2602739726027397, + "eval_loss": 2.2854597568511963, + "eval_runtime": 37.72, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 1150 + }, + { + "epoch": 1.2613698630136987, + "grad_norm": 1.653172492980957, + "learning_rate": 3.2514216084484162e-06, + "loss": 2.1278, + "step": 1151 + }, + { + "epoch": 1.2624657534246575, + "grad_norm": 1.6651742458343506, + "learning_rate": 3.249390739236393e-06, + "loss": 2.1401, + "step": 1152 + }, + { + "epoch": 1.2635616438356165, + "grad_norm": 1.6661403179168701, + "learning_rate": 3.2473598700243708e-06, + "loss": 2.1067, + "step": 1153 + }, + { + "epoch": 1.2646575342465753, + "grad_norm": 1.6470211744308472, + "learning_rate": 3.2453290008123476e-06, + "loss": 2.1309, + "step": 1154 + }, + { + "epoch": 1.2657534246575342, + "grad_norm": 1.8089711666107178, + "learning_rate": 3.2432981316003253e-06, + "loss": 2.139, + "step": 1155 + }, + { + "epoch": 1.2668493150684932, + "grad_norm": 1.7904900312423706, + "learning_rate": 3.241267262388302e-06, + "loss": 2.0906, + "step": 1156 + }, + { + "epoch": 1.267945205479452, + "grad_norm": 1.6502841711044312, + "learning_rate": 3.23923639317628e-06, + "loss": 2.1263, + "step": 1157 + }, + { + "epoch": 1.269041095890411, + "grad_norm": 1.5825791358947754, + "learning_rate": 3.2372055239642567e-06, + "loss": 2.098, + "step": 1158 + }, + { + "epoch": 1.2701369863013698, + "grad_norm": 1.7173693180084229, + "learning_rate": 3.2351746547522344e-06, + "loss": 2.1367, + "step": 1159 + }, + { + "epoch": 1.2712328767123289, + "grad_norm": 1.766738772392273, + "learning_rate": 3.2331437855402113e-06, + "loss": 2.1914, + "step": 1160 + }, + { + "epoch": 1.2712328767123289, + "eval_loss": 2.282374858856201, + "eval_runtime": 37.7395, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1160 + }, + { + "epoch": 1.2723287671232877, + "grad_norm": 1.6669079065322876, + "learning_rate": 3.2311129163281885e-06, + "loss": 2.0959, + "step": 1161 + }, + { + "epoch": 1.2734246575342465, + "grad_norm": 1.7150713205337524, + "learning_rate": 3.229082047116166e-06, + "loss": 2.1431, + "step": 1162 + }, + { + "epoch": 1.2745205479452055, + "grad_norm": 1.5927042961120605, + "learning_rate": 3.227051177904143e-06, + "loss": 2.1239, + "step": 1163 + }, + { + "epoch": 1.2756164383561643, + "grad_norm": 1.8086377382278442, + "learning_rate": 3.2250203086921204e-06, + "loss": 2.1307, + "step": 1164 + }, + { + "epoch": 1.2767123287671232, + "grad_norm": 1.8078030347824097, + "learning_rate": 3.2229894394800977e-06, + "loss": 2.1632, + "step": 1165 + }, + { + "epoch": 1.2778082191780822, + "grad_norm": 1.7087128162384033, + "learning_rate": 3.220958570268075e-06, + "loss": 2.0663, + "step": 1166 + }, + { + "epoch": 1.2789041095890412, + "grad_norm": 1.8242524862289429, + "learning_rate": 3.218927701056052e-06, + "loss": 2.1225, + "step": 1167 + }, + { + "epoch": 1.28, + "grad_norm": 1.6094539165496826, + "learning_rate": 3.2168968318440295e-06, + "loss": 2.0586, + "step": 1168 + }, + { + "epoch": 1.2810958904109588, + "grad_norm": 1.8592051267623901, + "learning_rate": 3.2148659626320068e-06, + "loss": 2.1632, + "step": 1169 + }, + { + "epoch": 1.2821917808219179, + "grad_norm": 1.6238080263137817, + "learning_rate": 3.212835093419984e-06, + "loss": 2.1002, + "step": 1170 + }, + { + "epoch": 1.2821917808219179, + "eval_loss": 2.285175323486328, + "eval_runtime": 37.7569, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1170 + }, + { + "epoch": 1.2832876712328767, + "grad_norm": 1.631516933441162, + "learning_rate": 3.2108042242079613e-06, + "loss": 2.1094, + "step": 1171 + }, + { + "epoch": 1.2843835616438355, + "grad_norm": 1.6518285274505615, + "learning_rate": 3.2087733549959386e-06, + "loss": 2.1614, + "step": 1172 + }, + { + "epoch": 1.2854794520547945, + "grad_norm": 1.64463210105896, + "learning_rate": 3.206742485783916e-06, + "loss": 2.0863, + "step": 1173 + }, + { + "epoch": 1.2865753424657536, + "grad_norm": 1.5903335809707642, + "learning_rate": 3.2047116165718927e-06, + "loss": 2.0795, + "step": 1174 + }, + { + "epoch": 1.2876712328767124, + "grad_norm": 1.6504883766174316, + "learning_rate": 3.2026807473598704e-06, + "loss": 2.1352, + "step": 1175 + }, + { + "epoch": 1.2887671232876712, + "grad_norm": 1.7765742540359497, + "learning_rate": 3.2006498781478473e-06, + "loss": 2.1511, + "step": 1176 + }, + { + "epoch": 1.2898630136986302, + "grad_norm": 1.7253892421722412, + "learning_rate": 3.198619008935825e-06, + "loss": 2.0768, + "step": 1177 + }, + { + "epoch": 1.290958904109589, + "grad_norm": 1.7327924966812134, + "learning_rate": 3.196588139723802e-06, + "loss": 2.1196, + "step": 1178 + }, + { + "epoch": 1.2920547945205478, + "grad_norm": 1.6659244298934937, + "learning_rate": 3.1945572705117795e-06, + "loss": 2.1398, + "step": 1179 + }, + { + "epoch": 1.2931506849315069, + "grad_norm": 1.8803845643997192, + "learning_rate": 3.1925264012997564e-06, + "loss": 2.1382, + "step": 1180 + }, + { + "epoch": 1.2931506849315069, + "eval_loss": 2.284956455230713, + "eval_runtime": 37.7442, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1180 + }, + { + "epoch": 1.2942465753424657, + "grad_norm": 1.6562074422836304, + "learning_rate": 3.190495532087734e-06, + "loss": 2.1403, + "step": 1181 + }, + { + "epoch": 1.2953424657534247, + "grad_norm": 1.6407029628753662, + "learning_rate": 3.188464662875711e-06, + "loss": 2.1782, + "step": 1182 + }, + { + "epoch": 1.2964383561643835, + "grad_norm": 1.638464331626892, + "learning_rate": 3.1864337936636886e-06, + "loss": 2.1129, + "step": 1183 + }, + { + "epoch": 1.2975342465753426, + "grad_norm": 1.6085008382797241, + "learning_rate": 3.1844029244516655e-06, + "loss": 2.0967, + "step": 1184 + }, + { + "epoch": 1.2986301369863014, + "grad_norm": 1.9340208768844604, + "learning_rate": 3.182372055239643e-06, + "loss": 2.0906, + "step": 1185 + }, + { + "epoch": 1.2997260273972602, + "grad_norm": 1.6958277225494385, + "learning_rate": 3.18034118602762e-06, + "loss": 2.1519, + "step": 1186 + }, + { + "epoch": 1.3008219178082192, + "grad_norm": 1.6682950258255005, + "learning_rate": 3.1783103168155973e-06, + "loss": 2.1286, + "step": 1187 + }, + { + "epoch": 1.301917808219178, + "grad_norm": 1.7564821243286133, + "learning_rate": 3.1762794476035746e-06, + "loss": 2.1153, + "step": 1188 + }, + { + "epoch": 1.303013698630137, + "grad_norm": 1.765539526939392, + "learning_rate": 3.174248578391552e-06, + "loss": 2.112, + "step": 1189 + }, + { + "epoch": 1.3041095890410959, + "grad_norm": 1.8028367757797241, + "learning_rate": 3.172217709179529e-06, + "loss": 2.1169, + "step": 1190 + }, + { + "epoch": 1.3041095890410959, + "eval_loss": 2.2839744091033936, + "eval_runtime": 37.7782, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 1190 + }, + { + "epoch": 1.305205479452055, + "grad_norm": 1.6870243549346924, + "learning_rate": 3.1701868399675064e-06, + "loss": 2.1147, + "step": 1191 + }, + { + "epoch": 1.3063013698630137, + "grad_norm": 1.6161645650863647, + "learning_rate": 3.1681559707554837e-06, + "loss": 2.081, + "step": 1192 + }, + { + "epoch": 1.3073972602739725, + "grad_norm": 1.5438032150268555, + "learning_rate": 3.166125101543461e-06, + "loss": 2.139, + "step": 1193 + }, + { + "epoch": 1.3084931506849315, + "grad_norm": 1.6329416036605835, + "learning_rate": 3.1640942323314382e-06, + "loss": 2.1346, + "step": 1194 + }, + { + "epoch": 1.3095890410958904, + "grad_norm": 1.6933001279830933, + "learning_rate": 3.1620633631194155e-06, + "loss": 2.0906, + "step": 1195 + }, + { + "epoch": 1.3106849315068494, + "grad_norm": 1.5968612432479858, + "learning_rate": 3.1600324939073928e-06, + "loss": 2.1191, + "step": 1196 + }, + { + "epoch": 1.3117808219178082, + "grad_norm": 1.654739499092102, + "learning_rate": 3.15800162469537e-06, + "loss": 2.1253, + "step": 1197 + }, + { + "epoch": 1.3128767123287672, + "grad_norm": 1.6595433950424194, + "learning_rate": 3.155970755483347e-06, + "loss": 2.1326, + "step": 1198 + }, + { + "epoch": 1.313972602739726, + "grad_norm": 1.5709439516067505, + "learning_rate": 3.1539398862713246e-06, + "loss": 2.1427, + "step": 1199 + }, + { + "epoch": 1.3150684931506849, + "grad_norm": 1.7586764097213745, + "learning_rate": 3.1519090170593015e-06, + "loss": 2.1311, + "step": 1200 + }, + { + "epoch": 1.3150684931506849, + "eval_loss": 2.2803795337677, + "eval_runtime": 37.7752, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 1200 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 3.740771291234304e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/latest new file mode 100644 index 0000000000000000000000000000000000000000..10b09fac99bc80ff931649e8b3378aab683b28be --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/latest @@ -0,0 +1 @@ +global_step1600 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..99e3e04b82badc26523041bbb4ea661b7f5b9544 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/trainer_state.json @@ -0,0 +1,12513 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.7534246575342465, + "eval_steps": 10, + "global_step": 1600, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9794018268585205, + "learning_rate": 1.551094890510949e-06, + "loss": 2.3047, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.396454334259033, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.3671, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 2.0826210975646973, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.4105, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.3590970039367676, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.3405, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 2.3945844173431396, + "learning_rate": 1.624087591240876e-06, + "loss": 2.2922, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 2.185126781463623, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.3399, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.2687461376190186, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.3547, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.920499324798584, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.28, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.127774715423584, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.3409, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 1.8677122592926025, + "learning_rate": 1.715328467153285e-06, + "loss": 2.3406, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.4506173133850098, + "eval_runtime": 37.5008, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9803308248519897, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.3642, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 1.7733794450759888, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.3666, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 1.7639747858047485, + "learning_rate": 1.77007299270073e-06, + "loss": 2.3926, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 1.9172815084457397, + "learning_rate": 1.788321167883212e-06, + "loss": 2.3431, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 1.580621361732483, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.2967, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.0889124870300293, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.331, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.6659903526306152, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.3181, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.028651237487793, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.3237, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.681212306022644, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.32, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 1.9986696243286133, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.3425, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.4408469200134277, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 1.832582950592041, + "learning_rate": 1.916058394160584e-06, + "loss": 2.2548, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 1.8609144687652588, + "learning_rate": 1.934306569343066e-06, + "loss": 2.3024, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 1.8699817657470703, + "learning_rate": 1.952554744525548e-06, + "loss": 2.3044, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 1.81354558467865, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.3098, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 1.7789452075958252, + "learning_rate": 1.989051094890511e-06, + "loss": 2.3623, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 1.6322330236434937, + "learning_rate": 2.007299270072993e-06, + "loss": 2.303, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 1.8742916584014893, + "learning_rate": 2.025547445255475e-06, + "loss": 2.3183, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.7790557146072388, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.3168, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 1.7624462842941284, + "learning_rate": 2.062043795620438e-06, + "loss": 2.3629, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.781114101409912, + "learning_rate": 2.08029197080292e-06, + "loss": 2.2813, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.426335334777832, + "eval_runtime": 37.7146, + "eval_samples_per_second": 26.515, + "eval_steps_per_second": 1.114, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.769988775253296, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.3396, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.0640275478363037, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.3688, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.7884094715118408, + "learning_rate": 2.135036496350365e-06, + "loss": 2.2775, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.831847071647644, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.2966, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9044525623321533, + "learning_rate": 2.171532846715329e-06, + "loss": 2.3359, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.9899096488952637, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.3099, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8570183515548706, + "learning_rate": 2.208029197080292e-06, + "loss": 2.3016, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 1.8255733251571655, + "learning_rate": 2.226277372262774e-06, + "loss": 2.335, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.8759729862213135, + "learning_rate": 2.244525547445256e-06, + "loss": 2.3357, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.9031389951705933, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.3014, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.4213266372680664, + "eval_runtime": 37.6757, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.1044118404388428, + "learning_rate": 2.281021897810219e-06, + "loss": 2.2946, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.005735158920288, + "learning_rate": 2.299270072992701e-06, + "loss": 2.3132, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.8821312189102173, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.2816, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7866829633712769, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.309, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.8452801704406738, + "learning_rate": 2.354014598540146e-06, + "loss": 2.3222, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.5781638622283936, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.2688, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 1.9746729135513306, + "learning_rate": 2.39051094890511e-06, + "loss": 2.2864, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 2.0312447547912598, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.2914, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.675674557685852, + "learning_rate": 2.427007299270073e-06, + "loss": 2.3441, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 1.8987798690795898, + "learning_rate": 2.445255474452555e-06, + "loss": 2.2589, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.4149818420410156, + "eval_runtime": 37.6627, + "eval_samples_per_second": 26.551, + "eval_steps_per_second": 1.115, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 2.041752338409424, + "learning_rate": 2.463503649635037e-06, + "loss": 2.2457, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.7845569849014282, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.3376, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 2.489330530166626, + "learning_rate": 2.5e-06, + "loss": 2.258, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 2.3793296813964844, + "learning_rate": 2.518248175182482e-06, + "loss": 2.2561, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 2.0474460124969482, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.2992, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.7181849479675293, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.3003, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 2.3854947090148926, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.2866, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.7228457927703857, + "learning_rate": 2.591240875912409e-06, + "loss": 2.3027, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.3928253650665283, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.3233, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.136934757232666, + "learning_rate": 2.627737226277373e-06, + "loss": 2.2949, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.410449266433716, + "eval_runtime": 37.6837, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.8906159400939941, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.2693, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 1.8668863773345947, + "learning_rate": 2.664233576642336e-06, + "loss": 2.2944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.319363594055176, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.2791, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0869688987731934, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.32, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 1.8910208940505981, + "learning_rate": 2.718978102189781e-06, + "loss": 2.3059, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.157984495162964, + "learning_rate": 2.737226277372263e-06, + "loss": 2.2756, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.387314558029175, + "learning_rate": 2.755474452554745e-06, + "loss": 2.3597, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.183333158493042, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.322, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.158015489578247, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.2663, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 1.9594391584396362, + "learning_rate": 2.81021897810219e-06, + "loss": 2.3, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.411637306213379, + "eval_runtime": 37.6428, + "eval_samples_per_second": 26.565, + "eval_steps_per_second": 1.116, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.28957462310791, + "learning_rate": 2.828467153284672e-06, + "loss": 2.2258, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.8085683584213257, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.2739, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 1.9662787914276123, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.2398, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 2.337609052658081, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.3216, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 1.6008479595184326, + "learning_rate": 2.901459854014599e-06, + "loss": 2.2997, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 2.177690029144287, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.2893, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8746743202209473, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.3303, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 1.7862998247146606, + "learning_rate": 2.956204379562044e-06, + "loss": 2.3007, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.7578285932540894, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.2591, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.1532702445983887, + "learning_rate": 2.992700729927008e-06, + "loss": 2.2544, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.3992533683776855, + "eval_runtime": 37.6571, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.0168659687042236, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.3108, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 1.8336135149002075, + "learning_rate": 3.029197080291971e-06, + "loss": 2.3044, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.6675384044647217, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.2978, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 1.8866558074951172, + "learning_rate": 3.065693430656935e-06, + "loss": 2.3169, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0588760375976562, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.3041, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.9300280809402466, + "learning_rate": 3.102189781021898e-06, + "loss": 2.2494, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 1.837496042251587, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.2339, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.155080795288086, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.324, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 1.9562510251998901, + "learning_rate": 3.156934306569343e-06, + "loss": 2.2937, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 2.522388219833374, + "learning_rate": 3.175182481751825e-06, + "loss": 2.2683, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.3936216831207275, + "eval_runtime": 37.6499, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 1.7409396171569824, + "learning_rate": 3.193430656934307e-06, + "loss": 2.2833, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.176687002182007, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.2748, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 1.8576542139053345, + "learning_rate": 3.2299270072992707e-06, + "loss": 2.2144, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.9564870595932007, + "learning_rate": 3.248175182481752e-06, + "loss": 2.2898, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.896238088607788, + "learning_rate": 3.266423357664234e-06, + "loss": 2.2793, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.735960841178894, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.2869, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.018303394317627, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.2775, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 1.7448145151138306, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.2736, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8245689868927002, + "learning_rate": 3.339416058394161e-06, + "loss": 2.2799, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 1.821118950843811, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.2823, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.3893656730651855, + "eval_runtime": 37.676, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 1.9112446308135986, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.287, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 1.6978867053985596, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.2865, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9376300573349, + "learning_rate": 3.412408759124088e-06, + "loss": 2.2326, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.8067563772201538, + "learning_rate": 3.43065693430657e-06, + "loss": 2.2739, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.8476370573043823, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.2328, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.927923321723938, + "learning_rate": 3.467153284671533e-06, + "loss": 2.2679, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7956453561782837, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.2607, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 1.857413649559021, + "learning_rate": 3.503649635036497e-06, + "loss": 2.2854, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 1.8384093046188354, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.3021, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.7346371412277222, + "learning_rate": 3.54014598540146e-06, + "loss": 2.3402, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.390045642852783, + "eval_runtime": 37.4358, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.7546677589416504, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.3214, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 1.7437881231307983, + "learning_rate": 3.576642335766424e-06, + "loss": 2.2004, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9509227275848389, + "learning_rate": 3.5948905109489054e-06, + "loss": 2.2044, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 1.6918139457702637, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.2369, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9664885997772217, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.254, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 1.7942848205566406, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.2296, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 1.8610197305679321, + "learning_rate": 3.667883211678833e-06, + "loss": 2.2273, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 1.8405956029891968, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.3034, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 1.7151877880096436, + "learning_rate": 3.704379562043796e-06, + "loss": 2.2689, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 1.8795795440673828, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.2363, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.384481430053711, + "eval_runtime": 37.6512, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 1.7494690418243408, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.257, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 1.7700639963150024, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.2457, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 1.846092700958252, + "learning_rate": 3.777372262773723e-06, + "loss": 2.2425, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 1.6699751615524292, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.2734, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 1.8978508710861206, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.2435, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.6387466192245483, + "learning_rate": 3.832116788321168e-06, + "loss": 2.2241, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 1.6999541521072388, + "learning_rate": 3.85036496350365e-06, + "loss": 2.2468, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 1.6322710514068604, + "learning_rate": 3.868613138686132e-06, + "loss": 2.2642, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.5633572340011597, + "learning_rate": 3.886861313868613e-06, + "loss": 2.2862, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 1.7642648220062256, + "learning_rate": 3.905109489051096e-06, + "loss": 2.2414, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.3775289058685303, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.76724374294281, + "learning_rate": 3.923357664233577e-06, + "loss": 2.2976, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.6049190759658813, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.3034, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.6591981649398804, + "learning_rate": 3.95985401459854e-06, + "loss": 2.1957, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.650676965713501, + "learning_rate": 3.978102189781022e-06, + "loss": 2.275, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.6665960550308228, + "learning_rate": 3.996350364963504e-06, + "loss": 2.1853, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.5785901546478271, + "learning_rate": 4.014598540145986e-06, + "loss": 2.2399, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.976162075996399, + "learning_rate": 4.032846715328467e-06, + "loss": 2.2499, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.642327904701233, + "learning_rate": 4.05109489051095e-06, + "loss": 2.235, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 2.0126261711120605, + "learning_rate": 4.069343065693431e-06, + "loss": 2.2466, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.6737161874771118, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.2984, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.375901222229004, + "eval_runtime": 37.7072, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 2.094257354736328, + "learning_rate": 4.105839416058395e-06, + "loss": 2.251, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.655393123626709, + "learning_rate": 4.124087591240876e-06, + "loss": 2.2366, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 2.5909082889556885, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.2986, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.8278532028198242, + "learning_rate": 4.16058394160584e-06, + "loss": 2.1977, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9404895305633545, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.2578, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.6525582075119019, + "learning_rate": 4.197080291970803e-06, + "loss": 2.2263, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 2.021822929382324, + "learning_rate": 4.215328467153285e-06, + "loss": 2.2326, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 1.9136033058166504, + "learning_rate": 4.233576642335767e-06, + "loss": 2.2457, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.389714002609253, + "learning_rate": 4.251824817518249e-06, + "loss": 2.2256, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.7886695861816406, + "learning_rate": 4.27007299270073e-06, + "loss": 2.2222, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.3729844093322754, + "eval_runtime": 37.7638, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.9499074220657349, + "learning_rate": 4.288321167883212e-06, + "loss": 2.2026, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 1.834797978401184, + "learning_rate": 4.306569343065693e-06, + "loss": 2.2564, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 2.0415048599243164, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.2462, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 1.6527886390686035, + "learning_rate": 4.343065693430658e-06, + "loss": 2.2493, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.0504889488220215, + "learning_rate": 4.361313868613139e-06, + "loss": 2.2734, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.004772424697876, + "learning_rate": 4.379562043795621e-06, + "loss": 2.2469, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.1973347663879395, + "learning_rate": 4.397810218978102e-06, + "loss": 2.2479, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.8222625255584717, + "learning_rate": 4.416058394160584e-06, + "loss": 2.2206, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 2.1759912967681885, + "learning_rate": 4.434306569343066e-06, + "loss": 2.261, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.8344624042510986, + "learning_rate": 4.452554744525548e-06, + "loss": 2.2605, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.3706893920898438, + "eval_runtime": 37.734, + "eval_samples_per_second": 26.501, + "eval_steps_per_second": 1.113, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.7783704996109009, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.284, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 2.101193428039551, + "learning_rate": 4.489051094890512e-06, + "loss": 2.2524, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 1.7923569679260254, + "learning_rate": 4.507299270072993e-06, + "loss": 2.2434, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 2.170883893966675, + "learning_rate": 4.525547445255475e-06, + "loss": 2.2173, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.6749192476272583, + "learning_rate": 4.543795620437956e-06, + "loss": 2.283, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.0484230518341064, + "learning_rate": 4.562043795620438e-06, + "loss": 2.2931, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.9212749004364014, + "learning_rate": 4.580291970802921e-06, + "loss": 2.2661, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.6729028224945068, + "learning_rate": 4.598540145985402e-06, + "loss": 2.2269, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 2.1190619468688965, + "learning_rate": 4.6167883211678835e-06, + "loss": 2.2278, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.7444697618484497, + "learning_rate": 4.635036496350365e-06, + "loss": 2.2634, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.3706092834472656, + "eval_runtime": 37.7391, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 1.8921526670455933, + "learning_rate": 4.653284671532847e-06, + "loss": 2.2697, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7347339391708374, + "learning_rate": 4.671532846715329e-06, + "loss": 2.2406, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 2.1617953777313232, + "learning_rate": 4.689781021897811e-06, + "loss": 2.2891, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.873384714126587, + "learning_rate": 4.708029197080292e-06, + "loss": 2.2554, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.801356554031372, + "learning_rate": 4.726277372262774e-06, + "loss": 2.2073, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.9250633716583252, + "learning_rate": 4.744525547445255e-06, + "loss": 2.2501, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 1.8360860347747803, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.2561, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 1.8711498975753784, + "learning_rate": 4.78102189781022e-06, + "loss": 2.2247, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.6604810953140259, + "learning_rate": 4.799270072992701e-06, + "loss": 2.2448, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 1.689201831817627, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.2494, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.3682427406311035, + "eval_runtime": 37.7233, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.720961093902588, + "learning_rate": 4.835766423357665e-06, + "loss": 2.2013, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.7863153219223022, + "learning_rate": 4.854014598540146e-06, + "loss": 2.2183, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 1.8583060503005981, + "learning_rate": 4.872262773722628e-06, + "loss": 2.2013, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 1.722428560256958, + "learning_rate": 4.89051094890511e-06, + "loss": 2.2399, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.8408993482589722, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.2489, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.887604832649231, + "learning_rate": 4.927007299270074e-06, + "loss": 2.2145, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 1.7112189531326294, + "learning_rate": 4.945255474452555e-06, + "loss": 2.2915, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 1.8314064741134644, + "learning_rate": 4.963503649635037e-06, + "loss": 2.2786, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.7001186609268188, + "learning_rate": 4.981751824817518e-06, + "loss": 2.249, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.5820505619049072, + "learning_rate": 5e-06, + "loss": 2.254, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.3624653816223145, + "eval_runtime": 37.7283, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 1.8516303300857544, + "learning_rate": 4.997969130787978e-06, + "loss": 2.1894, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 1.8591808080673218, + "learning_rate": 4.995938261575955e-06, + "loss": 2.2537, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.6822415590286255, + "learning_rate": 4.993907392363932e-06, + "loss": 2.1978, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 2.0779759883880615, + "learning_rate": 4.9918765231519095e-06, + "loss": 2.2286, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.6937508583068848, + "learning_rate": 4.989845653939887e-06, + "loss": 2.2339, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.9342437982559204, + "learning_rate": 4.987814784727864e-06, + "loss": 2.2145, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.773258924484253, + "learning_rate": 4.985783915515841e-06, + "loss": 2.2368, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.8196488618850708, + "learning_rate": 4.983753046303819e-06, + "loss": 2.2224, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.6786775588989258, + "learning_rate": 4.981722177091796e-06, + "loss": 2.2439, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.7761427164077759, + "learning_rate": 4.979691307879773e-06, + "loss": 2.236, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.3614282608032227, + "eval_runtime": 37.7314, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.8389159440994263, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.2745, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.7774744033813477, + "learning_rate": 4.975629569455728e-06, + "loss": 2.2122, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.7862017154693604, + "learning_rate": 4.973598700243705e-06, + "loss": 2.2097, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.8851722478866577, + "learning_rate": 4.971567831031681e-06, + "loss": 2.118, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.068608045578003, + "learning_rate": 4.9695369618196595e-06, + "loss": 2.1884, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.8885847330093384, + "learning_rate": 4.967506092607636e-06, + "loss": 2.2097, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.9441415071487427, + "learning_rate": 4.965475223395614e-06, + "loss": 2.2068, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 2.313577890396118, + "learning_rate": 4.9634443541835905e-06, + "loss": 2.2355, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 1.8542472124099731, + "learning_rate": 4.961413484971569e-06, + "loss": 2.2299, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9382838010787964, + "learning_rate": 4.959382615759545e-06, + "loss": 2.2072, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.361767053604126, + "eval_runtime": 37.4603, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.081242322921753, + "learning_rate": 4.957351746547523e-06, + "loss": 2.3167, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 1.9109065532684326, + "learning_rate": 4.9553208773355e-06, + "loss": 2.2336, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.124897003173828, + "learning_rate": 4.953290008123478e-06, + "loss": 2.184, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.8820433616638184, + "learning_rate": 4.951259138911454e-06, + "loss": 2.2836, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.016204595565796, + "learning_rate": 4.9492282696994315e-06, + "loss": 2.2221, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.0679776668548584, + "learning_rate": 4.947197400487409e-06, + "loss": 2.2113, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 1.8040891885757446, + "learning_rate": 4.945166531275386e-06, + "loss": 2.2292, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.0315871238708496, + "learning_rate": 4.943135662063363e-06, + "loss": 2.2657, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.71120023727417, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.2517, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0443193912506104, + "learning_rate": 4.939073923639318e-06, + "loss": 2.2105, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.3582265377044678, + "eval_runtime": 37.6749, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.9363837242126465, + "learning_rate": 4.937043054427295e-06, + "loss": 2.1944, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 2.0878167152404785, + "learning_rate": 4.935012185215272e-06, + "loss": 2.2477, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.9059010744094849, + "learning_rate": 4.93298131600325e-06, + "loss": 2.2278, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 2.220851421356201, + "learning_rate": 4.930950446791227e-06, + "loss": 2.2391, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.7555570602416992, + "learning_rate": 4.928919577579204e-06, + "loss": 2.2206, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 2.1546502113342285, + "learning_rate": 4.9268887083671815e-06, + "loss": 2.2216, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 1.9688911437988281, + "learning_rate": 4.924857839155159e-06, + "loss": 2.2423, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 1.983587622642517, + "learning_rate": 4.922826969943136e-06, + "loss": 2.1738, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 2.1990034580230713, + "learning_rate": 4.920796100731113e-06, + "loss": 2.2341, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 1.9690128564834595, + "learning_rate": 4.918765231519091e-06, + "loss": 2.1914, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.36171817779541, + "eval_runtime": 37.727, + "eval_samples_per_second": 26.506, + "eval_steps_per_second": 1.113, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 2.5950212478637695, + "learning_rate": 4.916734362307068e-06, + "loss": 2.1921, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 2.023130416870117, + "learning_rate": 4.914703493095045e-06, + "loss": 2.1932, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2486252784729004, + "learning_rate": 4.912672623883022e-06, + "loss": 2.2174, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 1.723677635192871, + "learning_rate": 4.910641754671e-06, + "loss": 2.2173, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.94509756565094, + "learning_rate": 4.908610885458977e-06, + "loss": 2.2539, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 1.6934322118759155, + "learning_rate": 4.906580016246954e-06, + "loss": 2.2345, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 1.6882652044296265, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.231, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.6730620861053467, + "learning_rate": 4.902518277822909e-06, + "loss": 2.2394, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.7261624336242676, + "learning_rate": 4.900487408610886e-06, + "loss": 2.2524, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.8310920000076294, + "learning_rate": 4.898456539398863e-06, + "loss": 2.3004, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.3536295890808105, + "eval_runtime": 37.7461, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.7025809288024902, + "learning_rate": 4.89642567018684e-06, + "loss": 2.1923, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.7285082340240479, + "learning_rate": 4.894394800974818e-06, + "loss": 2.241, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 1.7589807510375977, + "learning_rate": 4.892363931762794e-06, + "loss": 2.2019, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.6664931774139404, + "learning_rate": 4.8903330625507725e-06, + "loss": 2.1822, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 1.8308027982711792, + "learning_rate": 4.888302193338749e-06, + "loss": 2.2651, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.6426894664764404, + "learning_rate": 4.886271324126727e-06, + "loss": 2.2514, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.9765205383300781, + "learning_rate": 4.8842404549147034e-06, + "loss": 2.226, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.7872203588485718, + "learning_rate": 4.8822095857026816e-06, + "loss": 2.18, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 2.0645813941955566, + "learning_rate": 4.880178716490658e-06, + "loss": 2.2023, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 1.9823166131973267, + "learning_rate": 4.878147847278636e-06, + "loss": 2.1801, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.353776454925537, + "eval_runtime": 37.771, + "eval_samples_per_second": 26.475, + "eval_steps_per_second": 1.112, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.831152319908142, + "learning_rate": 4.8761169780666125e-06, + "loss": 2.1858, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 1.9321790933609009, + "learning_rate": 4.874086108854591e-06, + "loss": 2.1638, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.9097596406936646, + "learning_rate": 4.872055239642567e-06, + "loss": 2.2206, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8821815252304077, + "learning_rate": 4.870024370430544e-06, + "loss": 2.2338, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.7959749698638916, + "learning_rate": 4.867993501218522e-06, + "loss": 2.1882, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 1.625320315361023, + "learning_rate": 4.865962632006499e-06, + "loss": 2.2137, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 1.6914536952972412, + "learning_rate": 4.863931762794476e-06, + "loss": 2.2374, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0801584720611572, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.2517, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 1.6874804496765137, + "learning_rate": 4.859870024370431e-06, + "loss": 2.2621, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 1.9347832202911377, + "learning_rate": 4.857839155158408e-06, + "loss": 2.2326, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.3468236923217773, + "eval_runtime": 37.7184, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.6798334121704102, + "learning_rate": 4.855808285946385e-06, + "loss": 2.2222, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 1.8993884325027466, + "learning_rate": 4.8537774167343626e-06, + "loss": 2.2178, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 1.697892189025879, + "learning_rate": 4.85174654752234e-06, + "loss": 2.2217, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8484156131744385, + "learning_rate": 4.849715678310317e-06, + "loss": 2.2069, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 1.8125075101852417, + "learning_rate": 4.847684809098294e-06, + "loss": 2.2285, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 1.8215030431747437, + "learning_rate": 4.845653939886272e-06, + "loss": 2.2029, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 1.777129054069519, + "learning_rate": 4.843623070674249e-06, + "loss": 2.2512, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.7327351570129395, + "learning_rate": 4.841592201462226e-06, + "loss": 2.1767, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.7363202571868896, + "learning_rate": 4.8395613322502035e-06, + "loss": 2.2174, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9183942079544067, + "learning_rate": 4.837530463038181e-06, + "loss": 2.2383, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.345961332321167, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.7290767431259155, + "learning_rate": 4.835499593826158e-06, + "loss": 2.22, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 1.7892694473266602, + "learning_rate": 4.833468724614135e-06, + "loss": 2.1403, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.6671900749206543, + "learning_rate": 4.831437855402113e-06, + "loss": 2.1989, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 1.848751425743103, + "learning_rate": 4.82940698619009e-06, + "loss": 2.245, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.6957710981369019, + "learning_rate": 4.827376116978067e-06, + "loss": 2.2239, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.650583267211914, + "learning_rate": 4.8253452477660444e-06, + "loss": 2.2101, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.6510974168777466, + "learning_rate": 4.823314378554022e-06, + "loss": 2.204, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 1.9469026327133179, + "learning_rate": 4.821283509341998e-06, + "loss": 2.221, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 1.6031194925308228, + "learning_rate": 4.819252640129976e-06, + "loss": 2.1926, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.120325803756714, + "learning_rate": 4.817221770917953e-06, + "loss": 2.2166, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.3419017791748047, + "eval_runtime": 37.6854, + "eval_samples_per_second": 26.535, + "eval_steps_per_second": 1.114, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.5533722639083862, + "learning_rate": 4.815190901705931e-06, + "loss": 2.1981, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9650007486343384, + "learning_rate": 4.813160032493907e-06, + "loss": 2.2319, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 1.6874849796295166, + "learning_rate": 4.811129163281885e-06, + "loss": 2.2263, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.787996768951416, + "learning_rate": 4.809098294069862e-06, + "loss": 2.2546, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 1.7853440046310425, + "learning_rate": 4.80706742485784e-06, + "loss": 2.1994, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.7882641553878784, + "learning_rate": 4.805036555645816e-06, + "loss": 2.2449, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 1.8206003904342651, + "learning_rate": 4.8030056864337945e-06, + "loss": 2.2144, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.7012749910354614, + "learning_rate": 4.800974817221771e-06, + "loss": 2.2122, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 1.8711670637130737, + "learning_rate": 4.798943948009749e-06, + "loss": 2.2132, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 1.6880611181259155, + "learning_rate": 4.7969130787977254e-06, + "loss": 2.2336, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.340393543243408, + "eval_runtime": 37.6885, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 2.0543065071105957, + "learning_rate": 4.794882209585703e-06, + "loss": 2.1899, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 1.5774480104446411, + "learning_rate": 4.79285134037368e-06, + "loss": 2.1982, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 1.7154620885849, + "learning_rate": 4.790820471161657e-06, + "loss": 2.2096, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.8200151920318604, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.2391, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 1.6368298530578613, + "learning_rate": 4.786758732737612e-06, + "loss": 2.1855, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 1.6851340532302856, + "learning_rate": 4.784727863525589e-06, + "loss": 2.2114, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 1.8293818235397339, + "learning_rate": 4.782696994313566e-06, + "loss": 2.1636, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 1.6213513612747192, + "learning_rate": 4.780666125101544e-06, + "loss": 2.1884, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.7791825532913208, + "learning_rate": 4.778635255889521e-06, + "loss": 2.1977, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 1.6627994775772095, + "learning_rate": 4.776604386677498e-06, + "loss": 2.2444, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.3396637439727783, + "eval_runtime": 37.7069, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 1.7597277164459229, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.2369, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.7754322290420532, + "learning_rate": 4.772542648253453e-06, + "loss": 2.1859, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 1.7897601127624512, + "learning_rate": 4.77051177904143e-06, + "loss": 2.2245, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 1.7294282913208008, + "learning_rate": 4.768480909829407e-06, + "loss": 2.1998, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 1.8406974077224731, + "learning_rate": 4.766450040617385e-06, + "loss": 2.2353, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 1.849493384361267, + "learning_rate": 4.764419171405362e-06, + "loss": 2.2165, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.5568877458572388, + "learning_rate": 4.762388302193339e-06, + "loss": 2.1989, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 1.8816689252853394, + "learning_rate": 4.760357432981316e-06, + "loss": 2.2062, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 1.7648332118988037, + "learning_rate": 4.758326563769294e-06, + "loss": 2.2204, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.753944754600525, + "learning_rate": 4.756295694557271e-06, + "loss": 2.1748, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.339653253555298, + "eval_runtime": 37.4438, + "eval_samples_per_second": 26.707, + "eval_steps_per_second": 1.122, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.7343329191207886, + "learning_rate": 4.754264825345248e-06, + "loss": 2.1843, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 1.8125407695770264, + "learning_rate": 4.7522339561332255e-06, + "loss": 2.1492, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.5255184173583984, + "learning_rate": 4.750203086921203e-06, + "loss": 2.2178, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 1.6692545413970947, + "learning_rate": 4.74817221770918e-06, + "loss": 2.2516, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 1.7804819345474243, + "learning_rate": 4.7461413484971565e-06, + "loss": 2.1592, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8048933744430542, + "learning_rate": 4.744110479285135e-06, + "loss": 2.1869, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 1.6439207792282104, + "learning_rate": 4.742079610073111e-06, + "loss": 2.1995, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.7375173568725586, + "learning_rate": 4.740048740861089e-06, + "loss": 2.2067, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 1.9471659660339355, + "learning_rate": 4.738017871649066e-06, + "loss": 2.2513, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 1.6524564027786255, + "learning_rate": 4.735987002437044e-06, + "loss": 2.2643, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.337153434753418, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 1.8574563264846802, + "learning_rate": 4.73395613322502e-06, + "loss": 2.2373, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 1.836464524269104, + "learning_rate": 4.731925264012998e-06, + "loss": 2.1769, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 1.7012423276901245, + "learning_rate": 4.729894394800975e-06, + "loss": 2.1494, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 1.953180193901062, + "learning_rate": 4.727863525588953e-06, + "loss": 2.186, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 1.8468470573425293, + "learning_rate": 4.725832656376929e-06, + "loss": 2.2279, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 2.005077362060547, + "learning_rate": 4.723801787164907e-06, + "loss": 2.1798, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 1.8246952295303345, + "learning_rate": 4.721770917952885e-06, + "loss": 2.1916, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 1.8558839559555054, + "learning_rate": 4.719740048740861e-06, + "loss": 2.2237, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 1.7670999765396118, + "learning_rate": 4.717709179528839e-06, + "loss": 2.2179, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.8544219732284546, + "learning_rate": 4.715678310316816e-06, + "loss": 2.1811, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.333791732788086, + "eval_runtime": 37.7254, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 1.5792925357818604, + "learning_rate": 4.713647441104794e-06, + "loss": 2.1749, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 1.5678222179412842, + "learning_rate": 4.71161657189277e-06, + "loss": 2.2294, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 1.9467812776565552, + "learning_rate": 4.709585702680748e-06, + "loss": 2.2684, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 1.6647964715957642, + "learning_rate": 4.707554833468725e-06, + "loss": 2.1722, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 1.7037622928619385, + "learning_rate": 4.705523964256703e-06, + "loss": 2.1915, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.5857148170471191, + "learning_rate": 4.703493095044679e-06, + "loss": 2.1466, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 1.5792173147201538, + "learning_rate": 4.701462225832657e-06, + "loss": 2.2472, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 1.5998480319976807, + "learning_rate": 4.699431356620634e-06, + "loss": 2.2129, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.6728194952011108, + "learning_rate": 4.697400487408612e-06, + "loss": 2.2057, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 1.5402852296829224, + "learning_rate": 4.695369618196588e-06, + "loss": 2.1903, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.3325934410095215, + "eval_runtime": 37.7463, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 1.5890164375305176, + "learning_rate": 4.693338748984566e-06, + "loss": 2.1584, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 1.805923581123352, + "learning_rate": 4.691307879772543e-06, + "loss": 2.2263, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.5932700634002686, + "learning_rate": 4.68927701056052e-06, + "loss": 2.2109, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 1.5611685514450073, + "learning_rate": 4.6872461413484975e-06, + "loss": 2.192, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 1.6976194381713867, + "learning_rate": 4.685215272136475e-06, + "loss": 2.1998, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.7149630784988403, + "learning_rate": 4.683184402924452e-06, + "loss": 2.2599, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 1.7597423791885376, + "learning_rate": 4.681153533712429e-06, + "loss": 2.1765, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 1.539652943611145, + "learning_rate": 4.679122664500407e-06, + "loss": 2.1635, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.690402865409851, + "learning_rate": 4.677091795288384e-06, + "loss": 2.2417, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.7267673015594482, + "learning_rate": 4.675060926076361e-06, + "loss": 2.166, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.331728935241699, + "eval_runtime": 37.7133, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.6383073329925537, + "learning_rate": 4.673030056864338e-06, + "loss": 2.1624, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.5364140272140503, + "learning_rate": 4.670999187652316e-06, + "loss": 2.1696, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.591477870941162, + "learning_rate": 4.668968318440293e-06, + "loss": 2.2306, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 1.8121449947357178, + "learning_rate": 4.66693744922827e-06, + "loss": 2.1449, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.515539288520813, + "learning_rate": 4.6649065800162475e-06, + "loss": 2.2108, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 1.6521453857421875, + "learning_rate": 4.662875710804225e-06, + "loss": 2.1398, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 1.7074251174926758, + "learning_rate": 4.660844841592202e-06, + "loss": 2.2342, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.7688119411468506, + "learning_rate": 4.658813972380179e-06, + "loss": 2.1287, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.7170679569244385, + "learning_rate": 4.656783103168157e-06, + "loss": 2.2143, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 1.7401350736618042, + "learning_rate": 4.654752233956134e-06, + "loss": 2.1855, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.335416555404663, + "eval_runtime": 37.7115, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.6344956159591675, + "learning_rate": 4.652721364744111e-06, + "loss": 2.2205, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.629184603691101, + "learning_rate": 4.6506904955320885e-06, + "loss": 2.2271, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.769491195678711, + "learning_rate": 4.648659626320066e-06, + "loss": 2.1945, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.6513251066207886, + "learning_rate": 4.646628757108043e-06, + "loss": 2.1705, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9357609748840332, + "learning_rate": 4.6445978878960194e-06, + "loss": 2.1968, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 1.75972318649292, + "learning_rate": 4.6425670186839976e-06, + "loss": 2.2229, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 1.686574935913086, + "learning_rate": 4.640536149471974e-06, + "loss": 2.2361, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 1.655661702156067, + "learning_rate": 4.638505280259952e-06, + "loss": 2.2064, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.7583094835281372, + "learning_rate": 4.6364744110479285e-06, + "loss": 2.2352, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.5906383991241455, + "learning_rate": 4.634443541835907e-06, + "loss": 2.187, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.332923412322998, + "eval_runtime": 37.7496, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 1.708019733428955, + "learning_rate": 4.632412672623883e-06, + "loss": 2.1032, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 1.6350252628326416, + "learning_rate": 4.630381803411861e-06, + "loss": 2.1976, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.9037389755249023, + "learning_rate": 4.628350934199838e-06, + "loss": 2.224, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 1.7126816511154175, + "learning_rate": 4.626320064987816e-06, + "loss": 2.1925, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 1.5828406810760498, + "learning_rate": 4.624289195775792e-06, + "loss": 2.1928, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.6797102689743042, + "learning_rate": 4.62225832656377e-06, + "loss": 2.1902, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 1.6871609687805176, + "learning_rate": 4.620227457351747e-06, + "loss": 2.2024, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.8785804510116577, + "learning_rate": 4.618196588139724e-06, + "loss": 2.1778, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.6175488233566284, + "learning_rate": 4.616165718927701e-06, + "loss": 2.2438, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.7724019289016724, + "learning_rate": 4.6141348497156786e-06, + "loss": 2.2272, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.328923225402832, + "eval_runtime": 37.7443, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.7230926752090454, + "learning_rate": 4.612103980503656e-06, + "loss": 2.2021, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.7899917364120483, + "learning_rate": 4.610073111291633e-06, + "loss": 2.1854, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.5864417552947998, + "learning_rate": 4.60804224207961e-06, + "loss": 2.2042, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.875375509262085, + "learning_rate": 4.606011372867588e-06, + "loss": 2.1844, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 1.830597162246704, + "learning_rate": 4.603980503655565e-06, + "loss": 2.1474, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 1.8624461889266968, + "learning_rate": 4.601949634443542e-06, + "loss": 2.158, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.7161756753921509, + "learning_rate": 4.5999187652315195e-06, + "loss": 2.1838, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.6212890148162842, + "learning_rate": 4.597887896019497e-06, + "loss": 2.1911, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.8586416244506836, + "learning_rate": 4.595857026807474e-06, + "loss": 2.1794, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.653976559638977, + "learning_rate": 4.593826157595451e-06, + "loss": 2.2428, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.3279106616973877, + "eval_runtime": 37.739, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 2.054537057876587, + "learning_rate": 4.591795288383429e-06, + "loss": 2.2003, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.730942964553833, + "learning_rate": 4.589764419171406e-06, + "loss": 2.1816, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.7264728546142578, + "learning_rate": 4.587733549959383e-06, + "loss": 2.1615, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 1.8276281356811523, + "learning_rate": 4.5857026807473604e-06, + "loss": 2.1442, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 1.9947556257247925, + "learning_rate": 4.583671811535338e-06, + "loss": 2.2192, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 1.770601511001587, + "learning_rate": 4.581640942323315e-06, + "loss": 2.233, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 1.8946491479873657, + "learning_rate": 4.579610073111292e-06, + "loss": 2.2249, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 1.7839634418487549, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.2469, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.6298807859420776, + "learning_rate": 4.575548334687247e-06, + "loss": 2.1723, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 1.9310041666030884, + "learning_rate": 4.573517465475224e-06, + "loss": 2.2156, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.3249237537384033, + "eval_runtime": 37.7505, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 1.7901647090911865, + "learning_rate": 4.571486596263201e-06, + "loss": 2.2169, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.7739574909210205, + "learning_rate": 4.569455727051178e-06, + "loss": 2.2005, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 1.7310121059417725, + "learning_rate": 4.567424857839156e-06, + "loss": 2.2369, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.004469156265259, + "learning_rate": 4.565393988627132e-06, + "loss": 2.2051, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 1.7906666994094849, + "learning_rate": 4.5633631194151105e-06, + "loss": 2.1908, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 1.7224774360656738, + "learning_rate": 4.561332250203087e-06, + "loss": 2.1529, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 1.9950156211853027, + "learning_rate": 4.559301380991065e-06, + "loss": 2.1776, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 1.8257590532302856, + "learning_rate": 4.5572705117790414e-06, + "loss": 2.2267, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.038490056991577, + "learning_rate": 4.5552396425670196e-06, + "loss": 2.1703, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 1.7314448356628418, + "learning_rate": 4.553208773354996e-06, + "loss": 2.2409, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.327078104019165, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 2.0013606548309326, + "learning_rate": 4.551177904142974e-06, + "loss": 2.1732, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.7352690696716309, + "learning_rate": 4.5491470349309505e-06, + "loss": 2.2027, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9024865627288818, + "learning_rate": 4.547116165718929e-06, + "loss": 2.1806, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8870717287063599, + "learning_rate": 4.545085296506905e-06, + "loss": 2.1664, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.781226634979248, + "learning_rate": 4.543054427294882e-06, + "loss": 2.166, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.8994221687316895, + "learning_rate": 4.54102355808286e-06, + "loss": 2.1951, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.6999517679214478, + "learning_rate": 4.538992688870837e-06, + "loss": 2.1786, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.899590253829956, + "learning_rate": 4.536961819658814e-06, + "loss": 2.2051, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.8054970502853394, + "learning_rate": 4.5349309504467915e-06, + "loss": 2.1675, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.6732861995697021, + "learning_rate": 4.532900081234769e-06, + "loss": 2.1997, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.3242335319519043, + "eval_runtime": 37.6748, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 2.0063157081604004, + "learning_rate": 4.530869212022746e-06, + "loss": 2.1909, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.7750036716461182, + "learning_rate": 4.528838342810723e-06, + "loss": 2.1935, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9993029832839966, + "learning_rate": 4.526807473598701e-06, + "loss": 2.1753, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.6742994785308838, + "learning_rate": 4.524776604386678e-06, + "loss": 2.2205, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 1.9300428628921509, + "learning_rate": 4.522745735174655e-06, + "loss": 2.2081, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.8770017623901367, + "learning_rate": 4.520714865962632e-06, + "loss": 2.2074, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.847765326499939, + "learning_rate": 4.51868399675061e-06, + "loss": 2.1775, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.905117392539978, + "learning_rate": 4.516653127538587e-06, + "loss": 2.1511, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.7623670101165771, + "learning_rate": 4.514622258326564e-06, + "loss": 2.188, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.9232176542282104, + "learning_rate": 4.5125913891145415e-06, + "loss": 2.2109, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.3252475261688232, + "eval_runtime": 37.6923, + "eval_samples_per_second": 26.531, + "eval_steps_per_second": 1.114, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.6523535251617432, + "learning_rate": 4.510560519902519e-06, + "loss": 2.1499, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9934184551239014, + "learning_rate": 4.508529650690496e-06, + "loss": 2.1383, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.704522728919983, + "learning_rate": 4.506498781478473e-06, + "loss": 2.1892, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.9909989833831787, + "learning_rate": 4.504467912266451e-06, + "loss": 2.2422, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.7490800619125366, + "learning_rate": 4.502437043054428e-06, + "loss": 2.2008, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8620145320892334, + "learning_rate": 4.500406173842405e-06, + "loss": 2.1892, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.7069931030273438, + "learning_rate": 4.4983753046303824e-06, + "loss": 2.1816, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 2.152491807937622, + "learning_rate": 4.49634443541836e-06, + "loss": 2.2331, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.8526151180267334, + "learning_rate": 4.494313566206336e-06, + "loss": 2.1811, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.6865259408950806, + "learning_rate": 4.492282696994314e-06, + "loss": 2.1599, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.3265392780303955, + "eval_runtime": 37.7255, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.7710354328155518, + "learning_rate": 4.490251827782291e-06, + "loss": 2.1636, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.8444210290908813, + "learning_rate": 4.488220958570269e-06, + "loss": 2.1345, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.9095014333724976, + "learning_rate": 4.486190089358245e-06, + "loss": 2.2236, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.7406108379364014, + "learning_rate": 4.484159220146223e-06, + "loss": 2.2078, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.712534785270691, + "learning_rate": 4.4821283509342e-06, + "loss": 2.1809, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8921459913253784, + "learning_rate": 4.480097481722178e-06, + "loss": 2.158, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.805146336555481, + "learning_rate": 4.478066612510154e-06, + "loss": 2.1765, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 2.1620516777038574, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.2832, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.661618947982788, + "learning_rate": 4.474004874086109e-06, + "loss": 2.1397, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.732483983039856, + "learning_rate": 4.471974004874087e-06, + "loss": 2.225, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.3276846408843994, + "eval_runtime": 37.7156, + "eval_samples_per_second": 26.514, + "eval_steps_per_second": 1.114, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.8693495988845825, + "learning_rate": 4.4699431356620635e-06, + "loss": 2.1959, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8201690912246704, + "learning_rate": 4.467912266450041e-06, + "loss": 2.1882, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.9229968786239624, + "learning_rate": 4.465881397238018e-06, + "loss": 2.2006, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.6330770254135132, + "learning_rate": 4.463850528025995e-06, + "loss": 2.209, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9654744863510132, + "learning_rate": 4.4618196588139726e-06, + "loss": 2.1963, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.8808507919311523, + "learning_rate": 4.45978878960195e-06, + "loss": 2.1538, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 1.8481040000915527, + "learning_rate": 4.457757920389927e-06, + "loss": 2.1769, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 2.114439010620117, + "learning_rate": 4.455727051177904e-06, + "loss": 2.2241, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.5492331981658936, + "learning_rate": 4.453696181965882e-06, + "loss": 2.206, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 2.11787486076355, + "learning_rate": 4.451665312753859e-06, + "loss": 2.2328, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.3215034008026123, + "eval_runtime": 37.7471, + "eval_samples_per_second": 26.492, + "eval_steps_per_second": 1.113, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.6561832427978516, + "learning_rate": 4.449634443541836e-06, + "loss": 2.1607, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.7510966062545776, + "learning_rate": 4.4476035743298135e-06, + "loss": 2.1699, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.9424657821655273, + "learning_rate": 4.445572705117791e-06, + "loss": 2.2298, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 1.5836161375045776, + "learning_rate": 4.443541835905768e-06, + "loss": 2.1711, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.895216464996338, + "learning_rate": 4.441510966693745e-06, + "loss": 2.2252, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.3365776538848877, + "learning_rate": 4.439480097481723e-06, + "loss": 2.1309, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.6988815069198608, + "learning_rate": 4.4374492282697e-06, + "loss": 2.2023, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 2.228900671005249, + "learning_rate": 4.435418359057677e-06, + "loss": 2.2021, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.5536632537841797, + "learning_rate": 4.433387489845654e-06, + "loss": 2.1706, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.8463878631591797, + "learning_rate": 4.431356620633632e-06, + "loss": 2.1715, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.320664167404175, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.924741506576538, + "learning_rate": 4.429325751421609e-06, + "loss": 2.1894, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 1.5921828746795654, + "learning_rate": 4.427294882209586e-06, + "loss": 2.1358, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 2.4617693424224854, + "learning_rate": 4.4252640129975635e-06, + "loss": 2.1918, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.6239196062088013, + "learning_rate": 4.423233143785541e-06, + "loss": 2.1996, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.952112078666687, + "learning_rate": 4.421202274573518e-06, + "loss": 2.112, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.663409948348999, + "learning_rate": 4.4191714053614945e-06, + "loss": 2.149, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.831711769104004, + "learning_rate": 4.417140536149473e-06, + "loss": 2.209, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 1.9281738996505737, + "learning_rate": 4.415109666937449e-06, + "loss": 2.1637, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.5661600828170776, + "learning_rate": 4.413078797725427e-06, + "loss": 2.1832, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.6930522918701172, + "learning_rate": 4.411047928513404e-06, + "loss": 2.196, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.318861246109009, + "eval_runtime": 37.6994, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.74558424949646, + "learning_rate": 4.409017059301382e-06, + "loss": 2.2097, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.664880633354187, + "learning_rate": 4.406986190089358e-06, + "loss": 2.2076, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.5981497764587402, + "learning_rate": 4.404955320877336e-06, + "loss": 2.1706, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.6035112142562866, + "learning_rate": 4.402924451665313e-06, + "loss": 2.1538, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 1.6819255352020264, + "learning_rate": 4.400893582453291e-06, + "loss": 2.1816, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.7082457542419434, + "learning_rate": 4.398862713241267e-06, + "loss": 2.1804, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.6893151998519897, + "learning_rate": 4.396831844029245e-06, + "loss": 2.1783, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 1.8639923334121704, + "learning_rate": 4.394800974817222e-06, + "loss": 2.1989, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.6407119035720825, + "learning_rate": 4.392770105605199e-06, + "loss": 2.2013, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.6435848474502563, + "learning_rate": 4.390739236393176e-06, + "loss": 2.2049, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.318084478378296, + "eval_runtime": 37.6999, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.7974674701690674, + "learning_rate": 4.388708367181154e-06, + "loss": 2.188, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 1.8340431451797485, + "learning_rate": 4.386677497969131e-06, + "loss": 2.2305, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.55428946018219, + "learning_rate": 4.384646628757108e-06, + "loss": 2.1837, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.7258585691452026, + "learning_rate": 4.3826157595450855e-06, + "loss": 2.1426, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.7347694635391235, + "learning_rate": 4.380584890333063e-06, + "loss": 2.1161, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.6082992553710938, + "learning_rate": 4.37855402112104e-06, + "loss": 2.2284, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.9446425437927246, + "learning_rate": 4.376523151909017e-06, + "loss": 2.1571, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 1.8597826957702637, + "learning_rate": 4.3744922826969946e-06, + "loss": 2.1982, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.8589327335357666, + "learning_rate": 4.372461413484972e-06, + "loss": 2.2097, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.6825448274612427, + "learning_rate": 4.370430544272949e-06, + "loss": 2.2077, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.315889596939087, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 600 + }, + { + "epoch": 0.6586301369863014, + "grad_norm": 1.8007057905197144, + "learning_rate": 4.368399675060926e-06, + "loss": 2.1843, + "step": 601 + }, + { + "epoch": 0.6597260273972603, + "grad_norm": 2.057697057723999, + "learning_rate": 4.366368805848904e-06, + "loss": 2.13, + "step": 602 + }, + { + "epoch": 0.6608219178082192, + "grad_norm": 1.5714081525802612, + "learning_rate": 4.364337936636881e-06, + "loss": 2.1879, + "step": 603 + }, + { + "epoch": 0.6619178082191781, + "grad_norm": 1.855185866355896, + "learning_rate": 4.362307067424858e-06, + "loss": 2.1659, + "step": 604 + }, + { + "epoch": 0.663013698630137, + "grad_norm": 1.826479196548462, + "learning_rate": 4.3602761982128355e-06, + "loss": 2.1716, + "step": 605 + }, + { + "epoch": 0.6641095890410958, + "grad_norm": 1.6905351877212524, + "learning_rate": 4.358245329000813e-06, + "loss": 2.1825, + "step": 606 + }, + { + "epoch": 0.6652054794520548, + "grad_norm": 1.8157447576522827, + "learning_rate": 4.35621445978879e-06, + "loss": 2.1811, + "step": 607 + }, + { + "epoch": 0.6663013698630137, + "grad_norm": 1.6311850547790527, + "learning_rate": 4.354183590576767e-06, + "loss": 2.1507, + "step": 608 + }, + { + "epoch": 0.6673972602739726, + "grad_norm": 1.579039454460144, + "learning_rate": 4.352152721364745e-06, + "loss": 2.2031, + "step": 609 + }, + { + "epoch": 0.6684931506849315, + "grad_norm": 1.7773518562316895, + "learning_rate": 4.350121852152722e-06, + "loss": 2.1978, + "step": 610 + }, + { + "epoch": 0.6684931506849315, + "eval_loss": 2.314854383468628, + "eval_runtime": 37.4501, + "eval_samples_per_second": 26.702, + "eval_steps_per_second": 1.121, + "step": 610 + }, + { + "epoch": 0.6695890410958905, + "grad_norm": 1.7606587409973145, + "learning_rate": 4.348090982940699e-06, + "loss": 2.1891, + "step": 611 + }, + { + "epoch": 0.6706849315068493, + "grad_norm": 1.6145644187927246, + "learning_rate": 4.3460601137286764e-06, + "loss": 2.1799, + "step": 612 + }, + { + "epoch": 0.6717808219178082, + "grad_norm": 1.8070226907730103, + "learning_rate": 4.344029244516654e-06, + "loss": 2.241, + "step": 613 + }, + { + "epoch": 0.6728767123287671, + "grad_norm": 1.6185903549194336, + "learning_rate": 4.341998375304631e-06, + "loss": 2.1908, + "step": 614 + }, + { + "epoch": 0.673972602739726, + "grad_norm": 1.6690049171447754, + "learning_rate": 4.339967506092607e-06, + "loss": 2.1804, + "step": 615 + }, + { + "epoch": 0.675068493150685, + "grad_norm": 1.740003228187561, + "learning_rate": 4.3379366368805855e-06, + "loss": 2.1734, + "step": 616 + }, + { + "epoch": 0.6761643835616439, + "grad_norm": 1.7311010360717773, + "learning_rate": 4.335905767668562e-06, + "loss": 2.2015, + "step": 617 + }, + { + "epoch": 0.6772602739726027, + "grad_norm": 1.7491734027862549, + "learning_rate": 4.33387489845654e-06, + "loss": 2.1894, + "step": 618 + }, + { + "epoch": 0.6783561643835616, + "grad_norm": 1.6343051195144653, + "learning_rate": 4.3318440292445165e-06, + "loss": 2.1583, + "step": 619 + }, + { + "epoch": 0.6794520547945205, + "grad_norm": 1.584243893623352, + "learning_rate": 4.329813160032495e-06, + "loss": 2.1396, + "step": 620 + }, + { + "epoch": 0.6794520547945205, + "eval_loss": 2.3157904148101807, + "eval_runtime": 37.6734, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 620 + }, + { + "epoch": 0.6805479452054795, + "grad_norm": 1.6275745630264282, + "learning_rate": 4.327782290820471e-06, + "loss": 2.2064, + "step": 621 + }, + { + "epoch": 0.6816438356164384, + "grad_norm": 1.5353672504425049, + "learning_rate": 4.325751421608449e-06, + "loss": 2.1686, + "step": 622 + }, + { + "epoch": 0.6827397260273973, + "grad_norm": 1.6155133247375488, + "learning_rate": 4.323720552396426e-06, + "loss": 2.1652, + "step": 623 + }, + { + "epoch": 0.6838356164383562, + "grad_norm": 1.638383150100708, + "learning_rate": 4.321689683184404e-06, + "loss": 2.1747, + "step": 624 + }, + { + "epoch": 0.684931506849315, + "grad_norm": 1.6153279542922974, + "learning_rate": 4.31965881397238e-06, + "loss": 2.1975, + "step": 625 + }, + { + "epoch": 0.686027397260274, + "grad_norm": 2.023699998855591, + "learning_rate": 4.3176279447603574e-06, + "loss": 2.1904, + "step": 626 + }, + { + "epoch": 0.6871232876712329, + "grad_norm": 1.5969340801239014, + "learning_rate": 4.315597075548335e-06, + "loss": 2.2133, + "step": 627 + }, + { + "epoch": 0.6882191780821918, + "grad_norm": 2.138498067855835, + "learning_rate": 4.313566206336312e-06, + "loss": 2.1685, + "step": 628 + }, + { + "epoch": 0.6893150684931507, + "grad_norm": 1.7141013145446777, + "learning_rate": 4.311535337124289e-06, + "loss": 2.1818, + "step": 629 + }, + { + "epoch": 0.6904109589041096, + "grad_norm": 1.9214462041854858, + "learning_rate": 4.3095044679122665e-06, + "loss": 2.191, + "step": 630 + }, + { + "epoch": 0.6904109589041096, + "eval_loss": 2.3150644302368164, + "eval_runtime": 37.723, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 630 + }, + { + "epoch": 0.6915068493150685, + "grad_norm": 2.03241229057312, + "learning_rate": 4.307473598700244e-06, + "loss": 2.1398, + "step": 631 + }, + { + "epoch": 0.6926027397260274, + "grad_norm": 1.608492374420166, + "learning_rate": 4.305442729488221e-06, + "loss": 2.2113, + "step": 632 + }, + { + "epoch": 0.6936986301369863, + "grad_norm": 2.1223549842834473, + "learning_rate": 4.303411860276198e-06, + "loss": 2.2012, + "step": 633 + }, + { + "epoch": 0.6947945205479452, + "grad_norm": 1.5212339162826538, + "learning_rate": 4.301380991064176e-06, + "loss": 2.2201, + "step": 634 + }, + { + "epoch": 0.6958904109589041, + "grad_norm": 1.9565715789794922, + "learning_rate": 4.299350121852153e-06, + "loss": 2.2155, + "step": 635 + }, + { + "epoch": 0.6969863013698631, + "grad_norm": 1.700822114944458, + "learning_rate": 4.29731925264013e-06, + "loss": 2.1764, + "step": 636 + }, + { + "epoch": 0.6980821917808219, + "grad_norm": 1.5625321865081787, + "learning_rate": 4.2952883834281075e-06, + "loss": 2.1407, + "step": 637 + }, + { + "epoch": 0.6991780821917808, + "grad_norm": 2.3319756984710693, + "learning_rate": 4.293257514216085e-06, + "loss": 2.2058, + "step": 638 + }, + { + "epoch": 0.7002739726027397, + "grad_norm": 1.617644190788269, + "learning_rate": 4.291226645004062e-06, + "loss": 2.2157, + "step": 639 + }, + { + "epoch": 0.7013698630136986, + "grad_norm": 1.9357377290725708, + "learning_rate": 4.289195775792039e-06, + "loss": 2.1736, + "step": 640 + }, + { + "epoch": 0.7013698630136986, + "eval_loss": 2.310267925262451, + "eval_runtime": 37.731, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 640 + }, + { + "epoch": 0.7024657534246576, + "grad_norm": 1.6908962726593018, + "learning_rate": 4.287164906580017e-06, + "loss": 2.1523, + "step": 641 + }, + { + "epoch": 0.7035616438356165, + "grad_norm": 1.6051634550094604, + "learning_rate": 4.285134037367994e-06, + "loss": 2.1499, + "step": 642 + }, + { + "epoch": 0.7046575342465753, + "grad_norm": 1.760164499282837, + "learning_rate": 4.283103168155971e-06, + "loss": 2.1493, + "step": 643 + }, + { + "epoch": 0.7057534246575342, + "grad_norm": 1.6225790977478027, + "learning_rate": 4.281072298943948e-06, + "loss": 2.1737, + "step": 644 + }, + { + "epoch": 0.7068493150684931, + "grad_norm": 1.8128079175949097, + "learning_rate": 4.279041429731926e-06, + "loss": 2.1492, + "step": 645 + }, + { + "epoch": 0.707945205479452, + "grad_norm": 1.993295431137085, + "learning_rate": 4.277010560519903e-06, + "loss": 2.1702, + "step": 646 + }, + { + "epoch": 0.709041095890411, + "grad_norm": 1.7715190649032593, + "learning_rate": 4.27497969130788e-06, + "loss": 2.1939, + "step": 647 + }, + { + "epoch": 0.7101369863013699, + "grad_norm": 1.921738624572754, + "learning_rate": 4.2729488220958575e-06, + "loss": 2.201, + "step": 648 + }, + { + "epoch": 0.7112328767123288, + "grad_norm": 1.5670042037963867, + "learning_rate": 4.270917952883835e-06, + "loss": 2.152, + "step": 649 + }, + { + "epoch": 0.7123287671232876, + "grad_norm": 1.6405668258666992, + "learning_rate": 4.268887083671812e-06, + "loss": 2.2488, + "step": 650 + }, + { + "epoch": 0.7123287671232876, + "eval_loss": 2.3100948333740234, + "eval_runtime": 37.7393, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 650 + }, + { + "epoch": 0.7134246575342466, + "grad_norm": 1.79966402053833, + "learning_rate": 4.266856214459789e-06, + "loss": 2.1506, + "step": 651 + }, + { + "epoch": 0.7145205479452055, + "grad_norm": 1.6062074899673462, + "learning_rate": 4.264825345247766e-06, + "loss": 2.1818, + "step": 652 + }, + { + "epoch": 0.7156164383561644, + "grad_norm": 1.61439049243927, + "learning_rate": 4.262794476035744e-06, + "loss": 2.1269, + "step": 653 + }, + { + "epoch": 0.7167123287671233, + "grad_norm": 1.7649872303009033, + "learning_rate": 4.26076360682372e-06, + "loss": 2.1703, + "step": 654 + }, + { + "epoch": 0.7178082191780822, + "grad_norm": 1.6238528490066528, + "learning_rate": 4.2587327376116984e-06, + "loss": 2.1764, + "step": 655 + }, + { + "epoch": 0.718904109589041, + "grad_norm": 1.6847354173660278, + "learning_rate": 4.256701868399675e-06, + "loss": 2.1796, + "step": 656 + }, + { + "epoch": 0.72, + "grad_norm": 1.7208436727523804, + "learning_rate": 4.254670999187653e-06, + "loss": 2.1903, + "step": 657 + }, + { + "epoch": 0.7210958904109589, + "grad_norm": 1.5888099670410156, + "learning_rate": 4.252640129975629e-06, + "loss": 2.1686, + "step": 658 + }, + { + "epoch": 0.7221917808219178, + "grad_norm": 1.742118000984192, + "learning_rate": 4.2506092607636075e-06, + "loss": 2.1287, + "step": 659 + }, + { + "epoch": 0.7232876712328767, + "grad_norm": 1.8450883626937866, + "learning_rate": 4.248578391551584e-06, + "loss": 2.1761, + "step": 660 + }, + { + "epoch": 0.7232876712328767, + "eval_loss": 2.3134732246398926, + "eval_runtime": 37.7568, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 660 + }, + { + "epoch": 0.7243835616438357, + "grad_norm": 1.630985140800476, + "learning_rate": 4.246547522339562e-06, + "loss": 2.1743, + "step": 661 + }, + { + "epoch": 0.7254794520547945, + "grad_norm": 1.6984914541244507, + "learning_rate": 4.2445166531275385e-06, + "loss": 2.1356, + "step": 662 + }, + { + "epoch": 0.7265753424657534, + "grad_norm": 1.6732920408248901, + "learning_rate": 4.242485783915517e-06, + "loss": 2.1377, + "step": 663 + }, + { + "epoch": 0.7276712328767123, + "grad_norm": 1.6210153102874756, + "learning_rate": 4.240454914703493e-06, + "loss": 2.1455, + "step": 664 + }, + { + "epoch": 0.7287671232876712, + "grad_norm": 1.7643382549285889, + "learning_rate": 4.23842404549147e-06, + "loss": 2.1373, + "step": 665 + }, + { + "epoch": 0.7298630136986302, + "grad_norm": 1.6491565704345703, + "learning_rate": 4.236393176279448e-06, + "loss": 2.1897, + "step": 666 + }, + { + "epoch": 0.7309589041095891, + "grad_norm": 1.670323133468628, + "learning_rate": 4.234362307067425e-06, + "loss": 2.2081, + "step": 667 + }, + { + "epoch": 0.7320547945205479, + "grad_norm": 1.6350057125091553, + "learning_rate": 4.232331437855402e-06, + "loss": 2.2002, + "step": 668 + }, + { + "epoch": 0.7331506849315068, + "grad_norm": 1.6191266775131226, + "learning_rate": 4.2303005686433795e-06, + "loss": 2.1744, + "step": 669 + }, + { + "epoch": 0.7342465753424657, + "grad_norm": 1.5511432886123657, + "learning_rate": 4.228269699431357e-06, + "loss": 2.1797, + "step": 670 + }, + { + "epoch": 0.7342465753424657, + "eval_loss": 2.313626527786255, + "eval_runtime": 37.7763, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 670 + }, + { + "epoch": 0.7353424657534247, + "grad_norm": 1.5128833055496216, + "learning_rate": 4.226238830219334e-06, + "loss": 2.1729, + "step": 671 + }, + { + "epoch": 0.7364383561643836, + "grad_norm": 1.6752032041549683, + "learning_rate": 4.224207961007311e-06, + "loss": 2.1747, + "step": 672 + }, + { + "epoch": 0.7375342465753425, + "grad_norm": 1.8643760681152344, + "learning_rate": 4.2221770917952886e-06, + "loss": 2.174, + "step": 673 + }, + { + "epoch": 0.7386301369863014, + "grad_norm": 1.5659266710281372, + "learning_rate": 4.220146222583266e-06, + "loss": 2.1706, + "step": 674 + }, + { + "epoch": 0.7397260273972602, + "grad_norm": 1.9675986766815186, + "learning_rate": 4.218115353371243e-06, + "loss": 2.1558, + "step": 675 + }, + { + "epoch": 0.7408219178082192, + "grad_norm": 1.6091173887252808, + "learning_rate": 4.21608448415922e-06, + "loss": 2.2126, + "step": 676 + }, + { + "epoch": 0.7419178082191781, + "grad_norm": 1.6801071166992188, + "learning_rate": 4.214053614947198e-06, + "loss": 2.1352, + "step": 677 + }, + { + "epoch": 0.743013698630137, + "grad_norm": 1.6964598894119263, + "learning_rate": 4.212022745735175e-06, + "loss": 2.1863, + "step": 678 + }, + { + "epoch": 0.7441095890410959, + "grad_norm": 1.670597791671753, + "learning_rate": 4.209991876523152e-06, + "loss": 2.1478, + "step": 679 + }, + { + "epoch": 0.7452054794520548, + "grad_norm": 1.6591588258743286, + "learning_rate": 4.2079610073111295e-06, + "loss": 2.2198, + "step": 680 + }, + { + "epoch": 0.7452054794520548, + "eval_loss": 2.3091039657592773, + "eval_runtime": 37.7551, + "eval_samples_per_second": 26.486, + "eval_steps_per_second": 1.112, + "step": 680 + }, + { + "epoch": 0.7463013698630137, + "grad_norm": 1.5147838592529297, + "learning_rate": 4.205930138099107e-06, + "loss": 2.1836, + "step": 681 + }, + { + "epoch": 0.7473972602739726, + "grad_norm": 1.6278952360153198, + "learning_rate": 4.203899268887084e-06, + "loss": 2.196, + "step": 682 + }, + { + "epoch": 0.7484931506849315, + "grad_norm": 1.5964363813400269, + "learning_rate": 4.201868399675061e-06, + "loss": 2.1162, + "step": 683 + }, + { + "epoch": 0.7495890410958904, + "grad_norm": 1.6993104219436646, + "learning_rate": 4.199837530463039e-06, + "loss": 2.1731, + "step": 684 + }, + { + "epoch": 0.7506849315068493, + "grad_norm": 1.6196811199188232, + "learning_rate": 4.197806661251016e-06, + "loss": 2.16, + "step": 685 + }, + { + "epoch": 0.7517808219178083, + "grad_norm": 1.6826317310333252, + "learning_rate": 4.195775792038993e-06, + "loss": 2.1123, + "step": 686 + }, + { + "epoch": 0.7528767123287671, + "grad_norm": 1.7604600191116333, + "learning_rate": 4.19374492282697e-06, + "loss": 2.1962, + "step": 687 + }, + { + "epoch": 0.753972602739726, + "grad_norm": 1.5533616542816162, + "learning_rate": 4.191714053614948e-06, + "loss": 2.1593, + "step": 688 + }, + { + "epoch": 0.7550684931506849, + "grad_norm": 1.8652232885360718, + "learning_rate": 4.189683184402924e-06, + "loss": 2.2067, + "step": 689 + }, + { + "epoch": 0.7561643835616438, + "grad_norm": 1.6330097913742065, + "learning_rate": 4.187652315190902e-06, + "loss": 2.1483, + "step": 690 + }, + { + "epoch": 0.7561643835616438, + "eval_loss": 2.309835195541382, + "eval_runtime": 37.746, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 690 + }, + { + "epoch": 0.7572602739726028, + "grad_norm": 1.5989009141921997, + "learning_rate": 4.185621445978879e-06, + "loss": 2.1686, + "step": 691 + }, + { + "epoch": 0.7583561643835617, + "grad_norm": 1.7416167259216309, + "learning_rate": 4.183590576766857e-06, + "loss": 2.2111, + "step": 692 + }, + { + "epoch": 0.7594520547945205, + "grad_norm": 1.5840280055999756, + "learning_rate": 4.181559707554833e-06, + "loss": 2.1499, + "step": 693 + }, + { + "epoch": 0.7605479452054794, + "grad_norm": 1.839306354522705, + "learning_rate": 4.179528838342811e-06, + "loss": 2.1492, + "step": 694 + }, + { + "epoch": 0.7616438356164383, + "grad_norm": 1.54163658618927, + "learning_rate": 4.177497969130788e-06, + "loss": 2.2029, + "step": 695 + }, + { + "epoch": 0.7627397260273973, + "grad_norm": 1.8050168752670288, + "learning_rate": 4.175467099918766e-06, + "loss": 2.1615, + "step": 696 + }, + { + "epoch": 0.7638356164383562, + "grad_norm": 1.6971197128295898, + "learning_rate": 4.173436230706742e-06, + "loss": 2.1476, + "step": 697 + }, + { + "epoch": 0.7649315068493151, + "grad_norm": 1.6200546026229858, + "learning_rate": 4.1714053614947205e-06, + "loss": 2.1602, + "step": 698 + }, + { + "epoch": 0.766027397260274, + "grad_norm": 1.7440853118896484, + "learning_rate": 4.169374492282697e-06, + "loss": 2.1949, + "step": 699 + }, + { + "epoch": 0.7671232876712328, + "grad_norm": 1.5590331554412842, + "learning_rate": 4.167343623070675e-06, + "loss": 2.1656, + "step": 700 + }, + { + "epoch": 0.7671232876712328, + "eval_loss": 2.309752941131592, + "eval_runtime": 37.7663, + "eval_samples_per_second": 26.479, + "eval_steps_per_second": 1.112, + "step": 700 + }, + { + "epoch": 0.7682191780821918, + "grad_norm": 1.7314256429672241, + "learning_rate": 4.165312753858652e-06, + "loss": 2.1516, + "step": 701 + }, + { + "epoch": 0.7693150684931507, + "grad_norm": 1.5145938396453857, + "learning_rate": 4.163281884646629e-06, + "loss": 2.1795, + "step": 702 + }, + { + "epoch": 0.7704109589041096, + "grad_norm": 1.5300369262695312, + "learning_rate": 4.161251015434607e-06, + "loss": 2.1356, + "step": 703 + }, + { + "epoch": 0.7715068493150685, + "grad_norm": 1.6113280057907104, + "learning_rate": 4.159220146222583e-06, + "loss": 2.157, + "step": 704 + }, + { + "epoch": 0.7726027397260274, + "grad_norm": 1.5458502769470215, + "learning_rate": 4.157189277010561e-06, + "loss": 2.2017, + "step": 705 + }, + { + "epoch": 0.7736986301369863, + "grad_norm": 1.5456187725067139, + "learning_rate": 4.155158407798538e-06, + "loss": 2.1758, + "step": 706 + }, + { + "epoch": 0.7747945205479452, + "grad_norm": 1.6051454544067383, + "learning_rate": 4.153127538586516e-06, + "loss": 2.183, + "step": 707 + }, + { + "epoch": 0.7758904109589041, + "grad_norm": 1.565788745880127, + "learning_rate": 4.151096669374492e-06, + "loss": 2.1538, + "step": 708 + }, + { + "epoch": 0.776986301369863, + "grad_norm": 1.7484439611434937, + "learning_rate": 4.1490658001624705e-06, + "loss": 2.169, + "step": 709 + }, + { + "epoch": 0.7780821917808219, + "grad_norm": 1.7206509113311768, + "learning_rate": 4.147034930950447e-06, + "loss": 2.1661, + "step": 710 + }, + { + "epoch": 0.7780821917808219, + "eval_loss": 2.307927131652832, + "eval_runtime": 37.5145, + "eval_samples_per_second": 26.656, + "eval_steps_per_second": 1.12, + "step": 710 + }, + { + "epoch": 0.7791780821917809, + "grad_norm": 1.8865350484848022, + "learning_rate": 4.145004061738425e-06, + "loss": 2.093, + "step": 711 + }, + { + "epoch": 0.7802739726027397, + "grad_norm": 1.935110092163086, + "learning_rate": 4.1429731925264015e-06, + "loss": 2.1298, + "step": 712 + }, + { + "epoch": 0.7813698630136986, + "grad_norm": 1.6938892602920532, + "learning_rate": 4.140942323314379e-06, + "loss": 2.1508, + "step": 713 + }, + { + "epoch": 0.7824657534246575, + "grad_norm": 1.5774824619293213, + "learning_rate": 4.138911454102356e-06, + "loss": 2.1808, + "step": 714 + }, + { + "epoch": 0.7835616438356164, + "grad_norm": 1.8637802600860596, + "learning_rate": 4.136880584890333e-06, + "loss": 2.1822, + "step": 715 + }, + { + "epoch": 0.7846575342465754, + "grad_norm": 1.6216785907745361, + "learning_rate": 4.1348497156783106e-06, + "loss": 2.1807, + "step": 716 + }, + { + "epoch": 0.7857534246575343, + "grad_norm": 1.7688151597976685, + "learning_rate": 4.132818846466288e-06, + "loss": 2.1967, + "step": 717 + }, + { + "epoch": 0.7868493150684932, + "grad_norm": 1.6824712753295898, + "learning_rate": 4.130787977254265e-06, + "loss": 2.1952, + "step": 718 + }, + { + "epoch": 0.787945205479452, + "grad_norm": 1.911048412322998, + "learning_rate": 4.128757108042242e-06, + "loss": 2.2408, + "step": 719 + }, + { + "epoch": 0.7890410958904109, + "grad_norm": 1.62872314453125, + "learning_rate": 4.12672623883022e-06, + "loss": 2.1647, + "step": 720 + }, + { + "epoch": 0.7890410958904109, + "eval_loss": 2.3071532249450684, + "eval_runtime": 37.6838, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 720 + }, + { + "epoch": 0.7901369863013699, + "grad_norm": 2.080674171447754, + "learning_rate": 4.124695369618197e-06, + "loss": 2.1614, + "step": 721 + }, + { + "epoch": 0.7912328767123288, + "grad_norm": 1.7225316762924194, + "learning_rate": 4.122664500406174e-06, + "loss": 2.1784, + "step": 722 + }, + { + "epoch": 0.7923287671232877, + "grad_norm": 1.838957667350769, + "learning_rate": 4.1206336311941515e-06, + "loss": 2.1281, + "step": 723 + }, + { + "epoch": 0.7934246575342466, + "grad_norm": 1.9360777139663696, + "learning_rate": 4.118602761982129e-06, + "loss": 2.1168, + "step": 724 + }, + { + "epoch": 0.7945205479452054, + "grad_norm": 1.7712862491607666, + "learning_rate": 4.116571892770106e-06, + "loss": 2.2004, + "step": 725 + }, + { + "epoch": 0.7956164383561644, + "grad_norm": 1.6759192943572998, + "learning_rate": 4.114541023558083e-06, + "loss": 2.1588, + "step": 726 + }, + { + "epoch": 0.7967123287671233, + "grad_norm": 1.8396296501159668, + "learning_rate": 4.112510154346061e-06, + "loss": 2.1777, + "step": 727 + }, + { + "epoch": 0.7978082191780822, + "grad_norm": 1.6775752305984497, + "learning_rate": 4.110479285134038e-06, + "loss": 2.1923, + "step": 728 + }, + { + "epoch": 0.7989041095890411, + "grad_norm": 1.7111576795578003, + "learning_rate": 4.108448415922015e-06, + "loss": 2.1854, + "step": 729 + }, + { + "epoch": 0.8, + "grad_norm": 1.7400974035263062, + "learning_rate": 4.1064175467099924e-06, + "loss": 2.0897, + "step": 730 + }, + { + "epoch": 0.8, + "eval_loss": 2.3070170879364014, + "eval_runtime": 37.7602, + "eval_samples_per_second": 26.483, + "eval_steps_per_second": 1.112, + "step": 730 + }, + { + "epoch": 0.8010958904109589, + "grad_norm": 1.9166353940963745, + "learning_rate": 4.10438667749797e-06, + "loss": 2.2417, + "step": 731 + }, + { + "epoch": 0.8021917808219178, + "grad_norm": 1.7914725542068481, + "learning_rate": 4.102355808285947e-06, + "loss": 2.1677, + "step": 732 + }, + { + "epoch": 0.8032876712328767, + "grad_norm": 1.7602828741073608, + "learning_rate": 4.100324939073924e-06, + "loss": 2.1623, + "step": 733 + }, + { + "epoch": 0.8043835616438356, + "grad_norm": 1.6864640712738037, + "learning_rate": 4.0982940698619015e-06, + "loss": 2.1748, + "step": 734 + }, + { + "epoch": 0.8054794520547945, + "grad_norm": 1.564070463180542, + "learning_rate": 4.096263200649879e-06, + "loss": 2.1458, + "step": 735 + }, + { + "epoch": 0.8065753424657535, + "grad_norm": 1.9004014730453491, + "learning_rate": 4.094232331437856e-06, + "loss": 2.1536, + "step": 736 + }, + { + "epoch": 0.8076712328767123, + "grad_norm": 1.6670420169830322, + "learning_rate": 4.092201462225833e-06, + "loss": 2.1809, + "step": 737 + }, + { + "epoch": 0.8087671232876712, + "grad_norm": 1.6553101539611816, + "learning_rate": 4.090170593013811e-06, + "loss": 2.1343, + "step": 738 + }, + { + "epoch": 0.8098630136986301, + "grad_norm": 1.9083986282348633, + "learning_rate": 4.088139723801787e-06, + "loss": 2.1455, + "step": 739 + }, + { + "epoch": 0.810958904109589, + "grad_norm": 1.6069375276565552, + "learning_rate": 4.086108854589765e-06, + "loss": 2.1621, + "step": 740 + }, + { + "epoch": 0.810958904109589, + "eval_loss": 2.3045499324798584, + "eval_runtime": 37.7804, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 740 + }, + { + "epoch": 0.812054794520548, + "grad_norm": 1.6774693727493286, + "learning_rate": 4.084077985377742e-06, + "loss": 2.2079, + "step": 741 + }, + { + "epoch": 0.8131506849315069, + "grad_norm": 1.772355556488037, + "learning_rate": 4.08204711616572e-06, + "loss": 2.1276, + "step": 742 + }, + { + "epoch": 0.8142465753424658, + "grad_norm": 1.6998035907745361, + "learning_rate": 4.080016246953696e-06, + "loss": 2.1903, + "step": 743 + }, + { + "epoch": 0.8153424657534246, + "grad_norm": 1.850844383239746, + "learning_rate": 4.077985377741674e-06, + "loss": 2.1965, + "step": 744 + }, + { + "epoch": 0.8164383561643835, + "grad_norm": 1.5916879177093506, + "learning_rate": 4.075954508529651e-06, + "loss": 2.1982, + "step": 745 + }, + { + "epoch": 0.8175342465753425, + "grad_norm": 1.8195812702178955, + "learning_rate": 4.073923639317629e-06, + "loss": 2.1507, + "step": 746 + }, + { + "epoch": 0.8186301369863014, + "grad_norm": 1.5145621299743652, + "learning_rate": 4.071892770105605e-06, + "loss": 2.1236, + "step": 747 + }, + { + "epoch": 0.8197260273972603, + "grad_norm": 1.5750877857208252, + "learning_rate": 4.069861900893583e-06, + "loss": 2.142, + "step": 748 + }, + { + "epoch": 0.8208219178082192, + "grad_norm": 1.708666443824768, + "learning_rate": 4.06783103168156e-06, + "loss": 2.1577, + "step": 749 + }, + { + "epoch": 0.821917808219178, + "grad_norm": 1.5988762378692627, + "learning_rate": 4.065800162469538e-06, + "loss": 2.1795, + "step": 750 + }, + { + "epoch": 0.821917808219178, + "eval_loss": 2.3052196502685547, + "eval_runtime": 37.7983, + "eval_samples_per_second": 26.456, + "eval_steps_per_second": 1.111, + "step": 750 + }, + { + "epoch": 0.823013698630137, + "grad_norm": 1.608738899230957, + "learning_rate": 4.063769293257514e-06, + "loss": 2.1578, + "step": 751 + }, + { + "epoch": 0.8241095890410959, + "grad_norm": 1.5578334331512451, + "learning_rate": 4.061738424045492e-06, + "loss": 2.1724, + "step": 752 + }, + { + "epoch": 0.8252054794520548, + "grad_norm": 1.6957333087921143, + "learning_rate": 4.059707554833469e-06, + "loss": 2.1975, + "step": 753 + }, + { + "epoch": 0.8263013698630137, + "grad_norm": 1.691512107849121, + "learning_rate": 4.057676685621446e-06, + "loss": 2.1705, + "step": 754 + }, + { + "epoch": 0.8273972602739726, + "grad_norm": 1.6061747074127197, + "learning_rate": 4.0556458164094235e-06, + "loss": 2.1031, + "step": 755 + }, + { + "epoch": 0.8284931506849315, + "grad_norm": 1.6074447631835938, + "learning_rate": 4.053614947197401e-06, + "loss": 2.1738, + "step": 756 + }, + { + "epoch": 0.8295890410958904, + "grad_norm": 1.9722917079925537, + "learning_rate": 4.051584077985378e-06, + "loss": 2.2043, + "step": 757 + }, + { + "epoch": 0.8306849315068493, + "grad_norm": 1.597692847251892, + "learning_rate": 4.049553208773355e-06, + "loss": 2.1621, + "step": 758 + }, + { + "epoch": 0.8317808219178082, + "grad_norm": 1.85458242893219, + "learning_rate": 4.047522339561333e-06, + "loss": 2.1504, + "step": 759 + }, + { + "epoch": 0.8328767123287671, + "grad_norm": 1.8328344821929932, + "learning_rate": 4.04549147034931e-06, + "loss": 2.2541, + "step": 760 + }, + { + "epoch": 0.8328767123287671, + "eval_loss": 2.3033809661865234, + "eval_runtime": 37.8088, + "eval_samples_per_second": 26.449, + "eval_steps_per_second": 1.111, + "step": 760 + }, + { + "epoch": 0.8339726027397261, + "grad_norm": 1.6544418334960938, + "learning_rate": 4.043460601137287e-06, + "loss": 2.1605, + "step": 761 + }, + { + "epoch": 0.8350684931506849, + "grad_norm": 1.9443641901016235, + "learning_rate": 4.041429731925264e-06, + "loss": 2.1621, + "step": 762 + }, + { + "epoch": 0.8361643835616438, + "grad_norm": 1.731277346611023, + "learning_rate": 4.039398862713242e-06, + "loss": 2.2168, + "step": 763 + }, + { + "epoch": 0.8372602739726027, + "grad_norm": 1.671072244644165, + "learning_rate": 4.037367993501219e-06, + "loss": 2.1612, + "step": 764 + }, + { + "epoch": 0.8383561643835616, + "grad_norm": 1.8777449131011963, + "learning_rate": 4.035337124289196e-06, + "loss": 2.1339, + "step": 765 + }, + { + "epoch": 0.8394520547945206, + "grad_norm": 1.5941543579101562, + "learning_rate": 4.0333062550771735e-06, + "loss": 2.1545, + "step": 766 + }, + { + "epoch": 0.8405479452054795, + "grad_norm": 1.9245301485061646, + "learning_rate": 4.031275385865151e-06, + "loss": 2.187, + "step": 767 + }, + { + "epoch": 0.8416438356164384, + "grad_norm": 1.69217848777771, + "learning_rate": 4.029244516653128e-06, + "loss": 2.1168, + "step": 768 + }, + { + "epoch": 0.8427397260273972, + "grad_norm": 1.7042242288589478, + "learning_rate": 4.027213647441105e-06, + "loss": 2.1353, + "step": 769 + }, + { + "epoch": 0.8438356164383561, + "grad_norm": 1.6391927003860474, + "learning_rate": 4.025182778229083e-06, + "loss": 2.1568, + "step": 770 + }, + { + "epoch": 0.8438356164383561, + "eval_loss": 2.310056686401367, + "eval_runtime": 37.7573, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 770 + }, + { + "epoch": 0.8449315068493151, + "grad_norm": 1.7833209037780762, + "learning_rate": 4.02315190901706e-06, + "loss": 2.1664, + "step": 771 + }, + { + "epoch": 0.846027397260274, + "grad_norm": 1.8203891515731812, + "learning_rate": 4.021121039805037e-06, + "loss": 2.2342, + "step": 772 + }, + { + "epoch": 0.8471232876712329, + "grad_norm": 1.572940707206726, + "learning_rate": 4.0190901705930144e-06, + "loss": 2.1418, + "step": 773 + }, + { + "epoch": 0.8482191780821918, + "grad_norm": 1.6871955394744873, + "learning_rate": 4.017059301380992e-06, + "loss": 2.2023, + "step": 774 + }, + { + "epoch": 0.8493150684931506, + "grad_norm": 1.6343133449554443, + "learning_rate": 4.015028432168969e-06, + "loss": 2.1302, + "step": 775 + }, + { + "epoch": 0.8504109589041096, + "grad_norm": 1.4492065906524658, + "learning_rate": 4.012997562956945e-06, + "loss": 2.1427, + "step": 776 + }, + { + "epoch": 0.8515068493150685, + "grad_norm": 1.6599146127700806, + "learning_rate": 4.0109666937449235e-06, + "loss": 2.1479, + "step": 777 + }, + { + "epoch": 0.8526027397260274, + "grad_norm": 1.6836353540420532, + "learning_rate": 4.0089358245329e-06, + "loss": 2.148, + "step": 778 + }, + { + "epoch": 0.8536986301369863, + "grad_norm": 1.5977990627288818, + "learning_rate": 4.006904955320878e-06, + "loss": 2.1946, + "step": 779 + }, + { + "epoch": 0.8547945205479452, + "grad_norm": 1.6259957551956177, + "learning_rate": 4.0048740861088545e-06, + "loss": 2.1761, + "step": 780 + }, + { + "epoch": 0.8547945205479452, + "eval_loss": 2.303910970687866, + "eval_runtime": 37.7728, + "eval_samples_per_second": 26.474, + "eval_steps_per_second": 1.112, + "step": 780 + }, + { + "epoch": 0.8558904109589041, + "grad_norm": 1.6990300416946411, + "learning_rate": 4.002843216896833e-06, + "loss": 2.1309, + "step": 781 + }, + { + "epoch": 0.856986301369863, + "grad_norm": 1.6556029319763184, + "learning_rate": 4.000812347684809e-06, + "loss": 2.151, + "step": 782 + }, + { + "epoch": 0.8580821917808219, + "grad_norm": 1.5659290552139282, + "learning_rate": 3.998781478472787e-06, + "loss": 2.1317, + "step": 783 + }, + { + "epoch": 0.8591780821917808, + "grad_norm": 1.6553704738616943, + "learning_rate": 3.996750609260764e-06, + "loss": 2.2344, + "step": 784 + }, + { + "epoch": 0.8602739726027397, + "grad_norm": 1.6608418226242065, + "learning_rate": 3.994719740048742e-06, + "loss": 2.1807, + "step": 785 + }, + { + "epoch": 0.8613698630136987, + "grad_norm": 1.5862030982971191, + "learning_rate": 3.992688870836718e-06, + "loss": 2.1351, + "step": 786 + }, + { + "epoch": 0.8624657534246576, + "grad_norm": 1.767526388168335, + "learning_rate": 3.990658001624696e-06, + "loss": 2.1856, + "step": 787 + }, + { + "epoch": 0.8635616438356164, + "grad_norm": 1.6902577877044678, + "learning_rate": 3.988627132412673e-06, + "loss": 2.1592, + "step": 788 + }, + { + "epoch": 0.8646575342465753, + "grad_norm": 1.7760623693466187, + "learning_rate": 3.98659626320065e-06, + "loss": 2.1645, + "step": 789 + }, + { + "epoch": 0.8657534246575342, + "grad_norm": 1.6791038513183594, + "learning_rate": 3.984565393988627e-06, + "loss": 2.1611, + "step": 790 + }, + { + "epoch": 0.8657534246575342, + "eval_loss": 2.3027942180633545, + "eval_runtime": 37.736, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 790 + }, + { + "epoch": 0.8668493150684932, + "grad_norm": 1.6695773601531982, + "learning_rate": 3.9825345247766046e-06, + "loss": 2.1747, + "step": 791 + }, + { + "epoch": 0.8679452054794521, + "grad_norm": 1.8736445903778076, + "learning_rate": 3.980503655564582e-06, + "loss": 2.1806, + "step": 792 + }, + { + "epoch": 0.869041095890411, + "grad_norm": 1.6695518493652344, + "learning_rate": 3.978472786352559e-06, + "loss": 2.1631, + "step": 793 + }, + { + "epoch": 0.8701369863013698, + "grad_norm": 1.6782366037368774, + "learning_rate": 3.976441917140536e-06, + "loss": 2.2175, + "step": 794 + }, + { + "epoch": 0.8712328767123287, + "grad_norm": 1.678619623184204, + "learning_rate": 3.974411047928514e-06, + "loss": 2.1883, + "step": 795 + }, + { + "epoch": 0.8723287671232877, + "grad_norm": 1.6274909973144531, + "learning_rate": 3.972380178716491e-06, + "loss": 2.1866, + "step": 796 + }, + { + "epoch": 0.8734246575342466, + "grad_norm": 1.5609477758407593, + "learning_rate": 3.970349309504468e-06, + "loss": 2.1257, + "step": 797 + }, + { + "epoch": 0.8745205479452055, + "grad_norm": 1.7419973611831665, + "learning_rate": 3.9683184402924455e-06, + "loss": 2.1253, + "step": 798 + }, + { + "epoch": 0.8756164383561644, + "grad_norm": 1.6296330690383911, + "learning_rate": 3.966287571080423e-06, + "loss": 2.1755, + "step": 799 + }, + { + "epoch": 0.8767123287671232, + "grad_norm": 1.5995864868164062, + "learning_rate": 3.9642567018684e-06, + "loss": 2.169, + "step": 800 + }, + { + "epoch": 0.8767123287671232, + "eval_loss": 2.301098108291626, + "eval_runtime": 37.7893, + "eval_samples_per_second": 26.463, + "eval_steps_per_second": 1.111, + "step": 800 + }, + { + "epoch": 0.8778082191780822, + "grad_norm": 1.653707504272461, + "learning_rate": 3.962225832656377e-06, + "loss": 2.115, + "step": 801 + }, + { + "epoch": 0.8789041095890411, + "grad_norm": 1.5908581018447876, + "learning_rate": 3.960194963444355e-06, + "loss": 2.1659, + "step": 802 + }, + { + "epoch": 0.88, + "grad_norm": 1.6865545511245728, + "learning_rate": 3.958164094232332e-06, + "loss": 2.1666, + "step": 803 + }, + { + "epoch": 0.8810958904109589, + "grad_norm": 1.6234707832336426, + "learning_rate": 3.956133225020309e-06, + "loss": 2.1594, + "step": 804 + }, + { + "epoch": 0.8821917808219178, + "grad_norm": 1.7375986576080322, + "learning_rate": 3.954102355808286e-06, + "loss": 2.1596, + "step": 805 + }, + { + "epoch": 0.8832876712328767, + "grad_norm": 1.6529247760772705, + "learning_rate": 3.952071486596264e-06, + "loss": 2.1495, + "step": 806 + }, + { + "epoch": 0.8843835616438356, + "grad_norm": 1.517357349395752, + "learning_rate": 3.950040617384241e-06, + "loss": 2.1394, + "step": 807 + }, + { + "epoch": 0.8854794520547945, + "grad_norm": 1.7367429733276367, + "learning_rate": 3.948009748172218e-06, + "loss": 2.206, + "step": 808 + }, + { + "epoch": 0.8865753424657534, + "grad_norm": 1.5467783212661743, + "learning_rate": 3.9459788789601955e-06, + "loss": 2.1672, + "step": 809 + }, + { + "epoch": 0.8876712328767123, + "grad_norm": 1.5647791624069214, + "learning_rate": 3.943948009748173e-06, + "loss": 2.1695, + "step": 810 + }, + { + "epoch": 0.8876712328767123, + "eval_loss": 2.2975995540618896, + "eval_runtime": 37.4912, + "eval_samples_per_second": 26.673, + "eval_steps_per_second": 1.12, + "step": 810 + }, + { + "epoch": 0.8887671232876713, + "grad_norm": 1.597876787185669, + "learning_rate": 3.94191714053615e-06, + "loss": 2.1808, + "step": 811 + }, + { + "epoch": 0.8898630136986302, + "grad_norm": 1.7645189762115479, + "learning_rate": 3.939886271324127e-06, + "loss": 2.1291, + "step": 812 + }, + { + "epoch": 0.890958904109589, + "grad_norm": 1.5209035873413086, + "learning_rate": 3.937855402112104e-06, + "loss": 2.1269, + "step": 813 + }, + { + "epoch": 0.8920547945205479, + "grad_norm": 1.5804905891418457, + "learning_rate": 3.935824532900082e-06, + "loss": 2.1751, + "step": 814 + }, + { + "epoch": 0.8931506849315068, + "grad_norm": 1.687865972518921, + "learning_rate": 3.933793663688058e-06, + "loss": 2.1917, + "step": 815 + }, + { + "epoch": 0.8942465753424658, + "grad_norm": 1.6767957210540771, + "learning_rate": 3.9317627944760365e-06, + "loss": 2.118, + "step": 816 + }, + { + "epoch": 0.8953424657534247, + "grad_norm": 1.657046914100647, + "learning_rate": 3.929731925264013e-06, + "loss": 2.1479, + "step": 817 + }, + { + "epoch": 0.8964383561643836, + "grad_norm": 1.5638110637664795, + "learning_rate": 3.927701056051991e-06, + "loss": 2.1565, + "step": 818 + }, + { + "epoch": 0.8975342465753424, + "grad_norm": 1.6865980625152588, + "learning_rate": 3.9256701868399674e-06, + "loss": 2.1382, + "step": 819 + }, + { + "epoch": 0.8986301369863013, + "grad_norm": 1.6738817691802979, + "learning_rate": 3.9236393176279456e-06, + "loss": 2.1257, + "step": 820 + }, + { + "epoch": 0.8986301369863013, + "eval_loss": 2.2953474521636963, + "eval_runtime": 37.7016, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 820 + }, + { + "epoch": 0.8997260273972603, + "grad_norm": 1.568678617477417, + "learning_rate": 3.921608448415922e-06, + "loss": 2.169, + "step": 821 + }, + { + "epoch": 0.9008219178082192, + "grad_norm": 1.6803818941116333, + "learning_rate": 3.9195775792039e-06, + "loss": 2.1489, + "step": 822 + }, + { + "epoch": 0.9019178082191781, + "grad_norm": 1.7010858058929443, + "learning_rate": 3.9175467099918765e-06, + "loss": 2.1846, + "step": 823 + }, + { + "epoch": 0.903013698630137, + "grad_norm": 1.5782372951507568, + "learning_rate": 3.915515840779855e-06, + "loss": 2.1788, + "step": 824 + }, + { + "epoch": 0.9041095890410958, + "grad_norm": 1.7804685831069946, + "learning_rate": 3.913484971567831e-06, + "loss": 2.1058, + "step": 825 + }, + { + "epoch": 0.9052054794520548, + "grad_norm": 1.5909292697906494, + "learning_rate": 3.911454102355808e-06, + "loss": 2.1716, + "step": 826 + }, + { + "epoch": 0.9063013698630137, + "grad_norm": 1.7001307010650635, + "learning_rate": 3.909423233143786e-06, + "loss": 2.1647, + "step": 827 + }, + { + "epoch": 0.9073972602739726, + "grad_norm": 1.5576646327972412, + "learning_rate": 3.907392363931763e-06, + "loss": 2.2174, + "step": 828 + }, + { + "epoch": 0.9084931506849315, + "grad_norm": 1.6621390581130981, + "learning_rate": 3.90536149471974e-06, + "loss": 2.2344, + "step": 829 + }, + { + "epoch": 0.9095890410958904, + "grad_norm": 1.592091679573059, + "learning_rate": 3.9033306255077175e-06, + "loss": 2.2501, + "step": 830 + }, + { + "epoch": 0.9095890410958904, + "eval_loss": 2.294229507446289, + "eval_runtime": 37.7195, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.113, + "step": 830 + }, + { + "epoch": 0.9106849315068493, + "grad_norm": 1.5698535442352295, + "learning_rate": 3.901299756295695e-06, + "loss": 2.1534, + "step": 831 + }, + { + "epoch": 0.9117808219178082, + "grad_norm": 1.7159922122955322, + "learning_rate": 3.899268887083672e-06, + "loss": 2.1601, + "step": 832 + }, + { + "epoch": 0.9128767123287671, + "grad_norm": 1.5601210594177246, + "learning_rate": 3.897238017871649e-06, + "loss": 2.1598, + "step": 833 + }, + { + "epoch": 0.913972602739726, + "grad_norm": 1.5987435579299927, + "learning_rate": 3.8952071486596266e-06, + "loss": 2.1307, + "step": 834 + }, + { + "epoch": 0.915068493150685, + "grad_norm": 1.7568409442901611, + "learning_rate": 3.893176279447604e-06, + "loss": 2.2209, + "step": 835 + }, + { + "epoch": 0.9161643835616439, + "grad_norm": 1.5044634342193604, + "learning_rate": 3.891145410235581e-06, + "loss": 2.1699, + "step": 836 + }, + { + "epoch": 0.9172602739726028, + "grad_norm": 1.7287421226501465, + "learning_rate": 3.889114541023558e-06, + "loss": 2.1489, + "step": 837 + }, + { + "epoch": 0.9183561643835616, + "grad_norm": 1.6595203876495361, + "learning_rate": 3.887083671811536e-06, + "loss": 2.1167, + "step": 838 + }, + { + "epoch": 0.9194520547945205, + "grad_norm": 1.5620248317718506, + "learning_rate": 3.885052802599513e-06, + "loss": 2.2119, + "step": 839 + }, + { + "epoch": 0.9205479452054794, + "grad_norm": 1.6492353677749634, + "learning_rate": 3.88302193338749e-06, + "loss": 2.1725, + "step": 840 + }, + { + "epoch": 0.9205479452054794, + "eval_loss": 2.2956349849700928, + "eval_runtime": 37.779, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 840 + }, + { + "epoch": 0.9216438356164384, + "grad_norm": 1.6601520776748657, + "learning_rate": 3.8809910641754675e-06, + "loss": 2.1532, + "step": 841 + }, + { + "epoch": 0.9227397260273973, + "grad_norm": 1.6723994016647339, + "learning_rate": 3.878960194963445e-06, + "loss": 2.1931, + "step": 842 + }, + { + "epoch": 0.9238356164383562, + "grad_norm": 1.7594473361968994, + "learning_rate": 3.876929325751422e-06, + "loss": 2.1219, + "step": 843 + }, + { + "epoch": 0.924931506849315, + "grad_norm": 1.546398401260376, + "learning_rate": 3.874898456539399e-06, + "loss": 2.1679, + "step": 844 + }, + { + "epoch": 0.9260273972602739, + "grad_norm": 1.8171741962432861, + "learning_rate": 3.872867587327377e-06, + "loss": 2.1434, + "step": 845 + }, + { + "epoch": 0.9271232876712329, + "grad_norm": 1.703196406364441, + "learning_rate": 3.870836718115354e-06, + "loss": 2.1574, + "step": 846 + }, + { + "epoch": 0.9282191780821918, + "grad_norm": 1.6290628910064697, + "learning_rate": 3.868805848903331e-06, + "loss": 2.1657, + "step": 847 + }, + { + "epoch": 0.9293150684931507, + "grad_norm": 1.5868806838989258, + "learning_rate": 3.8667749796913084e-06, + "loss": 2.1659, + "step": 848 + }, + { + "epoch": 0.9304109589041096, + "grad_norm": 1.551581621170044, + "learning_rate": 3.864744110479286e-06, + "loss": 2.1916, + "step": 849 + }, + { + "epoch": 0.9315068493150684, + "grad_norm": 1.6634963750839233, + "learning_rate": 3.862713241267262e-06, + "loss": 2.2027, + "step": 850 + }, + { + "epoch": 0.9315068493150684, + "eval_loss": 2.2988946437835693, + "eval_runtime": 37.7841, + "eval_samples_per_second": 26.466, + "eval_steps_per_second": 1.112, + "step": 850 + }, + { + "epoch": 0.9326027397260274, + "grad_norm": 1.6138510704040527, + "learning_rate": 3.86068237205524e-06, + "loss": 2.1634, + "step": 851 + }, + { + "epoch": 0.9336986301369863, + "grad_norm": 1.6294615268707275, + "learning_rate": 3.858651502843217e-06, + "loss": 2.1175, + "step": 852 + }, + { + "epoch": 0.9347945205479452, + "grad_norm": 1.7515549659729004, + "learning_rate": 3.856620633631195e-06, + "loss": 2.1562, + "step": 853 + }, + { + "epoch": 0.9358904109589041, + "grad_norm": 1.6304186582565308, + "learning_rate": 3.854589764419171e-06, + "loss": 2.2256, + "step": 854 + }, + { + "epoch": 0.936986301369863, + "grad_norm": 1.6189597845077515, + "learning_rate": 3.852558895207149e-06, + "loss": 2.1742, + "step": 855 + }, + { + "epoch": 0.938082191780822, + "grad_norm": 1.665515661239624, + "learning_rate": 3.850528025995126e-06, + "loss": 2.1481, + "step": 856 + }, + { + "epoch": 0.9391780821917808, + "grad_norm": 1.7066419124603271, + "learning_rate": 3.848497156783104e-06, + "loss": 2.1522, + "step": 857 + }, + { + "epoch": 0.9402739726027397, + "grad_norm": 1.884599208831787, + "learning_rate": 3.84646628757108e-06, + "loss": 2.1493, + "step": 858 + }, + { + "epoch": 0.9413698630136986, + "grad_norm": 1.6224873065948486, + "learning_rate": 3.8444354183590585e-06, + "loss": 2.1461, + "step": 859 + }, + { + "epoch": 0.9424657534246575, + "grad_norm": 1.916416049003601, + "learning_rate": 3.842404549147035e-06, + "loss": 2.1265, + "step": 860 + }, + { + "epoch": 0.9424657534246575, + "eval_loss": 2.297154426574707, + "eval_runtime": 37.77, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 860 + }, + { + "epoch": 0.9435616438356165, + "grad_norm": 1.5281857252120972, + "learning_rate": 3.840373679935013e-06, + "loss": 2.1476, + "step": 861 + }, + { + "epoch": 0.9446575342465754, + "grad_norm": 1.7905207872390747, + "learning_rate": 3.8383428107229894e-06, + "loss": 2.1421, + "step": 862 + }, + { + "epoch": 0.9457534246575342, + "grad_norm": 1.8282849788665771, + "learning_rate": 3.836311941510967e-06, + "loss": 2.1807, + "step": 863 + }, + { + "epoch": 0.9468493150684931, + "grad_norm": 1.7861522436141968, + "learning_rate": 3.834281072298944e-06, + "loss": 2.1337, + "step": 864 + }, + { + "epoch": 0.947945205479452, + "grad_norm": 1.6107250452041626, + "learning_rate": 3.832250203086921e-06, + "loss": 2.1911, + "step": 865 + }, + { + "epoch": 0.949041095890411, + "grad_norm": 1.657909631729126, + "learning_rate": 3.8302193338748985e-06, + "loss": 2.2257, + "step": 866 + }, + { + "epoch": 0.9501369863013699, + "grad_norm": 1.8969722986221313, + "learning_rate": 3.828188464662876e-06, + "loss": 2.1778, + "step": 867 + }, + { + "epoch": 0.9512328767123288, + "grad_norm": 1.6498360633850098, + "learning_rate": 3.826157595450853e-06, + "loss": 2.127, + "step": 868 + }, + { + "epoch": 0.9523287671232876, + "grad_norm": 1.5750585794448853, + "learning_rate": 3.82412672623883e-06, + "loss": 2.1716, + "step": 869 + }, + { + "epoch": 0.9534246575342465, + "grad_norm": 1.6047886610031128, + "learning_rate": 3.822095857026808e-06, + "loss": 2.1354, + "step": 870 + }, + { + "epoch": 0.9534246575342465, + "eval_loss": 2.295785665512085, + "eval_runtime": 37.7419, + "eval_samples_per_second": 26.496, + "eval_steps_per_second": 1.113, + "step": 870 + }, + { + "epoch": 0.9545205479452055, + "grad_norm": 1.6426187753677368, + "learning_rate": 3.820064987814785e-06, + "loss": 2.1375, + "step": 871 + }, + { + "epoch": 0.9556164383561644, + "grad_norm": 1.5118632316589355, + "learning_rate": 3.818034118602762e-06, + "loss": 2.1491, + "step": 872 + }, + { + "epoch": 0.9567123287671233, + "grad_norm": 1.630476474761963, + "learning_rate": 3.8160032493907395e-06, + "loss": 2.1497, + "step": 873 + }, + { + "epoch": 0.9578082191780822, + "grad_norm": 1.6820838451385498, + "learning_rate": 3.813972380178717e-06, + "loss": 2.1746, + "step": 874 + }, + { + "epoch": 0.958904109589041, + "grad_norm": 1.5480482578277588, + "learning_rate": 3.811941510966694e-06, + "loss": 2.1226, + "step": 875 + }, + { + "epoch": 0.96, + "grad_norm": 1.601493239402771, + "learning_rate": 3.809910641754671e-06, + "loss": 2.0934, + "step": 876 + }, + { + "epoch": 0.9610958904109589, + "grad_norm": 1.5756622552871704, + "learning_rate": 3.8078797725426486e-06, + "loss": 2.1679, + "step": 877 + }, + { + "epoch": 0.9621917808219178, + "grad_norm": 1.6811070442199707, + "learning_rate": 3.8058489033306254e-06, + "loss": 2.1407, + "step": 878 + }, + { + "epoch": 0.9632876712328767, + "grad_norm": 1.5748940706253052, + "learning_rate": 3.803818034118603e-06, + "loss": 2.1985, + "step": 879 + }, + { + "epoch": 0.9643835616438357, + "grad_norm": 1.5412342548370361, + "learning_rate": 3.80178716490658e-06, + "loss": 2.1876, + "step": 880 + }, + { + "epoch": 0.9643835616438357, + "eval_loss": 2.291630268096924, + "eval_runtime": 37.7702, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 880 + }, + { + "epoch": 0.9654794520547946, + "grad_norm": 1.7080684900283813, + "learning_rate": 3.7997562956945577e-06, + "loss": 2.1453, + "step": 881 + }, + { + "epoch": 0.9665753424657534, + "grad_norm": 1.8705110549926758, + "learning_rate": 3.7977254264825345e-06, + "loss": 2.1298, + "step": 882 + }, + { + "epoch": 0.9676712328767123, + "grad_norm": 1.7332537174224854, + "learning_rate": 3.7956945572705122e-06, + "loss": 2.2117, + "step": 883 + }, + { + "epoch": 0.9687671232876712, + "grad_norm": 1.88144052028656, + "learning_rate": 3.793663688058489e-06, + "loss": 2.2005, + "step": 884 + }, + { + "epoch": 0.9698630136986301, + "grad_norm": 1.6263943910598755, + "learning_rate": 3.791632818846467e-06, + "loss": 2.1853, + "step": 885 + }, + { + "epoch": 0.9709589041095891, + "grad_norm": 1.5968067646026611, + "learning_rate": 3.7896019496344436e-06, + "loss": 2.1998, + "step": 886 + }, + { + "epoch": 0.972054794520548, + "grad_norm": 1.7149827480316162, + "learning_rate": 3.787571080422421e-06, + "loss": 2.1855, + "step": 887 + }, + { + "epoch": 0.9731506849315068, + "grad_norm": 1.7375026941299438, + "learning_rate": 3.785540211210398e-06, + "loss": 2.1664, + "step": 888 + }, + { + "epoch": 0.9742465753424657, + "grad_norm": 1.6411423683166504, + "learning_rate": 3.7835093419983755e-06, + "loss": 2.1492, + "step": 889 + }, + { + "epoch": 0.9753424657534246, + "grad_norm": 1.6208266019821167, + "learning_rate": 3.7814784727863527e-06, + "loss": 2.2341, + "step": 890 + }, + { + "epoch": 0.9753424657534246, + "eval_loss": 2.2929985523223877, + "eval_runtime": 37.6693, + "eval_samples_per_second": 26.547, + "eval_steps_per_second": 1.115, + "step": 890 + }, + { + "epoch": 0.9764383561643836, + "grad_norm": 1.6871188879013062, + "learning_rate": 3.77944760357433e-06, + "loss": 2.1985, + "step": 891 + }, + { + "epoch": 0.9775342465753425, + "grad_norm": 1.8475587368011475, + "learning_rate": 3.7774167343623073e-06, + "loss": 2.1423, + "step": 892 + }, + { + "epoch": 0.9786301369863014, + "grad_norm": 1.6115065813064575, + "learning_rate": 3.7753858651502846e-06, + "loss": 2.1944, + "step": 893 + }, + { + "epoch": 0.9797260273972602, + "grad_norm": 1.6232231855392456, + "learning_rate": 3.773354995938262e-06, + "loss": 2.1751, + "step": 894 + }, + { + "epoch": 0.9808219178082191, + "grad_norm": 1.8564099073410034, + "learning_rate": 3.771324126726239e-06, + "loss": 2.2021, + "step": 895 + }, + { + "epoch": 0.9819178082191781, + "grad_norm": 1.7270197868347168, + "learning_rate": 3.7692932575142164e-06, + "loss": 2.1351, + "step": 896 + }, + { + "epoch": 0.983013698630137, + "grad_norm": 1.7187668085098267, + "learning_rate": 3.7672623883021937e-06, + "loss": 2.0856, + "step": 897 + }, + { + "epoch": 0.9841095890410959, + "grad_norm": 1.7766387462615967, + "learning_rate": 3.765231519090171e-06, + "loss": 2.168, + "step": 898 + }, + { + "epoch": 0.9852054794520548, + "grad_norm": 1.8063374757766724, + "learning_rate": 3.7632006498781482e-06, + "loss": 2.1317, + "step": 899 + }, + { + "epoch": 0.9863013698630136, + "grad_norm": 1.7601871490478516, + "learning_rate": 3.761169780666125e-06, + "loss": 2.1742, + "step": 900 + }, + { + "epoch": 0.9863013698630136, + "eval_loss": 2.292574167251587, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 900 + }, + { + "epoch": 0.9873972602739726, + "grad_norm": 1.8501280546188354, + "learning_rate": 3.7591389114541028e-06, + "loss": 2.1856, + "step": 901 + }, + { + "epoch": 0.9884931506849315, + "grad_norm": 1.8786728382110596, + "learning_rate": 3.7571080422420796e-06, + "loss": 2.1432, + "step": 902 + }, + { + "epoch": 0.9895890410958904, + "grad_norm": 1.7662389278411865, + "learning_rate": 3.7550771730300573e-06, + "loss": 2.2011, + "step": 903 + }, + { + "epoch": 0.9906849315068493, + "grad_norm": 1.6810210943222046, + "learning_rate": 3.753046303818034e-06, + "loss": 2.1719, + "step": 904 + }, + { + "epoch": 0.9917808219178083, + "grad_norm": 2.034982681274414, + "learning_rate": 3.751015434606012e-06, + "loss": 2.2206, + "step": 905 + }, + { + "epoch": 0.9928767123287672, + "grad_norm": 1.804049015045166, + "learning_rate": 3.7489845653939887e-06, + "loss": 2.1248, + "step": 906 + }, + { + "epoch": 0.993972602739726, + "grad_norm": 1.6039106845855713, + "learning_rate": 3.7469536961819664e-06, + "loss": 2.1545, + "step": 907 + }, + { + "epoch": 0.9950684931506849, + "grad_norm": 1.9352415800094604, + "learning_rate": 3.7449228269699433e-06, + "loss": 2.1842, + "step": 908 + }, + { + "epoch": 0.9961643835616438, + "grad_norm": 1.5759649276733398, + "learning_rate": 3.742891957757921e-06, + "loss": 2.1588, + "step": 909 + }, + { + "epoch": 0.9972602739726028, + "grad_norm": 1.6169486045837402, + "learning_rate": 3.740861088545898e-06, + "loss": 2.1871, + "step": 910 + }, + { + "epoch": 0.9972602739726028, + "eval_loss": 2.2923426628112793, + "eval_runtime": 37.4584, + "eval_samples_per_second": 26.696, + "eval_steps_per_second": 1.121, + "step": 910 + }, + { + "epoch": 0.9983561643835617, + "grad_norm": 1.748138427734375, + "learning_rate": 3.7388302193338755e-06, + "loss": 2.144, + "step": 911 + }, + { + "epoch": 0.9994520547945206, + "grad_norm": 1.8182991743087769, + "learning_rate": 3.7367993501218524e-06, + "loss": 2.1474, + "step": 912 + }, + { + "epoch": 1.0005479452054795, + "grad_norm": 1.6532196998596191, + "learning_rate": 3.7347684809098297e-06, + "loss": 2.1146, + "step": 913 + }, + { + "epoch": 1.0016438356164383, + "grad_norm": 2.0338973999023438, + "learning_rate": 3.732737611697807e-06, + "loss": 2.0851, + "step": 914 + }, + { + "epoch": 1.0027397260273974, + "grad_norm": 1.71133553981781, + "learning_rate": 3.7307067424857842e-06, + "loss": 2.1197, + "step": 915 + }, + { + "epoch": 1.0038356164383562, + "grad_norm": 1.6081736087799072, + "learning_rate": 3.7286758732737615e-06, + "loss": 2.1106, + "step": 916 + }, + { + "epoch": 1.004931506849315, + "grad_norm": 2.100734233856201, + "learning_rate": 3.7266450040617388e-06, + "loss": 2.2125, + "step": 917 + }, + { + "epoch": 1.006027397260274, + "grad_norm": 1.758898377418518, + "learning_rate": 3.724614134849716e-06, + "loss": 2.0985, + "step": 918 + }, + { + "epoch": 1.0071232876712328, + "grad_norm": 1.7335586547851562, + "learning_rate": 3.7225832656376933e-06, + "loss": 2.1911, + "step": 919 + }, + { + "epoch": 1.0082191780821919, + "grad_norm": 1.7879023551940918, + "learning_rate": 3.7205523964256706e-06, + "loss": 2.1632, + "step": 920 + }, + { + "epoch": 1.0082191780821919, + "eval_loss": 2.294053554534912, + "eval_runtime": 37.6132, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 920 + }, + { + "epoch": 1.0093150684931507, + "grad_norm": 1.7201987504959106, + "learning_rate": 3.718521527213648e-06, + "loss": 2.1501, + "step": 921 + }, + { + "epoch": 1.0104109589041095, + "grad_norm": 1.7638986110687256, + "learning_rate": 3.716490658001625e-06, + "loss": 2.0661, + "step": 922 + }, + { + "epoch": 1.0115068493150685, + "grad_norm": 1.7376288175582886, + "learning_rate": 3.7144597887896024e-06, + "loss": 2.1613, + "step": 923 + }, + { + "epoch": 1.0126027397260273, + "grad_norm": 1.9208155870437622, + "learning_rate": 3.7124289195775797e-06, + "loss": 2.1562, + "step": 924 + }, + { + "epoch": 1.0136986301369864, + "grad_norm": 1.6709166765213013, + "learning_rate": 3.710398050365557e-06, + "loss": 2.1309, + "step": 925 + }, + { + "epoch": 1.0147945205479452, + "grad_norm": 1.7070904970169067, + "learning_rate": 3.708367181153534e-06, + "loss": 2.1538, + "step": 926 + }, + { + "epoch": 1.0158904109589042, + "grad_norm": 1.7832821607589722, + "learning_rate": 3.7063363119415115e-06, + "loss": 2.1262, + "step": 927 + }, + { + "epoch": 1.016986301369863, + "grad_norm": 1.8703486919403076, + "learning_rate": 3.7043054427294884e-06, + "loss": 2.1142, + "step": 928 + }, + { + "epoch": 1.0180821917808218, + "grad_norm": 1.692858338356018, + "learning_rate": 3.702274573517466e-06, + "loss": 2.1182, + "step": 929 + }, + { + "epoch": 1.0191780821917809, + "grad_norm": 1.7156164646148682, + "learning_rate": 3.700243704305443e-06, + "loss": 2.1189, + "step": 930 + }, + { + "epoch": 1.0191780821917809, + "eval_loss": 2.2969048023223877, + "eval_runtime": 37.6673, + "eval_samples_per_second": 26.548, + "eval_steps_per_second": 1.115, + "step": 930 + }, + { + "epoch": 1.0202739726027397, + "grad_norm": 1.7443044185638428, + "learning_rate": 3.6982128350934206e-06, + "loss": 2.1058, + "step": 931 + }, + { + "epoch": 1.0213698630136987, + "grad_norm": 2.00412917137146, + "learning_rate": 3.6961819658813975e-06, + "loss": 2.2071, + "step": 932 + }, + { + "epoch": 1.0224657534246575, + "grad_norm": 1.7170531749725342, + "learning_rate": 3.694151096669375e-06, + "loss": 2.1254, + "step": 933 + }, + { + "epoch": 1.0235616438356165, + "grad_norm": 1.679018497467041, + "learning_rate": 3.692120227457352e-06, + "loss": 2.1487, + "step": 934 + }, + { + "epoch": 1.0246575342465754, + "grad_norm": 1.8477305173873901, + "learning_rate": 3.6900893582453297e-06, + "loss": 2.062, + "step": 935 + }, + { + "epoch": 1.0257534246575342, + "grad_norm": 1.664095401763916, + "learning_rate": 3.6880584890333066e-06, + "loss": 2.1418, + "step": 936 + }, + { + "epoch": 1.0268493150684932, + "grad_norm": 1.8831602334976196, + "learning_rate": 3.6860276198212834e-06, + "loss": 2.1483, + "step": 937 + }, + { + "epoch": 1.027945205479452, + "grad_norm": 1.7625365257263184, + "learning_rate": 3.683996750609261e-06, + "loss": 2.146, + "step": 938 + }, + { + "epoch": 1.029041095890411, + "grad_norm": 1.8168137073516846, + "learning_rate": 3.681965881397238e-06, + "loss": 2.1594, + "step": 939 + }, + { + "epoch": 1.0301369863013699, + "grad_norm": 1.6803343296051025, + "learning_rate": 3.6799350121852157e-06, + "loss": 2.1168, + "step": 940 + }, + { + "epoch": 1.0301369863013699, + "eval_loss": 2.2941982746124268, + "eval_runtime": 37.7112, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 940 + }, + { + "epoch": 1.0312328767123287, + "grad_norm": 2.0829122066497803, + "learning_rate": 3.6779041429731925e-06, + "loss": 2.1399, + "step": 941 + }, + { + "epoch": 1.0323287671232877, + "grad_norm": 1.6813585758209229, + "learning_rate": 3.6758732737611702e-06, + "loss": 2.1312, + "step": 942 + }, + { + "epoch": 1.0334246575342465, + "grad_norm": 1.7511613368988037, + "learning_rate": 3.673842404549147e-06, + "loss": 2.1273, + "step": 943 + }, + { + "epoch": 1.0345205479452055, + "grad_norm": 1.8264555931091309, + "learning_rate": 3.671811535337125e-06, + "loss": 2.1343, + "step": 944 + }, + { + "epoch": 1.0356164383561643, + "grad_norm": 1.7675082683563232, + "learning_rate": 3.6697806661251016e-06, + "loss": 2.2153, + "step": 945 + }, + { + "epoch": 1.0367123287671234, + "grad_norm": 1.6589823961257935, + "learning_rate": 3.6677497969130793e-06, + "loss": 2.1575, + "step": 946 + }, + { + "epoch": 1.0378082191780822, + "grad_norm": 1.6730010509490967, + "learning_rate": 3.665718927701056e-06, + "loss": 2.1773, + "step": 947 + }, + { + "epoch": 1.038904109589041, + "grad_norm": 1.627983570098877, + "learning_rate": 3.663688058489034e-06, + "loss": 2.1394, + "step": 948 + }, + { + "epoch": 1.04, + "grad_norm": 1.6363394260406494, + "learning_rate": 3.6616571892770107e-06, + "loss": 2.0828, + "step": 949 + }, + { + "epoch": 1.0410958904109588, + "grad_norm": 1.6270133256912231, + "learning_rate": 3.659626320064988e-06, + "loss": 2.1308, + "step": 950 + }, + { + "epoch": 1.0410958904109588, + "eval_loss": 2.293911933898926, + "eval_runtime": 37.7095, + "eval_samples_per_second": 26.519, + "eval_steps_per_second": 1.114, + "step": 950 + }, + { + "epoch": 1.0421917808219179, + "grad_norm": 1.784549593925476, + "learning_rate": 3.6575954508529653e-06, + "loss": 2.1239, + "step": 951 + }, + { + "epoch": 1.0432876712328767, + "grad_norm": 1.7265311479568481, + "learning_rate": 3.6555645816409426e-06, + "loss": 2.1573, + "step": 952 + }, + { + "epoch": 1.0443835616438357, + "grad_norm": 1.8133327960968018, + "learning_rate": 3.65353371242892e-06, + "loss": 2.046, + "step": 953 + }, + { + "epoch": 1.0454794520547945, + "grad_norm": 1.7085450887680054, + "learning_rate": 3.651502843216897e-06, + "loss": 2.1471, + "step": 954 + }, + { + "epoch": 1.0465753424657533, + "grad_norm": 2.0099172592163086, + "learning_rate": 3.6494719740048744e-06, + "loss": 2.1193, + "step": 955 + }, + { + "epoch": 1.0476712328767124, + "grad_norm": 1.582085132598877, + "learning_rate": 3.6474411047928517e-06, + "loss": 2.1522, + "step": 956 + }, + { + "epoch": 1.0487671232876712, + "grad_norm": 1.7923028469085693, + "learning_rate": 3.645410235580829e-06, + "loss": 2.1358, + "step": 957 + }, + { + "epoch": 1.0498630136986302, + "grad_norm": 1.6636080741882324, + "learning_rate": 3.6433793663688062e-06, + "loss": 2.1263, + "step": 958 + }, + { + "epoch": 1.050958904109589, + "grad_norm": 1.681675672531128, + "learning_rate": 3.6413484971567835e-06, + "loss": 2.1369, + "step": 959 + }, + { + "epoch": 1.0520547945205478, + "grad_norm": 1.7007646560668945, + "learning_rate": 3.6393176279447608e-06, + "loss": 2.1606, + "step": 960 + }, + { + "epoch": 1.0520547945205478, + "eval_loss": 2.291552782058716, + "eval_runtime": 37.704, + "eval_samples_per_second": 26.522, + "eval_steps_per_second": 1.114, + "step": 960 + }, + { + "epoch": 1.0531506849315069, + "grad_norm": 1.632997989654541, + "learning_rate": 3.637286758732738e-06, + "loss": 2.1742, + "step": 961 + }, + { + "epoch": 1.0542465753424657, + "grad_norm": 1.6881494522094727, + "learning_rate": 3.6352558895207153e-06, + "loss": 2.1555, + "step": 962 + }, + { + "epoch": 1.0553424657534247, + "grad_norm": 1.6624116897583008, + "learning_rate": 3.633225020308692e-06, + "loss": 2.1278, + "step": 963 + }, + { + "epoch": 1.0564383561643835, + "grad_norm": 1.7851325273513794, + "learning_rate": 3.63119415109667e-06, + "loss": 2.1599, + "step": 964 + }, + { + "epoch": 1.0575342465753426, + "grad_norm": 1.7592188119888306, + "learning_rate": 3.6291632818846467e-06, + "loss": 2.0898, + "step": 965 + }, + { + "epoch": 1.0586301369863014, + "grad_norm": 1.9573521614074707, + "learning_rate": 3.6271324126726244e-06, + "loss": 2.0876, + "step": 966 + }, + { + "epoch": 1.0597260273972602, + "grad_norm": 1.6684988737106323, + "learning_rate": 3.6251015434606013e-06, + "loss": 2.1288, + "step": 967 + }, + { + "epoch": 1.0608219178082192, + "grad_norm": 2.0011937618255615, + "learning_rate": 3.623070674248579e-06, + "loss": 2.1166, + "step": 968 + }, + { + "epoch": 1.061917808219178, + "grad_norm": 1.7490673065185547, + "learning_rate": 3.621039805036556e-06, + "loss": 2.1532, + "step": 969 + }, + { + "epoch": 1.063013698630137, + "grad_norm": 1.756422758102417, + "learning_rate": 3.6190089358245335e-06, + "loss": 2.1486, + "step": 970 + }, + { + "epoch": 1.063013698630137, + "eval_loss": 2.29402494430542, + "eval_runtime": 37.7078, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 970 + }, + { + "epoch": 1.0641095890410959, + "grad_norm": 2.014153480529785, + "learning_rate": 3.6169780666125104e-06, + "loss": 2.1512, + "step": 971 + }, + { + "epoch": 1.0652054794520547, + "grad_norm": 1.6026027202606201, + "learning_rate": 3.614947197400488e-06, + "loss": 2.0787, + "step": 972 + }, + { + "epoch": 1.0663013698630137, + "grad_norm": 1.8149386644363403, + "learning_rate": 3.612916328188465e-06, + "loss": 2.1675, + "step": 973 + }, + { + "epoch": 1.0673972602739725, + "grad_norm": 1.697019338607788, + "learning_rate": 3.610885458976442e-06, + "loss": 2.1584, + "step": 974 + }, + { + "epoch": 1.0684931506849316, + "grad_norm": 1.7330286502838135, + "learning_rate": 3.6088545897644195e-06, + "loss": 2.1048, + "step": 975 + }, + { + "epoch": 1.0695890410958904, + "grad_norm": 1.650336742401123, + "learning_rate": 3.6068237205523963e-06, + "loss": 2.1078, + "step": 976 + }, + { + "epoch": 1.0706849315068494, + "grad_norm": 1.7817063331604004, + "learning_rate": 3.604792851340374e-06, + "loss": 2.1656, + "step": 977 + }, + { + "epoch": 1.0717808219178082, + "grad_norm": 1.818617820739746, + "learning_rate": 3.602761982128351e-06, + "loss": 2.1388, + "step": 978 + }, + { + "epoch": 1.072876712328767, + "grad_norm": 1.7448735237121582, + "learning_rate": 3.6007311129163286e-06, + "loss": 2.1022, + "step": 979 + }, + { + "epoch": 1.073972602739726, + "grad_norm": 1.8401628732681274, + "learning_rate": 3.5987002437043054e-06, + "loss": 2.1181, + "step": 980 + }, + { + "epoch": 1.073972602739726, + "eval_loss": 2.2948012351989746, + "eval_runtime": 37.6902, + "eval_samples_per_second": 26.532, + "eval_steps_per_second": 1.114, + "step": 980 + }, + { + "epoch": 1.0750684931506849, + "grad_norm": 1.979794979095459, + "learning_rate": 3.596669374492283e-06, + "loss": 2.1766, + "step": 981 + }, + { + "epoch": 1.076164383561644, + "grad_norm": 1.7657036781311035, + "learning_rate": 3.59463850528026e-06, + "loss": 2.0889, + "step": 982 + }, + { + "epoch": 1.0772602739726027, + "grad_norm": 1.8293733596801758, + "learning_rate": 3.5926076360682377e-06, + "loss": 2.159, + "step": 983 + }, + { + "epoch": 1.0783561643835617, + "grad_norm": 1.7979159355163574, + "learning_rate": 3.5905767668562145e-06, + "loss": 2.1391, + "step": 984 + }, + { + "epoch": 1.0794520547945206, + "grad_norm": 1.847598910331726, + "learning_rate": 3.5885458976441922e-06, + "loss": 2.1695, + "step": 985 + }, + { + "epoch": 1.0805479452054794, + "grad_norm": 1.608178734779358, + "learning_rate": 3.586515028432169e-06, + "loss": 2.1745, + "step": 986 + }, + { + "epoch": 1.0816438356164384, + "grad_norm": 1.8166873455047607, + "learning_rate": 3.5844841592201464e-06, + "loss": 2.0882, + "step": 987 + }, + { + "epoch": 1.0827397260273972, + "grad_norm": 1.847463607788086, + "learning_rate": 3.5824532900081237e-06, + "loss": 2.1244, + "step": 988 + }, + { + "epoch": 1.0838356164383562, + "grad_norm": 1.6194696426391602, + "learning_rate": 3.580422420796101e-06, + "loss": 2.1398, + "step": 989 + }, + { + "epoch": 1.084931506849315, + "grad_norm": 1.9055372476577759, + "learning_rate": 3.578391551584078e-06, + "loss": 2.0774, + "step": 990 + }, + { + "epoch": 1.084931506849315, + "eval_loss": 2.2899444103240967, + "eval_runtime": 37.7237, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 990 + }, + { + "epoch": 1.086027397260274, + "grad_norm": 1.6451077461242676, + "learning_rate": 3.5763606823720555e-06, + "loss": 2.1351, + "step": 991 + }, + { + "epoch": 1.087123287671233, + "grad_norm": 1.8144199848175049, + "learning_rate": 3.5743298131600328e-06, + "loss": 2.1536, + "step": 992 + }, + { + "epoch": 1.0882191780821917, + "grad_norm": 1.760130763053894, + "learning_rate": 3.57229894394801e-06, + "loss": 2.1485, + "step": 993 + }, + { + "epoch": 1.0893150684931507, + "grad_norm": 1.7544556856155396, + "learning_rate": 3.5702680747359873e-06, + "loss": 2.1136, + "step": 994 + }, + { + "epoch": 1.0904109589041096, + "grad_norm": 1.9884154796600342, + "learning_rate": 3.5682372055239646e-06, + "loss": 2.1366, + "step": 995 + }, + { + "epoch": 1.0915068493150686, + "grad_norm": 1.72066330909729, + "learning_rate": 3.566206336311942e-06, + "loss": 2.1103, + "step": 996 + }, + { + "epoch": 1.0926027397260274, + "grad_norm": 1.807808756828308, + "learning_rate": 3.564175467099919e-06, + "loss": 2.0905, + "step": 997 + }, + { + "epoch": 1.0936986301369862, + "grad_norm": 2.1075172424316406, + "learning_rate": 3.5621445978878964e-06, + "loss": 2.1427, + "step": 998 + }, + { + "epoch": 1.0947945205479452, + "grad_norm": 1.7037009000778198, + "learning_rate": 3.5601137286758737e-06, + "loss": 2.1124, + "step": 999 + }, + { + "epoch": 1.095890410958904, + "grad_norm": 1.842721700668335, + "learning_rate": 3.5580828594638505e-06, + "loss": 2.1379, + "step": 1000 + }, + { + "epoch": 1.095890410958904, + "eval_loss": 2.291067361831665, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 1000 + }, + { + "epoch": 1.096986301369863, + "grad_norm": 1.802979588508606, + "learning_rate": 3.5560519902518282e-06, + "loss": 2.156, + "step": 1001 + }, + { + "epoch": 1.098082191780822, + "grad_norm": 1.699763298034668, + "learning_rate": 3.554021121039805e-06, + "loss": 2.1353, + "step": 1002 + }, + { + "epoch": 1.099178082191781, + "grad_norm": 1.695544958114624, + "learning_rate": 3.551990251827783e-06, + "loss": 2.1039, + "step": 1003 + }, + { + "epoch": 1.1002739726027397, + "grad_norm": 1.9509766101837158, + "learning_rate": 3.5499593826157596e-06, + "loss": 2.1201, + "step": 1004 + }, + { + "epoch": 1.1013698630136985, + "grad_norm": 1.7657231092453003, + "learning_rate": 3.5479285134037373e-06, + "loss": 2.1149, + "step": 1005 + }, + { + "epoch": 1.1024657534246576, + "grad_norm": 1.7416653633117676, + "learning_rate": 3.545897644191714e-06, + "loss": 2.1116, + "step": 1006 + }, + { + "epoch": 1.1035616438356164, + "grad_norm": 1.804185390472412, + "learning_rate": 3.543866774979692e-06, + "loss": 2.1426, + "step": 1007 + }, + { + "epoch": 1.1046575342465754, + "grad_norm": 1.8011984825134277, + "learning_rate": 3.5418359057676687e-06, + "loss": 2.1132, + "step": 1008 + }, + { + "epoch": 1.1057534246575342, + "grad_norm": 1.9039791822433472, + "learning_rate": 3.5398050365556464e-06, + "loss": 2.1479, + "step": 1009 + }, + { + "epoch": 1.106849315068493, + "grad_norm": 1.9513524770736694, + "learning_rate": 3.5377741673436233e-06, + "loss": 2.1093, + "step": 1010 + }, + { + "epoch": 1.106849315068493, + "eval_loss": 2.290522575378418, + "eval_runtime": 37.4617, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 1010 + }, + { + "epoch": 1.107945205479452, + "grad_norm": 1.7262980937957764, + "learning_rate": 3.535743298131601e-06, + "loss": 2.138, + "step": 1011 + }, + { + "epoch": 1.1090410958904109, + "grad_norm": 1.9514356851577759, + "learning_rate": 3.533712428919578e-06, + "loss": 2.1669, + "step": 1012 + }, + { + "epoch": 1.11013698630137, + "grad_norm": 1.7947099208831787, + "learning_rate": 3.5316815597075547e-06, + "loss": 2.105, + "step": 1013 + }, + { + "epoch": 1.1112328767123287, + "grad_norm": 1.8680298328399658, + "learning_rate": 3.5296506904955324e-06, + "loss": 2.1306, + "step": 1014 + }, + { + "epoch": 1.1123287671232878, + "grad_norm": 2.121732711791992, + "learning_rate": 3.5276198212835093e-06, + "loss": 2.1681, + "step": 1015 + }, + { + "epoch": 1.1134246575342466, + "grad_norm": 1.662564754486084, + "learning_rate": 3.525588952071487e-06, + "loss": 2.1681, + "step": 1016 + }, + { + "epoch": 1.1145205479452054, + "grad_norm": 1.8358488082885742, + "learning_rate": 3.523558082859464e-06, + "loss": 2.1363, + "step": 1017 + }, + { + "epoch": 1.1156164383561644, + "grad_norm": 1.8742910623550415, + "learning_rate": 3.5215272136474415e-06, + "loss": 2.1179, + "step": 1018 + }, + { + "epoch": 1.1167123287671232, + "grad_norm": 1.8780633211135864, + "learning_rate": 3.5194963444354184e-06, + "loss": 2.103, + "step": 1019 + }, + { + "epoch": 1.1178082191780823, + "grad_norm": 1.6979540586471558, + "learning_rate": 3.517465475223396e-06, + "loss": 2.0885, + "step": 1020 + }, + { + "epoch": 1.1178082191780823, + "eval_loss": 2.2939751148223877, + "eval_runtime": 37.66, + "eval_samples_per_second": 26.553, + "eval_steps_per_second": 1.115, + "step": 1020 + }, + { + "epoch": 1.118904109589041, + "grad_norm": 2.494971752166748, + "learning_rate": 3.515434606011373e-06, + "loss": 2.1582, + "step": 1021 + }, + { + "epoch": 1.12, + "grad_norm": 1.9183858633041382, + "learning_rate": 3.5134037367993506e-06, + "loss": 2.189, + "step": 1022 + }, + { + "epoch": 1.121095890410959, + "grad_norm": 1.9055982828140259, + "learning_rate": 3.5113728675873275e-06, + "loss": 2.1503, + "step": 1023 + }, + { + "epoch": 1.1221917808219177, + "grad_norm": 2.118046522140503, + "learning_rate": 3.5093419983753047e-06, + "loss": 2.1414, + "step": 1024 + }, + { + "epoch": 1.1232876712328768, + "grad_norm": 1.8413280248641968, + "learning_rate": 3.507311129163282e-06, + "loss": 2.09, + "step": 1025 + }, + { + "epoch": 1.1243835616438356, + "grad_norm": 2.0261356830596924, + "learning_rate": 3.5052802599512593e-06, + "loss": 2.132, + "step": 1026 + }, + { + "epoch": 1.1254794520547946, + "grad_norm": 1.8393104076385498, + "learning_rate": 3.5032493907392366e-06, + "loss": 2.1366, + "step": 1027 + }, + { + "epoch": 1.1265753424657534, + "grad_norm": 2.0282111167907715, + "learning_rate": 3.501218521527214e-06, + "loss": 2.0758, + "step": 1028 + }, + { + "epoch": 1.1276712328767124, + "grad_norm": 1.8438650369644165, + "learning_rate": 3.499187652315191e-06, + "loss": 2.1294, + "step": 1029 + }, + { + "epoch": 1.1287671232876713, + "grad_norm": 1.890739917755127, + "learning_rate": 3.4971567831031684e-06, + "loss": 2.1591, + "step": 1030 + }, + { + "epoch": 1.1287671232876713, + "eval_loss": 2.2894623279571533, + "eval_runtime": 37.6941, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1030 + }, + { + "epoch": 1.12986301369863, + "grad_norm": 1.877964973449707, + "learning_rate": 3.4951259138911457e-06, + "loss": 2.0823, + "step": 1031 + }, + { + "epoch": 1.130958904109589, + "grad_norm": 1.7460334300994873, + "learning_rate": 3.493095044679123e-06, + "loss": 2.1758, + "step": 1032 + }, + { + "epoch": 1.132054794520548, + "grad_norm": 1.668459415435791, + "learning_rate": 3.4910641754671002e-06, + "loss": 2.0461, + "step": 1033 + }, + { + "epoch": 1.133150684931507, + "grad_norm": 1.9560495615005493, + "learning_rate": 3.4890333062550775e-06, + "loss": 2.1394, + "step": 1034 + }, + { + "epoch": 1.1342465753424658, + "grad_norm": 1.820432424545288, + "learning_rate": 3.4870024370430548e-06, + "loss": 2.0735, + "step": 1035 + }, + { + "epoch": 1.1353424657534246, + "grad_norm": 1.7064971923828125, + "learning_rate": 3.484971567831032e-06, + "loss": 2.1196, + "step": 1036 + }, + { + "epoch": 1.1364383561643836, + "grad_norm": 1.837594747543335, + "learning_rate": 3.482940698619009e-06, + "loss": 2.1562, + "step": 1037 + }, + { + "epoch": 1.1375342465753424, + "grad_norm": 1.7570106983184814, + "learning_rate": 3.4809098294069866e-06, + "loss": 2.1253, + "step": 1038 + }, + { + "epoch": 1.1386301369863014, + "grad_norm": 1.7931634187698364, + "learning_rate": 3.4788789601949634e-06, + "loss": 2.1369, + "step": 1039 + }, + { + "epoch": 1.1397260273972603, + "grad_norm": 1.957241177558899, + "learning_rate": 3.476848090982941e-06, + "loss": 2.1204, + "step": 1040 + }, + { + "epoch": 1.1397260273972603, + "eval_loss": 2.2888705730438232, + "eval_runtime": 37.7002, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 1040 + }, + { + "epoch": 1.140821917808219, + "grad_norm": 1.7801663875579834, + "learning_rate": 3.474817221770918e-06, + "loss": 2.2008, + "step": 1041 + }, + { + "epoch": 1.141917808219178, + "grad_norm": 1.654398798942566, + "learning_rate": 3.4727863525588957e-06, + "loss": 2.1451, + "step": 1042 + }, + { + "epoch": 1.143013698630137, + "grad_norm": 1.8888044357299805, + "learning_rate": 3.4707554833468725e-06, + "loss": 2.1318, + "step": 1043 + }, + { + "epoch": 1.144109589041096, + "grad_norm": 1.8398077487945557, + "learning_rate": 3.4687246141348502e-06, + "loss": 2.1051, + "step": 1044 + }, + { + "epoch": 1.1452054794520548, + "grad_norm": 1.6971104145050049, + "learning_rate": 3.466693744922827e-06, + "loss": 2.119, + "step": 1045 + }, + { + "epoch": 1.1463013698630138, + "grad_norm": 1.605700969696045, + "learning_rate": 3.464662875710805e-06, + "loss": 2.1343, + "step": 1046 + }, + { + "epoch": 1.1473972602739726, + "grad_norm": 1.8513860702514648, + "learning_rate": 3.4626320064987817e-06, + "loss": 2.1331, + "step": 1047 + }, + { + "epoch": 1.1484931506849314, + "grad_norm": 1.7160199880599976, + "learning_rate": 3.4606011372867594e-06, + "loss": 2.1367, + "step": 1048 + }, + { + "epoch": 1.1495890410958904, + "grad_norm": 1.5926575660705566, + "learning_rate": 3.458570268074736e-06, + "loss": 2.1183, + "step": 1049 + }, + { + "epoch": 1.1506849315068493, + "grad_norm": 1.663843035697937, + "learning_rate": 3.4565393988627135e-06, + "loss": 2.1107, + "step": 1050 + }, + { + "epoch": 1.1506849315068493, + "eval_loss": 2.286076545715332, + "eval_runtime": 37.6944, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1050 + }, + { + "epoch": 1.1517808219178083, + "grad_norm": 1.7697075605392456, + "learning_rate": 3.4545085296506908e-06, + "loss": 2.0912, + "step": 1051 + }, + { + "epoch": 1.152876712328767, + "grad_norm": 1.626094937324524, + "learning_rate": 3.452477660438668e-06, + "loss": 2.1508, + "step": 1052 + }, + { + "epoch": 1.1539726027397261, + "grad_norm": 1.9503792524337769, + "learning_rate": 3.4504467912266453e-06, + "loss": 2.0819, + "step": 1053 + }, + { + "epoch": 1.155068493150685, + "grad_norm": 1.7748280763626099, + "learning_rate": 3.4484159220146226e-06, + "loss": 2.1233, + "step": 1054 + }, + { + "epoch": 1.1561643835616437, + "grad_norm": 2.099339008331299, + "learning_rate": 3.4463850528026e-06, + "loss": 2.1245, + "step": 1055 + }, + { + "epoch": 1.1572602739726028, + "grad_norm": 1.6200031042099, + "learning_rate": 3.444354183590577e-06, + "loss": 2.1437, + "step": 1056 + }, + { + "epoch": 1.1583561643835616, + "grad_norm": 1.7703361511230469, + "learning_rate": 3.4423233143785544e-06, + "loss": 2.0874, + "step": 1057 + }, + { + "epoch": 1.1594520547945206, + "grad_norm": 1.6505805253982544, + "learning_rate": 3.4402924451665317e-06, + "loss": 2.0655, + "step": 1058 + }, + { + "epoch": 1.1605479452054794, + "grad_norm": 1.7199327945709229, + "learning_rate": 3.438261575954509e-06, + "loss": 2.1383, + "step": 1059 + }, + { + "epoch": 1.1616438356164385, + "grad_norm": 1.772292137145996, + "learning_rate": 3.4362307067424862e-06, + "loss": 2.1673, + "step": 1060 + }, + { + "epoch": 1.1616438356164385, + "eval_loss": 2.2869999408721924, + "eval_runtime": 37.7398, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1060 + }, + { + "epoch": 1.1627397260273973, + "grad_norm": 1.5849649906158447, + "learning_rate": 3.434199837530463e-06, + "loss": 2.1331, + "step": 1061 + }, + { + "epoch": 1.163835616438356, + "grad_norm": 1.8389405012130737, + "learning_rate": 3.432168968318441e-06, + "loss": 2.0832, + "step": 1062 + }, + { + "epoch": 1.1649315068493151, + "grad_norm": 1.7142471075057983, + "learning_rate": 3.4301380991064176e-06, + "loss": 2.1447, + "step": 1063 + }, + { + "epoch": 1.166027397260274, + "grad_norm": 1.7853225469589233, + "learning_rate": 3.4281072298943953e-06, + "loss": 2.1591, + "step": 1064 + }, + { + "epoch": 1.167123287671233, + "grad_norm": 1.8228695392608643, + "learning_rate": 3.426076360682372e-06, + "loss": 2.1684, + "step": 1065 + }, + { + "epoch": 1.1682191780821918, + "grad_norm": 1.7090480327606201, + "learning_rate": 3.42404549147035e-06, + "loss": 2.1339, + "step": 1066 + }, + { + "epoch": 1.1693150684931506, + "grad_norm": 1.642682433128357, + "learning_rate": 3.4220146222583267e-06, + "loss": 2.075, + "step": 1067 + }, + { + "epoch": 1.1704109589041096, + "grad_norm": 1.8230397701263428, + "learning_rate": 3.4199837530463044e-06, + "loss": 2.1128, + "step": 1068 + }, + { + "epoch": 1.1715068493150684, + "grad_norm": 1.761135458946228, + "learning_rate": 3.4179528838342813e-06, + "loss": 2.1885, + "step": 1069 + }, + { + "epoch": 1.1726027397260275, + "grad_norm": 1.7829216718673706, + "learning_rate": 3.415922014622259e-06, + "loss": 2.1584, + "step": 1070 + }, + { + "epoch": 1.1726027397260275, + "eval_loss": 2.2878382205963135, + "eval_runtime": 37.7433, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1070 + }, + { + "epoch": 1.1736986301369863, + "grad_norm": 1.5480386018753052, + "learning_rate": 3.413891145410236e-06, + "loss": 2.1602, + "step": 1071 + }, + { + "epoch": 1.174794520547945, + "grad_norm": 1.8957836627960205, + "learning_rate": 3.4118602761982135e-06, + "loss": 2.1706, + "step": 1072 + }, + { + "epoch": 1.1758904109589041, + "grad_norm": 1.924424648284912, + "learning_rate": 3.4098294069861904e-06, + "loss": 2.1339, + "step": 1073 + }, + { + "epoch": 1.176986301369863, + "grad_norm": 1.660370111465454, + "learning_rate": 3.4077985377741673e-06, + "loss": 2.1175, + "step": 1074 + }, + { + "epoch": 1.178082191780822, + "grad_norm": 1.7255185842514038, + "learning_rate": 3.405767668562145e-06, + "loss": 2.1672, + "step": 1075 + }, + { + "epoch": 1.1791780821917808, + "grad_norm": 1.8028507232666016, + "learning_rate": 3.403736799350122e-06, + "loss": 2.1527, + "step": 1076 + }, + { + "epoch": 1.1802739726027398, + "grad_norm": 1.7100828886032104, + "learning_rate": 3.4017059301380995e-06, + "loss": 2.1031, + "step": 1077 + }, + { + "epoch": 1.1813698630136986, + "grad_norm": 1.785955548286438, + "learning_rate": 3.3996750609260764e-06, + "loss": 2.1829, + "step": 1078 + }, + { + "epoch": 1.1824657534246574, + "grad_norm": 1.8583980798721313, + "learning_rate": 3.397644191714054e-06, + "loss": 2.1198, + "step": 1079 + }, + { + "epoch": 1.1835616438356165, + "grad_norm": 1.8848942518234253, + "learning_rate": 3.395613322502031e-06, + "loss": 2.123, + "step": 1080 + }, + { + "epoch": 1.1835616438356165, + "eval_loss": 2.288412094116211, + "eval_runtime": 37.7637, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 1080 + }, + { + "epoch": 1.1846575342465753, + "grad_norm": 1.6747279167175293, + "learning_rate": 3.3935824532900086e-06, + "loss": 2.1087, + "step": 1081 + }, + { + "epoch": 1.1857534246575343, + "grad_norm": 1.8993490934371948, + "learning_rate": 3.3915515840779855e-06, + "loss": 2.1316, + "step": 1082 + }, + { + "epoch": 1.1868493150684931, + "grad_norm": 1.7047377824783325, + "learning_rate": 3.389520714865963e-06, + "loss": 2.1242, + "step": 1083 + }, + { + "epoch": 1.1879452054794521, + "grad_norm": 1.767811894416809, + "learning_rate": 3.38748984565394e-06, + "loss": 2.1556, + "step": 1084 + }, + { + "epoch": 1.189041095890411, + "grad_norm": 1.665313720703125, + "learning_rate": 3.3854589764419177e-06, + "loss": 2.101, + "step": 1085 + }, + { + "epoch": 1.1901369863013698, + "grad_norm": 1.7752090692520142, + "learning_rate": 3.3834281072298946e-06, + "loss": 2.1307, + "step": 1086 + }, + { + "epoch": 1.1912328767123288, + "grad_norm": 1.9759247303009033, + "learning_rate": 3.381397238017872e-06, + "loss": 2.1093, + "step": 1087 + }, + { + "epoch": 1.1923287671232876, + "grad_norm": 1.8772896528244019, + "learning_rate": 3.379366368805849e-06, + "loss": 2.1851, + "step": 1088 + }, + { + "epoch": 1.1934246575342466, + "grad_norm": 1.7442713975906372, + "learning_rate": 3.3773354995938264e-06, + "loss": 2.1433, + "step": 1089 + }, + { + "epoch": 1.1945205479452055, + "grad_norm": 1.8106931447982788, + "learning_rate": 3.3753046303818037e-06, + "loss": 2.1246, + "step": 1090 + }, + { + "epoch": 1.1945205479452055, + "eval_loss": 2.2858333587646484, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1090 + }, + { + "epoch": 1.1956164383561645, + "grad_norm": 1.7938610315322876, + "learning_rate": 3.373273761169781e-06, + "loss": 2.1325, + "step": 1091 + }, + { + "epoch": 1.1967123287671233, + "grad_norm": 1.5760858058929443, + "learning_rate": 3.3712428919577582e-06, + "loss": 2.1036, + "step": 1092 + }, + { + "epoch": 1.1978082191780821, + "grad_norm": 1.8173094987869263, + "learning_rate": 3.3692120227457355e-06, + "loss": 2.1509, + "step": 1093 + }, + { + "epoch": 1.1989041095890411, + "grad_norm": 2.0316812992095947, + "learning_rate": 3.3671811535337128e-06, + "loss": 2.0883, + "step": 1094 + }, + { + "epoch": 1.2, + "grad_norm": 1.7735658884048462, + "learning_rate": 3.36515028432169e-06, + "loss": 2.1038, + "step": 1095 + }, + { + "epoch": 1.201095890410959, + "grad_norm": 1.752718210220337, + "learning_rate": 3.3631194151096673e-06, + "loss": 2.1393, + "step": 1096 + }, + { + "epoch": 1.2021917808219178, + "grad_norm": 1.778838872909546, + "learning_rate": 3.3610885458976446e-06, + "loss": 2.1641, + "step": 1097 + }, + { + "epoch": 1.2032876712328768, + "grad_norm": 1.7422083616256714, + "learning_rate": 3.359057676685622e-06, + "loss": 2.1624, + "step": 1098 + }, + { + "epoch": 1.2043835616438356, + "grad_norm": 1.7196860313415527, + "learning_rate": 3.357026807473599e-06, + "loss": 2.0845, + "step": 1099 + }, + { + "epoch": 1.2054794520547945, + "grad_norm": 1.8373215198516846, + "learning_rate": 3.354995938261576e-06, + "loss": 2.1092, + "step": 1100 + }, + { + "epoch": 1.2054794520547945, + "eval_loss": 2.285470485687256, + "eval_runtime": 37.728, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 1100 + }, + { + "epoch": 1.2065753424657535, + "grad_norm": 1.626450538635254, + "learning_rate": 3.3529650690495537e-06, + "loss": 2.0789, + "step": 1101 + }, + { + "epoch": 1.2076712328767123, + "grad_norm": 1.736264705657959, + "learning_rate": 3.3509341998375305e-06, + "loss": 2.1227, + "step": 1102 + }, + { + "epoch": 1.208767123287671, + "grad_norm": 1.6195942163467407, + "learning_rate": 3.3489033306255082e-06, + "loss": 2.1301, + "step": 1103 + }, + { + "epoch": 1.2098630136986301, + "grad_norm": 1.6958918571472168, + "learning_rate": 3.346872461413485e-06, + "loss": 2.1059, + "step": 1104 + }, + { + "epoch": 1.210958904109589, + "grad_norm": 1.6385109424591064, + "learning_rate": 3.344841592201463e-06, + "loss": 2.0917, + "step": 1105 + }, + { + "epoch": 1.212054794520548, + "grad_norm": 1.6833516359329224, + "learning_rate": 3.3428107229894397e-06, + "loss": 2.1204, + "step": 1106 + }, + { + "epoch": 1.2131506849315068, + "grad_norm": 1.724780559539795, + "learning_rate": 3.3407798537774174e-06, + "loss": 2.138, + "step": 1107 + }, + { + "epoch": 1.2142465753424658, + "grad_norm": 1.7483834028244019, + "learning_rate": 3.338748984565394e-06, + "loss": 2.1628, + "step": 1108 + }, + { + "epoch": 1.2153424657534246, + "grad_norm": 1.6398268938064575, + "learning_rate": 3.336718115353372e-06, + "loss": 2.1738, + "step": 1109 + }, + { + "epoch": 1.2164383561643834, + "grad_norm": 1.6916258335113525, + "learning_rate": 3.3346872461413488e-06, + "loss": 2.1182, + "step": 1110 + }, + { + "epoch": 1.2164383561643834, + "eval_loss": 2.285295009613037, + "eval_runtime": 37.508, + "eval_samples_per_second": 26.661, + "eval_steps_per_second": 1.12, + "step": 1110 + }, + { + "epoch": 1.2175342465753425, + "grad_norm": 1.764147400856018, + "learning_rate": 3.3326563769293256e-06, + "loss": 2.1085, + "step": 1111 + }, + { + "epoch": 1.2186301369863013, + "grad_norm": 1.5984162092208862, + "learning_rate": 3.3306255077173033e-06, + "loss": 2.1208, + "step": 1112 + }, + { + "epoch": 1.2197260273972603, + "grad_norm": 1.6280403137207031, + "learning_rate": 3.32859463850528e-06, + "loss": 2.1226, + "step": 1113 + }, + { + "epoch": 1.2208219178082191, + "grad_norm": 1.5922425985336304, + "learning_rate": 3.326563769293258e-06, + "loss": 2.071, + "step": 1114 + }, + { + "epoch": 1.2219178082191782, + "grad_norm": 1.5941506624221802, + "learning_rate": 3.3245329000812347e-06, + "loss": 2.1059, + "step": 1115 + }, + { + "epoch": 1.223013698630137, + "grad_norm": 1.7812453508377075, + "learning_rate": 3.3225020308692124e-06, + "loss": 2.1192, + "step": 1116 + }, + { + "epoch": 1.2241095890410958, + "grad_norm": 1.6033459901809692, + "learning_rate": 3.3204711616571893e-06, + "loss": 2.0898, + "step": 1117 + }, + { + "epoch": 1.2252054794520548, + "grad_norm": 1.7053911685943604, + "learning_rate": 3.318440292445167e-06, + "loss": 2.1237, + "step": 1118 + }, + { + "epoch": 1.2263013698630136, + "grad_norm": 1.699777364730835, + "learning_rate": 3.316409423233144e-06, + "loss": 2.1014, + "step": 1119 + }, + { + "epoch": 1.2273972602739727, + "grad_norm": 1.7440428733825684, + "learning_rate": 3.3143785540211215e-06, + "loss": 2.1777, + "step": 1120 + }, + { + "epoch": 1.2273972602739727, + "eval_loss": 2.2857329845428467, + "eval_runtime": 37.7298, + "eval_samples_per_second": 26.504, + "eval_steps_per_second": 1.113, + "step": 1120 + }, + { + "epoch": 1.2284931506849315, + "grad_norm": 1.6901912689208984, + "learning_rate": 3.3123476848090984e-06, + "loss": 2.1373, + "step": 1121 + }, + { + "epoch": 1.2295890410958905, + "grad_norm": 1.5694825649261475, + "learning_rate": 3.310316815597076e-06, + "loss": 2.1285, + "step": 1122 + }, + { + "epoch": 1.2306849315068493, + "grad_norm": 1.6858313083648682, + "learning_rate": 3.308285946385053e-06, + "loss": 2.1858, + "step": 1123 + }, + { + "epoch": 1.2317808219178081, + "grad_norm": 1.6308691501617432, + "learning_rate": 3.30625507717303e-06, + "loss": 2.0793, + "step": 1124 + }, + { + "epoch": 1.2328767123287672, + "grad_norm": 1.6708226203918457, + "learning_rate": 3.3042242079610075e-06, + "loss": 2.1445, + "step": 1125 + }, + { + "epoch": 1.233972602739726, + "grad_norm": 1.7215708494186401, + "learning_rate": 3.3021933387489847e-06, + "loss": 2.159, + "step": 1126 + }, + { + "epoch": 1.235068493150685, + "grad_norm": 1.5764784812927246, + "learning_rate": 3.300162469536962e-06, + "loss": 2.0993, + "step": 1127 + }, + { + "epoch": 1.2361643835616438, + "grad_norm": 1.6322427988052368, + "learning_rate": 3.2981316003249393e-06, + "loss": 2.1643, + "step": 1128 + }, + { + "epoch": 1.2372602739726029, + "grad_norm": 1.6528997421264648, + "learning_rate": 3.2961007311129166e-06, + "loss": 2.1332, + "step": 1129 + }, + { + "epoch": 1.2383561643835617, + "grad_norm": 1.6434602737426758, + "learning_rate": 3.294069861900894e-06, + "loss": 2.1875, + "step": 1130 + }, + { + "epoch": 1.2383561643835617, + "eval_loss": 2.2840054035186768, + "eval_runtime": 37.7227, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 1130 + }, + { + "epoch": 1.2394520547945205, + "grad_norm": 1.6974412202835083, + "learning_rate": 3.292038992688871e-06, + "loss": 2.1396, + "step": 1131 + }, + { + "epoch": 1.2405479452054795, + "grad_norm": 1.7754731178283691, + "learning_rate": 3.2900081234768484e-06, + "loss": 2.1448, + "step": 1132 + }, + { + "epoch": 1.2416438356164383, + "grad_norm": 1.5928994417190552, + "learning_rate": 3.2879772542648257e-06, + "loss": 2.1665, + "step": 1133 + }, + { + "epoch": 1.2427397260273974, + "grad_norm": 1.609265685081482, + "learning_rate": 3.285946385052803e-06, + "loss": 2.0787, + "step": 1134 + }, + { + "epoch": 1.2438356164383562, + "grad_norm": 1.7366653680801392, + "learning_rate": 3.2839155158407802e-06, + "loss": 2.1453, + "step": 1135 + }, + { + "epoch": 1.244931506849315, + "grad_norm": 1.5945746898651123, + "learning_rate": 3.2818846466287575e-06, + "loss": 2.0705, + "step": 1136 + }, + { + "epoch": 1.246027397260274, + "grad_norm": 1.55001962184906, + "learning_rate": 3.2798537774167344e-06, + "loss": 2.1802, + "step": 1137 + }, + { + "epoch": 1.2471232876712328, + "grad_norm": 1.6575016975402832, + "learning_rate": 3.277822908204712e-06, + "loss": 2.1205, + "step": 1138 + }, + { + "epoch": 1.2482191780821918, + "grad_norm": 1.7479991912841797, + "learning_rate": 3.275792038992689e-06, + "loss": 2.0747, + "step": 1139 + }, + { + "epoch": 1.2493150684931507, + "grad_norm": 1.5351295471191406, + "learning_rate": 3.2737611697806666e-06, + "loss": 2.1263, + "step": 1140 + }, + { + "epoch": 1.2493150684931507, + "eval_loss": 2.286621570587158, + "eval_runtime": 37.744, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1140 + }, + { + "epoch": 1.2504109589041095, + "grad_norm": 1.6440327167510986, + "learning_rate": 3.2717303005686435e-06, + "loss": 2.0967, + "step": 1141 + }, + { + "epoch": 1.2515068493150685, + "grad_norm": 1.6527445316314697, + "learning_rate": 3.269699431356621e-06, + "loss": 2.1375, + "step": 1142 + }, + { + "epoch": 1.2526027397260273, + "grad_norm": 1.6786249876022339, + "learning_rate": 3.267668562144598e-06, + "loss": 2.1292, + "step": 1143 + }, + { + "epoch": 1.2536986301369863, + "grad_norm": 1.693476676940918, + "learning_rate": 3.2656376929325757e-06, + "loss": 2.1008, + "step": 1144 + }, + { + "epoch": 1.2547945205479452, + "grad_norm": 1.5634204149246216, + "learning_rate": 3.2636068237205526e-06, + "loss": 2.1504, + "step": 1145 + }, + { + "epoch": 1.2558904109589042, + "grad_norm": 1.6024060249328613, + "learning_rate": 3.2615759545085303e-06, + "loss": 2.1347, + "step": 1146 + }, + { + "epoch": 1.256986301369863, + "grad_norm": 1.7107738256454468, + "learning_rate": 3.259545085296507e-06, + "loss": 2.1722, + "step": 1147 + }, + { + "epoch": 1.2580821917808218, + "grad_norm": 1.6810829639434814, + "learning_rate": 3.257514216084484e-06, + "loss": 2.0948, + "step": 1148 + }, + { + "epoch": 1.2591780821917808, + "grad_norm": 1.6437987089157104, + "learning_rate": 3.2554833468724617e-06, + "loss": 2.1677, + "step": 1149 + }, + { + "epoch": 1.2602739726027397, + "grad_norm": 1.6707998514175415, + "learning_rate": 3.2534524776604385e-06, + "loss": 2.136, + "step": 1150 + }, + { + "epoch": 1.2602739726027397, + "eval_loss": 2.2854597568511963, + "eval_runtime": 37.72, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 1150 + }, + { + "epoch": 1.2613698630136987, + "grad_norm": 1.653172492980957, + "learning_rate": 3.2514216084484162e-06, + "loss": 2.1278, + "step": 1151 + }, + { + "epoch": 1.2624657534246575, + "grad_norm": 1.6651742458343506, + "learning_rate": 3.249390739236393e-06, + "loss": 2.1401, + "step": 1152 + }, + { + "epoch": 1.2635616438356165, + "grad_norm": 1.6661403179168701, + "learning_rate": 3.2473598700243708e-06, + "loss": 2.1067, + "step": 1153 + }, + { + "epoch": 1.2646575342465753, + "grad_norm": 1.6470211744308472, + "learning_rate": 3.2453290008123476e-06, + "loss": 2.1309, + "step": 1154 + }, + { + "epoch": 1.2657534246575342, + "grad_norm": 1.8089711666107178, + "learning_rate": 3.2432981316003253e-06, + "loss": 2.139, + "step": 1155 + }, + { + "epoch": 1.2668493150684932, + "grad_norm": 1.7904900312423706, + "learning_rate": 3.241267262388302e-06, + "loss": 2.0906, + "step": 1156 + }, + { + "epoch": 1.267945205479452, + "grad_norm": 1.6502841711044312, + "learning_rate": 3.23923639317628e-06, + "loss": 2.1263, + "step": 1157 + }, + { + "epoch": 1.269041095890411, + "grad_norm": 1.5825791358947754, + "learning_rate": 3.2372055239642567e-06, + "loss": 2.098, + "step": 1158 + }, + { + "epoch": 1.2701369863013698, + "grad_norm": 1.7173693180084229, + "learning_rate": 3.2351746547522344e-06, + "loss": 2.1367, + "step": 1159 + }, + { + "epoch": 1.2712328767123289, + "grad_norm": 1.766738772392273, + "learning_rate": 3.2331437855402113e-06, + "loss": 2.1914, + "step": 1160 + }, + { + "epoch": 1.2712328767123289, + "eval_loss": 2.282374858856201, + "eval_runtime": 37.7395, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1160 + }, + { + "epoch": 1.2723287671232877, + "grad_norm": 1.6669079065322876, + "learning_rate": 3.2311129163281885e-06, + "loss": 2.0959, + "step": 1161 + }, + { + "epoch": 1.2734246575342465, + "grad_norm": 1.7150713205337524, + "learning_rate": 3.229082047116166e-06, + "loss": 2.1431, + "step": 1162 + }, + { + "epoch": 1.2745205479452055, + "grad_norm": 1.5927042961120605, + "learning_rate": 3.227051177904143e-06, + "loss": 2.1239, + "step": 1163 + }, + { + "epoch": 1.2756164383561643, + "grad_norm": 1.8086377382278442, + "learning_rate": 3.2250203086921204e-06, + "loss": 2.1307, + "step": 1164 + }, + { + "epoch": 1.2767123287671232, + "grad_norm": 1.8078030347824097, + "learning_rate": 3.2229894394800977e-06, + "loss": 2.1632, + "step": 1165 + }, + { + "epoch": 1.2778082191780822, + "grad_norm": 1.7087128162384033, + "learning_rate": 3.220958570268075e-06, + "loss": 2.0663, + "step": 1166 + }, + { + "epoch": 1.2789041095890412, + "grad_norm": 1.8242524862289429, + "learning_rate": 3.218927701056052e-06, + "loss": 2.1225, + "step": 1167 + }, + { + "epoch": 1.28, + "grad_norm": 1.6094539165496826, + "learning_rate": 3.2168968318440295e-06, + "loss": 2.0586, + "step": 1168 + }, + { + "epoch": 1.2810958904109588, + "grad_norm": 1.8592051267623901, + "learning_rate": 3.2148659626320068e-06, + "loss": 2.1632, + "step": 1169 + }, + { + "epoch": 1.2821917808219179, + "grad_norm": 1.6238080263137817, + "learning_rate": 3.212835093419984e-06, + "loss": 2.1002, + "step": 1170 + }, + { + "epoch": 1.2821917808219179, + "eval_loss": 2.285175323486328, + "eval_runtime": 37.7569, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1170 + }, + { + "epoch": 1.2832876712328767, + "grad_norm": 1.631516933441162, + "learning_rate": 3.2108042242079613e-06, + "loss": 2.1094, + "step": 1171 + }, + { + "epoch": 1.2843835616438355, + "grad_norm": 1.6518285274505615, + "learning_rate": 3.2087733549959386e-06, + "loss": 2.1614, + "step": 1172 + }, + { + "epoch": 1.2854794520547945, + "grad_norm": 1.64463210105896, + "learning_rate": 3.206742485783916e-06, + "loss": 2.0863, + "step": 1173 + }, + { + "epoch": 1.2865753424657536, + "grad_norm": 1.5903335809707642, + "learning_rate": 3.2047116165718927e-06, + "loss": 2.0795, + "step": 1174 + }, + { + "epoch": 1.2876712328767124, + "grad_norm": 1.6504883766174316, + "learning_rate": 3.2026807473598704e-06, + "loss": 2.1352, + "step": 1175 + }, + { + "epoch": 1.2887671232876712, + "grad_norm": 1.7765742540359497, + "learning_rate": 3.2006498781478473e-06, + "loss": 2.1511, + "step": 1176 + }, + { + "epoch": 1.2898630136986302, + "grad_norm": 1.7253892421722412, + "learning_rate": 3.198619008935825e-06, + "loss": 2.0768, + "step": 1177 + }, + { + "epoch": 1.290958904109589, + "grad_norm": 1.7327924966812134, + "learning_rate": 3.196588139723802e-06, + "loss": 2.1196, + "step": 1178 + }, + { + "epoch": 1.2920547945205478, + "grad_norm": 1.6659244298934937, + "learning_rate": 3.1945572705117795e-06, + "loss": 2.1398, + "step": 1179 + }, + { + "epoch": 1.2931506849315069, + "grad_norm": 1.8803845643997192, + "learning_rate": 3.1925264012997564e-06, + "loss": 2.1382, + "step": 1180 + }, + { + "epoch": 1.2931506849315069, + "eval_loss": 2.284956455230713, + "eval_runtime": 37.7442, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1180 + }, + { + "epoch": 1.2942465753424657, + "grad_norm": 1.6562074422836304, + "learning_rate": 3.190495532087734e-06, + "loss": 2.1403, + "step": 1181 + }, + { + "epoch": 1.2953424657534247, + "grad_norm": 1.6407029628753662, + "learning_rate": 3.188464662875711e-06, + "loss": 2.1782, + "step": 1182 + }, + { + "epoch": 1.2964383561643835, + "grad_norm": 1.638464331626892, + "learning_rate": 3.1864337936636886e-06, + "loss": 2.1129, + "step": 1183 + }, + { + "epoch": 1.2975342465753426, + "grad_norm": 1.6085008382797241, + "learning_rate": 3.1844029244516655e-06, + "loss": 2.0967, + "step": 1184 + }, + { + "epoch": 1.2986301369863014, + "grad_norm": 1.9340208768844604, + "learning_rate": 3.182372055239643e-06, + "loss": 2.0906, + "step": 1185 + }, + { + "epoch": 1.2997260273972602, + "grad_norm": 1.6958277225494385, + "learning_rate": 3.18034118602762e-06, + "loss": 2.1519, + "step": 1186 + }, + { + "epoch": 1.3008219178082192, + "grad_norm": 1.6682950258255005, + "learning_rate": 3.1783103168155973e-06, + "loss": 2.1286, + "step": 1187 + }, + { + "epoch": 1.301917808219178, + "grad_norm": 1.7564821243286133, + "learning_rate": 3.1762794476035746e-06, + "loss": 2.1153, + "step": 1188 + }, + { + "epoch": 1.303013698630137, + "grad_norm": 1.765539526939392, + "learning_rate": 3.174248578391552e-06, + "loss": 2.112, + "step": 1189 + }, + { + "epoch": 1.3041095890410959, + "grad_norm": 1.8028367757797241, + "learning_rate": 3.172217709179529e-06, + "loss": 2.1169, + "step": 1190 + }, + { + "epoch": 1.3041095890410959, + "eval_loss": 2.2839744091033936, + "eval_runtime": 37.7782, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 1190 + }, + { + "epoch": 1.305205479452055, + "grad_norm": 1.6870243549346924, + "learning_rate": 3.1701868399675064e-06, + "loss": 2.1147, + "step": 1191 + }, + { + "epoch": 1.3063013698630137, + "grad_norm": 1.6161645650863647, + "learning_rate": 3.1681559707554837e-06, + "loss": 2.081, + "step": 1192 + }, + { + "epoch": 1.3073972602739725, + "grad_norm": 1.5438032150268555, + "learning_rate": 3.166125101543461e-06, + "loss": 2.139, + "step": 1193 + }, + { + "epoch": 1.3084931506849315, + "grad_norm": 1.6329416036605835, + "learning_rate": 3.1640942323314382e-06, + "loss": 2.1346, + "step": 1194 + }, + { + "epoch": 1.3095890410958904, + "grad_norm": 1.6933001279830933, + "learning_rate": 3.1620633631194155e-06, + "loss": 2.0906, + "step": 1195 + }, + { + "epoch": 1.3106849315068494, + "grad_norm": 1.5968612432479858, + "learning_rate": 3.1600324939073928e-06, + "loss": 2.1191, + "step": 1196 + }, + { + "epoch": 1.3117808219178082, + "grad_norm": 1.654739499092102, + "learning_rate": 3.15800162469537e-06, + "loss": 2.1253, + "step": 1197 + }, + { + "epoch": 1.3128767123287672, + "grad_norm": 1.6595433950424194, + "learning_rate": 3.155970755483347e-06, + "loss": 2.1326, + "step": 1198 + }, + { + "epoch": 1.313972602739726, + "grad_norm": 1.5709439516067505, + "learning_rate": 3.1539398862713246e-06, + "loss": 2.1427, + "step": 1199 + }, + { + "epoch": 1.3150684931506849, + "grad_norm": 1.7586764097213745, + "learning_rate": 3.1519090170593015e-06, + "loss": 2.1311, + "step": 1200 + }, + { + "epoch": 1.3150684931506849, + "eval_loss": 2.2803795337677, + "eval_runtime": 37.7752, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 1200 + }, + { + "epoch": 1.316164383561644, + "grad_norm": 1.675703763961792, + "learning_rate": 3.149878147847279e-06, + "loss": 2.1679, + "step": 1201 + }, + { + "epoch": 1.3172602739726027, + "grad_norm": 1.632204294204712, + "learning_rate": 3.147847278635256e-06, + "loss": 2.1889, + "step": 1202 + }, + { + "epoch": 1.3183561643835615, + "grad_norm": 1.6142032146453857, + "learning_rate": 3.1458164094232337e-06, + "loss": 2.1186, + "step": 1203 + }, + { + "epoch": 1.3194520547945205, + "grad_norm": 1.5145642757415771, + "learning_rate": 3.1437855402112106e-06, + "loss": 2.0699, + "step": 1204 + }, + { + "epoch": 1.3205479452054796, + "grad_norm": 1.6954951286315918, + "learning_rate": 3.1417546709991883e-06, + "loss": 2.1435, + "step": 1205 + }, + { + "epoch": 1.3216438356164384, + "grad_norm": 1.5725785493850708, + "learning_rate": 3.139723801787165e-06, + "loss": 2.1105, + "step": 1206 + }, + { + "epoch": 1.3227397260273972, + "grad_norm": 1.6296924352645874, + "learning_rate": 3.137692932575143e-06, + "loss": 2.1526, + "step": 1207 + }, + { + "epoch": 1.3238356164383562, + "grad_norm": 1.758583426475525, + "learning_rate": 3.1356620633631197e-06, + "loss": 2.0966, + "step": 1208 + }, + { + "epoch": 1.324931506849315, + "grad_norm": 1.6128432750701904, + "learning_rate": 3.1336311941510974e-06, + "loss": 2.1227, + "step": 1209 + }, + { + "epoch": 1.3260273972602739, + "grad_norm": 1.6251143217086792, + "learning_rate": 3.1316003249390742e-06, + "loss": 2.1557, + "step": 1210 + }, + { + "epoch": 1.3260273972602739, + "eval_loss": 2.277153253555298, + "eval_runtime": 37.4484, + "eval_samples_per_second": 26.703, + "eval_steps_per_second": 1.122, + "step": 1210 + }, + { + "epoch": 1.3271232876712329, + "grad_norm": 1.6224772930145264, + "learning_rate": 3.129569455727051e-06, + "loss": 2.1515, + "step": 1211 + }, + { + "epoch": 1.3282191780821917, + "grad_norm": 1.6408687829971313, + "learning_rate": 3.1275385865150288e-06, + "loss": 2.1415, + "step": 1212 + }, + { + "epoch": 1.3293150684931507, + "grad_norm": 1.5633060932159424, + "learning_rate": 3.1255077173030056e-06, + "loss": 2.1369, + "step": 1213 + }, + { + "epoch": 1.3304109589041095, + "grad_norm": 1.6440671682357788, + "learning_rate": 3.1234768480909833e-06, + "loss": 2.1302, + "step": 1214 + }, + { + "epoch": 1.3315068493150686, + "grad_norm": 1.6065458059310913, + "learning_rate": 3.12144597887896e-06, + "loss": 2.1359, + "step": 1215 + }, + { + "epoch": 1.3326027397260274, + "grad_norm": 1.6209193468093872, + "learning_rate": 3.119415109666938e-06, + "loss": 2.1669, + "step": 1216 + }, + { + "epoch": 1.3336986301369862, + "grad_norm": 1.745129942893982, + "learning_rate": 3.1173842404549147e-06, + "loss": 2.1318, + "step": 1217 + }, + { + "epoch": 1.3347945205479452, + "grad_norm": 1.6090255975723267, + "learning_rate": 3.1153533712428924e-06, + "loss": 2.1318, + "step": 1218 + }, + { + "epoch": 1.335890410958904, + "grad_norm": 1.5840481519699097, + "learning_rate": 3.1133225020308693e-06, + "loss": 2.152, + "step": 1219 + }, + { + "epoch": 1.336986301369863, + "grad_norm": 1.6378774642944336, + "learning_rate": 3.111291632818847e-06, + "loss": 2.1597, + "step": 1220 + }, + { + "epoch": 1.336986301369863, + "eval_loss": 2.2799103260040283, + "eval_runtime": 37.6598, + "eval_samples_per_second": 26.554, + "eval_steps_per_second": 1.115, + "step": 1220 + }, + { + "epoch": 1.3380821917808219, + "grad_norm": 1.7247775793075562, + "learning_rate": 3.109260763606824e-06, + "loss": 2.1328, + "step": 1221 + }, + { + "epoch": 1.339178082191781, + "grad_norm": 1.6469335556030273, + "learning_rate": 3.1072298943948015e-06, + "loss": 2.1434, + "step": 1222 + }, + { + "epoch": 1.3402739726027397, + "grad_norm": 1.5268198251724243, + "learning_rate": 3.1051990251827784e-06, + "loss": 2.1077, + "step": 1223 + }, + { + "epoch": 1.3413698630136985, + "grad_norm": 1.6413369178771973, + "learning_rate": 3.1031681559707557e-06, + "loss": 2.1583, + "step": 1224 + }, + { + "epoch": 1.3424657534246576, + "grad_norm": 1.7595359086990356, + "learning_rate": 3.101137286758733e-06, + "loss": 2.1625, + "step": 1225 + }, + { + "epoch": 1.3435616438356164, + "grad_norm": 1.6333140134811401, + "learning_rate": 3.09910641754671e-06, + "loss": 2.1555, + "step": 1226 + }, + { + "epoch": 1.3446575342465754, + "grad_norm": 1.8567262887954712, + "learning_rate": 3.0970755483346875e-06, + "loss": 2.1403, + "step": 1227 + }, + { + "epoch": 1.3457534246575342, + "grad_norm": 1.5342068672180176, + "learning_rate": 3.0950446791226648e-06, + "loss": 2.1686, + "step": 1228 + }, + { + "epoch": 1.3468493150684933, + "grad_norm": 1.6217644214630127, + "learning_rate": 3.093013809910642e-06, + "loss": 2.1345, + "step": 1229 + }, + { + "epoch": 1.347945205479452, + "grad_norm": 1.7755476236343384, + "learning_rate": 3.0909829406986193e-06, + "loss": 2.1769, + "step": 1230 + }, + { + "epoch": 1.347945205479452, + "eval_loss": 2.2792110443115234, + "eval_runtime": 37.6822, + "eval_samples_per_second": 26.538, + "eval_steps_per_second": 1.115, + "step": 1230 + }, + { + "epoch": 1.3490410958904109, + "grad_norm": 1.4750087261199951, + "learning_rate": 3.0889520714865966e-06, + "loss": 2.1501, + "step": 1231 + }, + { + "epoch": 1.35013698630137, + "grad_norm": 1.7657350301742554, + "learning_rate": 3.086921202274574e-06, + "loss": 2.1254, + "step": 1232 + }, + { + "epoch": 1.3512328767123287, + "grad_norm": 1.6447482109069824, + "learning_rate": 3.084890333062551e-06, + "loss": 2.1004, + "step": 1233 + }, + { + "epoch": 1.3523287671232875, + "grad_norm": 1.658786654472351, + "learning_rate": 3.0828594638505284e-06, + "loss": 2.1276, + "step": 1234 + }, + { + "epoch": 1.3534246575342466, + "grad_norm": 1.5626777410507202, + "learning_rate": 3.0808285946385057e-06, + "loss": 2.1067, + "step": 1235 + }, + { + "epoch": 1.3545205479452056, + "grad_norm": 1.7021020650863647, + "learning_rate": 3.078797725426483e-06, + "loss": 2.0656, + "step": 1236 + }, + { + "epoch": 1.3556164383561644, + "grad_norm": 1.5983797311782837, + "learning_rate": 3.07676685621446e-06, + "loss": 2.1319, + "step": 1237 + }, + { + "epoch": 1.3567123287671232, + "grad_norm": 1.5506508350372314, + "learning_rate": 3.0747359870024375e-06, + "loss": 2.1008, + "step": 1238 + }, + { + "epoch": 1.3578082191780823, + "grad_norm": 1.5998226404190063, + "learning_rate": 3.0727051177904144e-06, + "loss": 2.1102, + "step": 1239 + }, + { + "epoch": 1.358904109589041, + "grad_norm": 1.6222437620162964, + "learning_rate": 3.070674248578392e-06, + "loss": 2.1352, + "step": 1240 + }, + { + "epoch": 1.358904109589041, + "eval_loss": 2.278921127319336, + "eval_runtime": 37.6613, + "eval_samples_per_second": 26.552, + "eval_steps_per_second": 1.115, + "step": 1240 + }, + { + "epoch": 1.3599999999999999, + "grad_norm": 1.600468397140503, + "learning_rate": 3.068643379366369e-06, + "loss": 2.1115, + "step": 1241 + }, + { + "epoch": 1.361095890410959, + "grad_norm": 1.6324045658111572, + "learning_rate": 3.0666125101543466e-06, + "loss": 2.0887, + "step": 1242 + }, + { + "epoch": 1.362191780821918, + "grad_norm": 1.7005919218063354, + "learning_rate": 3.0645816409423235e-06, + "loss": 2.1344, + "step": 1243 + }, + { + "epoch": 1.3632876712328768, + "grad_norm": 1.6483793258666992, + "learning_rate": 3.062550771730301e-06, + "loss": 2.1165, + "step": 1244 + }, + { + "epoch": 1.3643835616438356, + "grad_norm": 1.774255633354187, + "learning_rate": 3.060519902518278e-06, + "loss": 2.1137, + "step": 1245 + }, + { + "epoch": 1.3654794520547946, + "grad_norm": 1.7157840728759766, + "learning_rate": 3.0584890333062557e-06, + "loss": 2.0656, + "step": 1246 + }, + { + "epoch": 1.3665753424657534, + "grad_norm": 1.6053105592727661, + "learning_rate": 3.0564581640942326e-06, + "loss": 2.128, + "step": 1247 + }, + { + "epoch": 1.3676712328767122, + "grad_norm": 1.8893406391143799, + "learning_rate": 3.0544272948822094e-06, + "loss": 2.106, + "step": 1248 + }, + { + "epoch": 1.3687671232876713, + "grad_norm": 1.6212185621261597, + "learning_rate": 3.052396425670187e-06, + "loss": 2.1422, + "step": 1249 + }, + { + "epoch": 1.36986301369863, + "grad_norm": 1.595241665840149, + "learning_rate": 3.050365556458164e-06, + "loss": 2.1132, + "step": 1250 + }, + { + "epoch": 1.36986301369863, + "eval_loss": 2.280195474624634, + "eval_runtime": 37.69, + "eval_samples_per_second": 26.532, + "eval_steps_per_second": 1.114, + "step": 1250 + }, + { + "epoch": 1.370958904109589, + "grad_norm": 1.7002182006835938, + "learning_rate": 3.0483346872461417e-06, + "loss": 2.1088, + "step": 1251 + }, + { + "epoch": 1.372054794520548, + "grad_norm": 1.7393778562545776, + "learning_rate": 3.0463038180341185e-06, + "loss": 2.1371, + "step": 1252 + }, + { + "epoch": 1.373150684931507, + "grad_norm": 1.5694235563278198, + "learning_rate": 3.0442729488220962e-06, + "loss": 2.1514, + "step": 1253 + }, + { + "epoch": 1.3742465753424657, + "grad_norm": 1.689393401145935, + "learning_rate": 3.042242079610073e-06, + "loss": 2.1512, + "step": 1254 + }, + { + "epoch": 1.3753424657534246, + "grad_norm": 1.6431982517242432, + "learning_rate": 3.0402112103980508e-06, + "loss": 2.0749, + "step": 1255 + }, + { + "epoch": 1.3764383561643836, + "grad_norm": 1.6122742891311646, + "learning_rate": 3.0381803411860276e-06, + "loss": 2.1024, + "step": 1256 + }, + { + "epoch": 1.3775342465753424, + "grad_norm": 1.5770806074142456, + "learning_rate": 3.0361494719740053e-06, + "loss": 2.1208, + "step": 1257 + }, + { + "epoch": 1.3786301369863014, + "grad_norm": 1.7665551900863647, + "learning_rate": 3.034118602761982e-06, + "loss": 2.1209, + "step": 1258 + }, + { + "epoch": 1.3797260273972602, + "grad_norm": 1.6896030902862549, + "learning_rate": 3.03208773354996e-06, + "loss": 2.0986, + "step": 1259 + }, + { + "epoch": 1.3808219178082193, + "grad_norm": 1.7461376190185547, + "learning_rate": 3.0300568643379367e-06, + "loss": 2.156, + "step": 1260 + }, + { + "epoch": 1.3808219178082193, + "eval_loss": 2.27716064453125, + "eval_runtime": 37.7183, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 1260 + }, + { + "epoch": 1.381917808219178, + "grad_norm": 1.6952084302902222, + "learning_rate": 3.028025995125914e-06, + "loss": 2.1449, + "step": 1261 + }, + { + "epoch": 1.383013698630137, + "grad_norm": 1.5834839344024658, + "learning_rate": 3.0259951259138913e-06, + "loss": 2.082, + "step": 1262 + }, + { + "epoch": 1.384109589041096, + "grad_norm": 1.7395371198654175, + "learning_rate": 3.0239642567018686e-06, + "loss": 2.126, + "step": 1263 + }, + { + "epoch": 1.3852054794520547, + "grad_norm": 1.6316404342651367, + "learning_rate": 3.021933387489846e-06, + "loss": 2.0833, + "step": 1264 + }, + { + "epoch": 1.3863013698630138, + "grad_norm": 1.6206353902816772, + "learning_rate": 3.019902518277823e-06, + "loss": 2.1311, + "step": 1265 + }, + { + "epoch": 1.3873972602739726, + "grad_norm": 1.632569432258606, + "learning_rate": 3.0178716490658004e-06, + "loss": 2.1468, + "step": 1266 + }, + { + "epoch": 1.3884931506849316, + "grad_norm": 1.64067804813385, + "learning_rate": 3.0158407798537777e-06, + "loss": 2.0859, + "step": 1267 + }, + { + "epoch": 1.3895890410958904, + "grad_norm": 1.54572331905365, + "learning_rate": 3.013809910641755e-06, + "loss": 2.0992, + "step": 1268 + }, + { + "epoch": 1.3906849315068492, + "grad_norm": 1.5950230360031128, + "learning_rate": 3.0117790414297322e-06, + "loss": 2.0895, + "step": 1269 + }, + { + "epoch": 1.3917808219178083, + "grad_norm": 1.8030891418457031, + "learning_rate": 3.0097481722177095e-06, + "loss": 2.1165, + "step": 1270 + }, + { + "epoch": 1.3917808219178083, + "eval_loss": 2.2781739234924316, + "eval_runtime": 37.7428, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1270 + }, + { + "epoch": 1.392876712328767, + "grad_norm": 1.5848240852355957, + "learning_rate": 3.0077173030056868e-06, + "loss": 2.1015, + "step": 1271 + }, + { + "epoch": 1.393972602739726, + "grad_norm": 1.682953953742981, + "learning_rate": 3.005686433793664e-06, + "loss": 2.1341, + "step": 1272 + }, + { + "epoch": 1.395068493150685, + "grad_norm": 1.8385032415390015, + "learning_rate": 3.0036555645816413e-06, + "loss": 2.111, + "step": 1273 + }, + { + "epoch": 1.396164383561644, + "grad_norm": 1.5879863500595093, + "learning_rate": 3.001624695369618e-06, + "loss": 2.1418, + "step": 1274 + }, + { + "epoch": 1.3972602739726028, + "grad_norm": 1.8076952695846558, + "learning_rate": 2.999593826157596e-06, + "loss": 2.1076, + "step": 1275 + }, + { + "epoch": 1.3983561643835616, + "grad_norm": 1.6944698095321655, + "learning_rate": 2.9975629569455727e-06, + "loss": 2.1415, + "step": 1276 + }, + { + "epoch": 1.3994520547945206, + "grad_norm": 1.7351915836334229, + "learning_rate": 2.9955320877335504e-06, + "loss": 2.1186, + "step": 1277 + }, + { + "epoch": 1.4005479452054794, + "grad_norm": 1.6877539157867432, + "learning_rate": 2.9935012185215273e-06, + "loss": 2.1374, + "step": 1278 + }, + { + "epoch": 1.4016438356164382, + "grad_norm": 1.6253048181533813, + "learning_rate": 2.991470349309505e-06, + "loss": 2.114, + "step": 1279 + }, + { + "epoch": 1.4027397260273973, + "grad_norm": 1.6599102020263672, + "learning_rate": 2.989439480097482e-06, + "loss": 2.1231, + "step": 1280 + }, + { + "epoch": 1.4027397260273973, + "eval_loss": 2.2765634059906006, + "eval_runtime": 37.7438, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1280 + }, + { + "epoch": 1.403835616438356, + "grad_norm": 1.6010860204696655, + "learning_rate": 2.9874086108854595e-06, + "loss": 2.108, + "step": 1281 + }, + { + "epoch": 1.4049315068493151, + "grad_norm": 1.7082704305648804, + "learning_rate": 2.9853777416734364e-06, + "loss": 2.143, + "step": 1282 + }, + { + "epoch": 1.406027397260274, + "grad_norm": 1.5962858200073242, + "learning_rate": 2.983346872461414e-06, + "loss": 2.1582, + "step": 1283 + }, + { + "epoch": 1.407123287671233, + "grad_norm": 1.678695559501648, + "learning_rate": 2.981316003249391e-06, + "loss": 2.1308, + "step": 1284 + }, + { + "epoch": 1.4082191780821918, + "grad_norm": 1.8608810901641846, + "learning_rate": 2.9792851340373678e-06, + "loss": 2.1192, + "step": 1285 + }, + { + "epoch": 1.4093150684931506, + "grad_norm": 1.6447912454605103, + "learning_rate": 2.9772542648253455e-06, + "loss": 2.149, + "step": 1286 + }, + { + "epoch": 1.4104109589041096, + "grad_norm": 1.8847064971923828, + "learning_rate": 2.9752233956133223e-06, + "loss": 2.1073, + "step": 1287 + }, + { + "epoch": 1.4115068493150684, + "grad_norm": 1.76127290725708, + "learning_rate": 2.9731925264013e-06, + "loss": 2.1123, + "step": 1288 + }, + { + "epoch": 1.4126027397260275, + "grad_norm": 1.8126407861709595, + "learning_rate": 2.971161657189277e-06, + "loss": 2.1123, + "step": 1289 + }, + { + "epoch": 1.4136986301369863, + "grad_norm": 1.8271949291229248, + "learning_rate": 2.9691307879772546e-06, + "loss": 2.1156, + "step": 1290 + }, + { + "epoch": 1.4136986301369863, + "eval_loss": 2.2737486362457275, + "eval_runtime": 37.6964, + "eval_samples_per_second": 26.528, + "eval_steps_per_second": 1.114, + "step": 1290 + }, + { + "epoch": 1.4147945205479453, + "grad_norm": 1.6088656187057495, + "learning_rate": 2.9670999187652314e-06, + "loss": 2.0904, + "step": 1291 + }, + { + "epoch": 1.415890410958904, + "grad_norm": 1.8934502601623535, + "learning_rate": 2.965069049553209e-06, + "loss": 2.1669, + "step": 1292 + }, + { + "epoch": 1.416986301369863, + "grad_norm": 1.770875334739685, + "learning_rate": 2.963038180341186e-06, + "loss": 2.0866, + "step": 1293 + }, + { + "epoch": 1.418082191780822, + "grad_norm": 1.6065635681152344, + "learning_rate": 2.9610073111291637e-06, + "loss": 2.1433, + "step": 1294 + }, + { + "epoch": 1.4191780821917808, + "grad_norm": 1.8257594108581543, + "learning_rate": 2.9589764419171405e-06, + "loss": 2.1611, + "step": 1295 + }, + { + "epoch": 1.4202739726027398, + "grad_norm": 1.7531297206878662, + "learning_rate": 2.9569455727051182e-06, + "loss": 2.0991, + "step": 1296 + }, + { + "epoch": 1.4213698630136986, + "grad_norm": 1.6781302690505981, + "learning_rate": 2.954914703493095e-06, + "loss": 2.121, + "step": 1297 + }, + { + "epoch": 1.4224657534246576, + "grad_norm": 1.7489228248596191, + "learning_rate": 2.9528838342810724e-06, + "loss": 2.1012, + "step": 1298 + }, + { + "epoch": 1.4235616438356165, + "grad_norm": 1.7776473760604858, + "learning_rate": 2.9508529650690496e-06, + "loss": 2.1066, + "step": 1299 + }, + { + "epoch": 1.4246575342465753, + "grad_norm": 1.7317523956298828, + "learning_rate": 2.948822095857027e-06, + "loss": 2.1278, + "step": 1300 + }, + { + "epoch": 1.4246575342465753, + "eval_loss": 2.2755467891693115, + "eval_runtime": 37.7024, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 1300 + }, + { + "epoch": 1.4257534246575343, + "grad_norm": 1.9439119100570679, + "learning_rate": 2.946791226645004e-06, + "loss": 2.0971, + "step": 1301 + }, + { + "epoch": 1.426849315068493, + "grad_norm": 1.683897614479065, + "learning_rate": 2.9447603574329815e-06, + "loss": 2.0895, + "step": 1302 + }, + { + "epoch": 1.427945205479452, + "grad_norm": 1.620665192604065, + "learning_rate": 2.9427294882209587e-06, + "loss": 2.1011, + "step": 1303 + }, + { + "epoch": 1.429041095890411, + "grad_norm": 1.9428532123565674, + "learning_rate": 2.940698619008936e-06, + "loss": 2.1253, + "step": 1304 + }, + { + "epoch": 1.43013698630137, + "grad_norm": 1.897630214691162, + "learning_rate": 2.9386677497969133e-06, + "loss": 2.1146, + "step": 1305 + }, + { + "epoch": 1.4312328767123288, + "grad_norm": 1.767519235610962, + "learning_rate": 2.9366368805848906e-06, + "loss": 2.1407, + "step": 1306 + }, + { + "epoch": 1.4323287671232876, + "grad_norm": 1.630615234375, + "learning_rate": 2.934606011372868e-06, + "loss": 2.1432, + "step": 1307 + }, + { + "epoch": 1.4334246575342466, + "grad_norm": 1.7037104368209839, + "learning_rate": 2.932575142160845e-06, + "loss": 2.1226, + "step": 1308 + }, + { + "epoch": 1.4345205479452054, + "grad_norm": 1.5496717691421509, + "learning_rate": 2.9305442729488224e-06, + "loss": 2.1235, + "step": 1309 + }, + { + "epoch": 1.4356164383561643, + "grad_norm": 1.6394745111465454, + "learning_rate": 2.9285134037367997e-06, + "loss": 2.1183, + "step": 1310 + }, + { + "epoch": 1.4356164383561643, + "eval_loss": 2.2730512619018555, + "eval_runtime": 37.4684, + "eval_samples_per_second": 26.689, + "eval_steps_per_second": 1.121, + "step": 1310 + }, + { + "epoch": 1.4367123287671233, + "grad_norm": 1.5624728202819824, + "learning_rate": 2.9264825345247765e-06, + "loss": 2.1297, + "step": 1311 + }, + { + "epoch": 1.4378082191780823, + "grad_norm": 1.7173740863800049, + "learning_rate": 2.9244516653127542e-06, + "loss": 2.1299, + "step": 1312 + }, + { + "epoch": 1.4389041095890411, + "grad_norm": 1.5682451725006104, + "learning_rate": 2.922420796100731e-06, + "loss": 2.0605, + "step": 1313 + }, + { + "epoch": 1.44, + "grad_norm": 1.6592832803726196, + "learning_rate": 2.9203899268887088e-06, + "loss": 2.1721, + "step": 1314 + }, + { + "epoch": 1.441095890410959, + "grad_norm": 1.7420763969421387, + "learning_rate": 2.9183590576766856e-06, + "loss": 2.0873, + "step": 1315 + }, + { + "epoch": 1.4421917808219178, + "grad_norm": 1.6949102878570557, + "learning_rate": 2.9163281884646633e-06, + "loss": 2.1213, + "step": 1316 + }, + { + "epoch": 1.4432876712328766, + "grad_norm": 1.7273104190826416, + "learning_rate": 2.91429731925264e-06, + "loss": 2.1501, + "step": 1317 + }, + { + "epoch": 1.4443835616438356, + "grad_norm": 1.684362530708313, + "learning_rate": 2.912266450040618e-06, + "loss": 2.1252, + "step": 1318 + }, + { + "epoch": 1.4454794520547944, + "grad_norm": 1.627016305923462, + "learning_rate": 2.9102355808285947e-06, + "loss": 2.1039, + "step": 1319 + }, + { + "epoch": 1.4465753424657535, + "grad_norm": 1.8114672899246216, + "learning_rate": 2.9082047116165724e-06, + "loss": 2.1049, + "step": 1320 + }, + { + "epoch": 1.4465753424657535, + "eval_loss": 2.273300886154175, + "eval_runtime": 37.6816, + "eval_samples_per_second": 26.538, + "eval_steps_per_second": 1.115, + "step": 1320 + }, + { + "epoch": 1.4476712328767123, + "grad_norm": 1.6053475141525269, + "learning_rate": 2.9061738424045493e-06, + "loss": 2.1026, + "step": 1321 + }, + { + "epoch": 1.4487671232876713, + "grad_norm": 1.757144570350647, + "learning_rate": 2.904142973192527e-06, + "loss": 2.1683, + "step": 1322 + }, + { + "epoch": 1.4498630136986301, + "grad_norm": 1.6230170726776123, + "learning_rate": 2.902112103980504e-06, + "loss": 2.1308, + "step": 1323 + }, + { + "epoch": 1.450958904109589, + "grad_norm": 1.5634756088256836, + "learning_rate": 2.900081234768481e-06, + "loss": 2.1194, + "step": 1324 + }, + { + "epoch": 1.452054794520548, + "grad_norm": 1.853056788444519, + "learning_rate": 2.8980503655564584e-06, + "loss": 2.1135, + "step": 1325 + }, + { + "epoch": 1.4531506849315068, + "grad_norm": 1.6028629541397095, + "learning_rate": 2.8960194963444357e-06, + "loss": 2.0872, + "step": 1326 + }, + { + "epoch": 1.4542465753424658, + "grad_norm": 1.657346487045288, + "learning_rate": 2.893988627132413e-06, + "loss": 2.1059, + "step": 1327 + }, + { + "epoch": 1.4553424657534246, + "grad_norm": 1.6872131824493408, + "learning_rate": 2.8919577579203902e-06, + "loss": 2.0726, + "step": 1328 + }, + { + "epoch": 1.4564383561643837, + "grad_norm": 1.677079439163208, + "learning_rate": 2.8899268887083675e-06, + "loss": 2.1199, + "step": 1329 + }, + { + "epoch": 1.4575342465753425, + "grad_norm": 1.6555675268173218, + "learning_rate": 2.8878960194963448e-06, + "loss": 2.1231, + "step": 1330 + }, + { + "epoch": 1.4575342465753425, + "eval_loss": 2.277087688446045, + "eval_runtime": 37.7399, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1330 + }, + { + "epoch": 1.4586301369863013, + "grad_norm": 1.6812067031860352, + "learning_rate": 2.885865150284322e-06, + "loss": 2.1397, + "step": 1331 + }, + { + "epoch": 1.4597260273972603, + "grad_norm": 1.5711122751235962, + "learning_rate": 2.8838342810722993e-06, + "loss": 2.1005, + "step": 1332 + }, + { + "epoch": 1.4608219178082191, + "grad_norm": 1.7067155838012695, + "learning_rate": 2.8818034118602766e-06, + "loss": 2.0874, + "step": 1333 + }, + { + "epoch": 1.4619178082191782, + "grad_norm": 1.7551555633544922, + "learning_rate": 2.879772542648254e-06, + "loss": 2.1376, + "step": 1334 + }, + { + "epoch": 1.463013698630137, + "grad_norm": 1.6309247016906738, + "learning_rate": 2.8777416734362307e-06, + "loss": 2.0595, + "step": 1335 + }, + { + "epoch": 1.464109589041096, + "grad_norm": 1.6252689361572266, + "learning_rate": 2.8757108042242084e-06, + "loss": 2.1778, + "step": 1336 + }, + { + "epoch": 1.4652054794520548, + "grad_norm": 1.5928912162780762, + "learning_rate": 2.8736799350121853e-06, + "loss": 2.1575, + "step": 1337 + }, + { + "epoch": 1.4663013698630136, + "grad_norm": 1.6079543828964233, + "learning_rate": 2.871649065800163e-06, + "loss": 2.1638, + "step": 1338 + }, + { + "epoch": 1.4673972602739727, + "grad_norm": 1.628999948501587, + "learning_rate": 2.86961819658814e-06, + "loss": 2.0999, + "step": 1339 + }, + { + "epoch": 1.4684931506849315, + "grad_norm": 1.5877981185913086, + "learning_rate": 2.8675873273761175e-06, + "loss": 2.1282, + "step": 1340 + }, + { + "epoch": 1.4684931506849315, + "eval_loss": 2.275458574295044, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 1340 + }, + { + "epoch": 1.4695890410958903, + "grad_norm": 1.6038731336593628, + "learning_rate": 2.8655564581640944e-06, + "loss": 2.1164, + "step": 1341 + }, + { + "epoch": 1.4706849315068493, + "grad_norm": 1.6492732763290405, + "learning_rate": 2.863525588952072e-06, + "loss": 2.1471, + "step": 1342 + }, + { + "epoch": 1.4717808219178083, + "grad_norm": 1.647605538368225, + "learning_rate": 2.861494719740049e-06, + "loss": 2.0737, + "step": 1343 + }, + { + "epoch": 1.4728767123287672, + "grad_norm": 1.5875589847564697, + "learning_rate": 2.8594638505280266e-06, + "loss": 2.0784, + "step": 1344 + }, + { + "epoch": 1.473972602739726, + "grad_norm": 1.660346508026123, + "learning_rate": 2.8574329813160035e-06, + "loss": 2.1166, + "step": 1345 + }, + { + "epoch": 1.475068493150685, + "grad_norm": 1.5677483081817627, + "learning_rate": 2.855402112103981e-06, + "loss": 2.1068, + "step": 1346 + }, + { + "epoch": 1.4761643835616438, + "grad_norm": 1.6814773082733154, + "learning_rate": 2.853371242891958e-06, + "loss": 2.1118, + "step": 1347 + }, + { + "epoch": 1.4772602739726026, + "grad_norm": 1.7251765727996826, + "learning_rate": 2.851340373679935e-06, + "loss": 2.1255, + "step": 1348 + }, + { + "epoch": 1.4783561643835617, + "grad_norm": 1.6479768753051758, + "learning_rate": 2.8493095044679126e-06, + "loss": 2.1675, + "step": 1349 + }, + { + "epoch": 1.4794520547945205, + "grad_norm": 1.5761340856552124, + "learning_rate": 2.8472786352558894e-06, + "loss": 2.1271, + "step": 1350 + }, + { + "epoch": 1.4794520547945205, + "eval_loss": 2.2767348289489746, + "eval_runtime": 37.7124, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 1350 + }, + { + "epoch": 1.4805479452054795, + "grad_norm": 1.577124834060669, + "learning_rate": 2.845247766043867e-06, + "loss": 2.0968, + "step": 1351 + }, + { + "epoch": 1.4816438356164383, + "grad_norm": 1.6901545524597168, + "learning_rate": 2.843216896831844e-06, + "loss": 2.092, + "step": 1352 + }, + { + "epoch": 1.4827397260273973, + "grad_norm": 1.582205891609192, + "learning_rate": 2.8411860276198217e-06, + "loss": 2.1067, + "step": 1353 + }, + { + "epoch": 1.4838356164383562, + "grad_norm": 1.6398444175720215, + "learning_rate": 2.8391551584077985e-06, + "loss": 2.1524, + "step": 1354 + }, + { + "epoch": 1.484931506849315, + "grad_norm": 1.6396856307983398, + "learning_rate": 2.8371242891957762e-06, + "loss": 2.1523, + "step": 1355 + }, + { + "epoch": 1.486027397260274, + "grad_norm": 1.7187350988388062, + "learning_rate": 2.835093419983753e-06, + "loss": 2.1186, + "step": 1356 + }, + { + "epoch": 1.4871232876712328, + "grad_norm": 1.6155378818511963, + "learning_rate": 2.833062550771731e-06, + "loss": 2.1119, + "step": 1357 + }, + { + "epoch": 1.4882191780821918, + "grad_norm": 1.6749935150146484, + "learning_rate": 2.8310316815597076e-06, + "loss": 2.0871, + "step": 1358 + }, + { + "epoch": 1.4893150684931507, + "grad_norm": 1.7644442319869995, + "learning_rate": 2.8290008123476853e-06, + "loss": 2.1296, + "step": 1359 + }, + { + "epoch": 1.4904109589041097, + "grad_norm": 1.7644187211990356, + "learning_rate": 2.826969943135662e-06, + "loss": 2.1168, + "step": 1360 + }, + { + "epoch": 1.4904109589041097, + "eval_loss": 2.275988817214966, + "eval_runtime": 37.7569, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1360 + }, + { + "epoch": 1.4915068493150685, + "grad_norm": 1.6250940561294556, + "learning_rate": 2.8249390739236395e-06, + "loss": 2.1126, + "step": 1361 + }, + { + "epoch": 1.4926027397260273, + "grad_norm": 1.545547366142273, + "learning_rate": 2.8229082047116167e-06, + "loss": 2.1014, + "step": 1362 + }, + { + "epoch": 1.4936986301369863, + "grad_norm": 1.6356806755065918, + "learning_rate": 2.820877335499594e-06, + "loss": 2.0994, + "step": 1363 + }, + { + "epoch": 1.4947945205479451, + "grad_norm": 1.6093254089355469, + "learning_rate": 2.8188464662875713e-06, + "loss": 2.0638, + "step": 1364 + }, + { + "epoch": 1.4958904109589042, + "grad_norm": 1.531909465789795, + "learning_rate": 2.8168155970755486e-06, + "loss": 2.0873, + "step": 1365 + }, + { + "epoch": 1.496986301369863, + "grad_norm": 1.7620573043823242, + "learning_rate": 2.814784727863526e-06, + "loss": 2.1124, + "step": 1366 + }, + { + "epoch": 1.498082191780822, + "grad_norm": 1.7096633911132812, + "learning_rate": 2.812753858651503e-06, + "loss": 2.1308, + "step": 1367 + }, + { + "epoch": 1.4991780821917808, + "grad_norm": 1.8180763721466064, + "learning_rate": 2.8107229894394804e-06, + "loss": 2.0961, + "step": 1368 + }, + { + "epoch": 1.5002739726027396, + "grad_norm": 1.7475993633270264, + "learning_rate": 2.8086921202274577e-06, + "loss": 2.0784, + "step": 1369 + }, + { + "epoch": 1.5013698630136987, + "grad_norm": 1.7892746925354004, + "learning_rate": 2.806661251015435e-06, + "loss": 2.1298, + "step": 1370 + }, + { + "epoch": 1.5013698630136987, + "eval_loss": 2.2772014141082764, + "eval_runtime": 37.709, + "eval_samples_per_second": 26.519, + "eval_steps_per_second": 1.114, + "step": 1370 + }, + { + "epoch": 1.5024657534246575, + "grad_norm": 1.700134038925171, + "learning_rate": 2.8046303818034122e-06, + "loss": 2.1571, + "step": 1371 + }, + { + "epoch": 1.5035616438356163, + "grad_norm": 1.7968006134033203, + "learning_rate": 2.802599512591389e-06, + "loss": 2.2039, + "step": 1372 + }, + { + "epoch": 1.5046575342465753, + "grad_norm": 1.5552955865859985, + "learning_rate": 2.8005686433793668e-06, + "loss": 2.1078, + "step": 1373 + }, + { + "epoch": 1.5057534246575344, + "grad_norm": 1.6243152618408203, + "learning_rate": 2.7985377741673436e-06, + "loss": 2.071, + "step": 1374 + }, + { + "epoch": 1.5068493150684932, + "grad_norm": 1.7453999519348145, + "learning_rate": 2.7965069049553213e-06, + "loss": 2.1228, + "step": 1375 + }, + { + "epoch": 1.507945205479452, + "grad_norm": 1.6567325592041016, + "learning_rate": 2.794476035743298e-06, + "loss": 2.158, + "step": 1376 + }, + { + "epoch": 1.509041095890411, + "grad_norm": 1.6160842180252075, + "learning_rate": 2.792445166531276e-06, + "loss": 2.1231, + "step": 1377 + }, + { + "epoch": 1.5101369863013698, + "grad_norm": 1.715888500213623, + "learning_rate": 2.7904142973192527e-06, + "loss": 2.1466, + "step": 1378 + }, + { + "epoch": 1.5112328767123286, + "grad_norm": 1.6263453960418701, + "learning_rate": 2.7883834281072304e-06, + "loss": 2.0926, + "step": 1379 + }, + { + "epoch": 1.5123287671232877, + "grad_norm": 1.7616580724716187, + "learning_rate": 2.7863525588952073e-06, + "loss": 2.0877, + "step": 1380 + }, + { + "epoch": 1.5123287671232877, + "eval_loss": 2.275289535522461, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 1380 + }, + { + "epoch": 1.5134246575342467, + "grad_norm": 1.6861932277679443, + "learning_rate": 2.784321689683185e-06, + "loss": 2.0867, + "step": 1381 + }, + { + "epoch": 1.5145205479452055, + "grad_norm": 1.570162296295166, + "learning_rate": 2.782290820471162e-06, + "loss": 2.0957, + "step": 1382 + }, + { + "epoch": 1.5156164383561643, + "grad_norm": 1.698014259338379, + "learning_rate": 2.7802599512591395e-06, + "loss": 2.1427, + "step": 1383 + }, + { + "epoch": 1.5167123287671234, + "grad_norm": 1.6773440837860107, + "learning_rate": 2.7782290820471164e-06, + "loss": 2.135, + "step": 1384 + }, + { + "epoch": 1.5178082191780822, + "grad_norm": 1.6762758493423462, + "learning_rate": 2.7761982128350932e-06, + "loss": 2.132, + "step": 1385 + }, + { + "epoch": 1.518904109589041, + "grad_norm": 1.5917284488677979, + "learning_rate": 2.774167343623071e-06, + "loss": 2.1616, + "step": 1386 + }, + { + "epoch": 1.52, + "grad_norm": 1.583791732788086, + "learning_rate": 2.772136474411048e-06, + "loss": 2.0901, + "step": 1387 + }, + { + "epoch": 1.521095890410959, + "grad_norm": 1.6780498027801514, + "learning_rate": 2.7701056051990255e-06, + "loss": 2.0577, + "step": 1388 + }, + { + "epoch": 1.5221917808219179, + "grad_norm": 1.6320325136184692, + "learning_rate": 2.7680747359870023e-06, + "loss": 2.1381, + "step": 1389 + }, + { + "epoch": 1.5232876712328767, + "grad_norm": 1.7060813903808594, + "learning_rate": 2.76604386677498e-06, + "loss": 2.115, + "step": 1390 + }, + { + "epoch": 1.5232876712328767, + "eval_loss": 2.27346134185791, + "eval_runtime": 37.778, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 1390 + }, + { + "epoch": 1.5243835616438357, + "grad_norm": 1.6142189502716064, + "learning_rate": 2.764012997562957e-06, + "loss": 2.1388, + "step": 1391 + }, + { + "epoch": 1.5254794520547945, + "grad_norm": 1.6605664491653442, + "learning_rate": 2.7619821283509346e-06, + "loss": 2.0784, + "step": 1392 + }, + { + "epoch": 1.5265753424657533, + "grad_norm": 1.7281928062438965, + "learning_rate": 2.7599512591389114e-06, + "loss": 2.1461, + "step": 1393 + }, + { + "epoch": 1.5276712328767124, + "grad_norm": 1.6661336421966553, + "learning_rate": 2.757920389926889e-06, + "loss": 2.0987, + "step": 1394 + }, + { + "epoch": 1.5287671232876714, + "grad_norm": 1.6454191207885742, + "learning_rate": 2.755889520714866e-06, + "loss": 2.0912, + "step": 1395 + }, + { + "epoch": 1.52986301369863, + "grad_norm": 1.6793323755264282, + "learning_rate": 2.7538586515028437e-06, + "loss": 2.0642, + "step": 1396 + }, + { + "epoch": 1.530958904109589, + "grad_norm": 1.5941072702407837, + "learning_rate": 2.7518277822908205e-06, + "loss": 2.1232, + "step": 1397 + }, + { + "epoch": 1.532054794520548, + "grad_norm": 1.5669342279434204, + "learning_rate": 2.749796913078798e-06, + "loss": 2.1406, + "step": 1398 + }, + { + "epoch": 1.5331506849315069, + "grad_norm": 1.5805522203445435, + "learning_rate": 2.747766043866775e-06, + "loss": 2.1096, + "step": 1399 + }, + { + "epoch": 1.5342465753424657, + "grad_norm": 1.6227574348449707, + "learning_rate": 2.7457351746547524e-06, + "loss": 2.15, + "step": 1400 + }, + { + "epoch": 1.5342465753424657, + "eval_loss": 2.27066707611084, + "eval_runtime": 37.708, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 1400 + }, + { + "epoch": 1.5353424657534247, + "grad_norm": 1.6500868797302246, + "learning_rate": 2.7437043054427297e-06, + "loss": 2.193, + "step": 1401 + }, + { + "epoch": 1.5364383561643835, + "grad_norm": 1.6337374448776245, + "learning_rate": 2.741673436230707e-06, + "loss": 2.1196, + "step": 1402 + }, + { + "epoch": 1.5375342465753423, + "grad_norm": 1.5784432888031006, + "learning_rate": 2.739642567018684e-06, + "loss": 2.0906, + "step": 1403 + }, + { + "epoch": 1.5386301369863014, + "grad_norm": 1.75803542137146, + "learning_rate": 2.7376116978066615e-06, + "loss": 2.124, + "step": 1404 + }, + { + "epoch": 1.5397260273972604, + "grad_norm": 1.7682627439498901, + "learning_rate": 2.7355808285946388e-06, + "loss": 2.121, + "step": 1405 + }, + { + "epoch": 1.5408219178082192, + "grad_norm": 1.7324312925338745, + "learning_rate": 2.733549959382616e-06, + "loss": 2.0975, + "step": 1406 + }, + { + "epoch": 1.541917808219178, + "grad_norm": 1.765071153640747, + "learning_rate": 2.7315190901705933e-06, + "loss": 2.1668, + "step": 1407 + }, + { + "epoch": 1.543013698630137, + "grad_norm": 1.7210100889205933, + "learning_rate": 2.7294882209585706e-06, + "loss": 2.1284, + "step": 1408 + }, + { + "epoch": 1.5441095890410959, + "grad_norm": 1.7064242362976074, + "learning_rate": 2.727457351746548e-06, + "loss": 2.1007, + "step": 1409 + }, + { + "epoch": 1.5452054794520547, + "grad_norm": 1.7347673177719116, + "learning_rate": 2.725426482534525e-06, + "loss": 2.0908, + "step": 1410 + }, + { + "epoch": 1.5452054794520547, + "eval_loss": 2.273494005203247, + "eval_runtime": 37.5003, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 1410 + }, + { + "epoch": 1.5463013698630137, + "grad_norm": 1.6207059621810913, + "learning_rate": 2.723395613322502e-06, + "loss": 2.1238, + "step": 1411 + }, + { + "epoch": 1.5473972602739727, + "grad_norm": 1.6258277893066406, + "learning_rate": 2.7213647441104797e-06, + "loss": 2.107, + "step": 1412 + }, + { + "epoch": 1.5484931506849315, + "grad_norm": 1.8286561965942383, + "learning_rate": 2.7193338748984565e-06, + "loss": 2.1198, + "step": 1413 + }, + { + "epoch": 1.5495890410958904, + "grad_norm": 1.6007434129714966, + "learning_rate": 2.7173030056864342e-06, + "loss": 2.2353, + "step": 1414 + }, + { + "epoch": 1.5506849315068494, + "grad_norm": 1.665379285812378, + "learning_rate": 2.715272136474411e-06, + "loss": 2.0716, + "step": 1415 + }, + { + "epoch": 1.5517808219178082, + "grad_norm": 1.6235698461532593, + "learning_rate": 2.713241267262389e-06, + "loss": 2.1239, + "step": 1416 + }, + { + "epoch": 1.552876712328767, + "grad_norm": 1.6042425632476807, + "learning_rate": 2.7112103980503656e-06, + "loss": 2.1033, + "step": 1417 + }, + { + "epoch": 1.553972602739726, + "grad_norm": 1.6660529375076294, + "learning_rate": 2.7091795288383433e-06, + "loss": 2.1326, + "step": 1418 + }, + { + "epoch": 1.555068493150685, + "grad_norm": 1.7682642936706543, + "learning_rate": 2.70714865962632e-06, + "loss": 2.0977, + "step": 1419 + }, + { + "epoch": 1.5561643835616439, + "grad_norm": 1.633492112159729, + "learning_rate": 2.705117790414298e-06, + "loss": 2.1078, + "step": 1420 + }, + { + "epoch": 1.5561643835616439, + "eval_loss": 2.2733564376831055, + "eval_runtime": 37.7364, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 1420 + }, + { + "epoch": 1.5572602739726027, + "grad_norm": 1.6661697626113892, + "learning_rate": 2.7030869212022747e-06, + "loss": 2.1395, + "step": 1421 + }, + { + "epoch": 1.5583561643835617, + "grad_norm": 1.678194284439087, + "learning_rate": 2.7010560519902516e-06, + "loss": 2.101, + "step": 1422 + }, + { + "epoch": 1.5594520547945205, + "grad_norm": 1.686339020729065, + "learning_rate": 2.6990251827782293e-06, + "loss": 2.1117, + "step": 1423 + }, + { + "epoch": 1.5605479452054793, + "grad_norm": 1.5932446718215942, + "learning_rate": 2.696994313566206e-06, + "loss": 2.1075, + "step": 1424 + }, + { + "epoch": 1.5616438356164384, + "grad_norm": 1.744378685951233, + "learning_rate": 2.694963444354184e-06, + "loss": 2.1658, + "step": 1425 + }, + { + "epoch": 1.5627397260273974, + "grad_norm": 1.5379083156585693, + "learning_rate": 2.6929325751421607e-06, + "loss": 2.1027, + "step": 1426 + }, + { + "epoch": 1.563835616438356, + "grad_norm": 1.582470417022705, + "learning_rate": 2.6909017059301384e-06, + "loss": 2.1343, + "step": 1427 + }, + { + "epoch": 1.564931506849315, + "grad_norm": 1.6151307821273804, + "learning_rate": 2.6888708367181153e-06, + "loss": 2.0837, + "step": 1428 + }, + { + "epoch": 1.566027397260274, + "grad_norm": 1.7137283086776733, + "learning_rate": 2.686839967506093e-06, + "loss": 2.135, + "step": 1429 + }, + { + "epoch": 1.5671232876712329, + "grad_norm": 1.6230741739273071, + "learning_rate": 2.68480909829407e-06, + "loss": 2.0593, + "step": 1430 + }, + { + "epoch": 1.5671232876712329, + "eval_loss": 2.2720754146575928, + "eval_runtime": 37.8152, + "eval_samples_per_second": 26.444, + "eval_steps_per_second": 1.111, + "step": 1430 + }, + { + "epoch": 1.5682191780821917, + "grad_norm": 1.638656497001648, + "learning_rate": 2.6827782290820475e-06, + "loss": 2.0652, + "step": 1431 + }, + { + "epoch": 1.5693150684931507, + "grad_norm": 1.565036654472351, + "learning_rate": 2.6807473598700244e-06, + "loss": 2.154, + "step": 1432 + }, + { + "epoch": 1.5704109589041098, + "grad_norm": 1.624617576599121, + "learning_rate": 2.678716490658002e-06, + "loss": 2.1264, + "step": 1433 + }, + { + "epoch": 1.5715068493150683, + "grad_norm": 1.5961869955062866, + "learning_rate": 2.676685621445979e-06, + "loss": 2.1189, + "step": 1434 + }, + { + "epoch": 1.5726027397260274, + "grad_norm": 1.5941061973571777, + "learning_rate": 2.674654752233956e-06, + "loss": 2.1229, + "step": 1435 + }, + { + "epoch": 1.5736986301369864, + "grad_norm": 1.5900241136550903, + "learning_rate": 2.6726238830219335e-06, + "loss": 2.1174, + "step": 1436 + }, + { + "epoch": 1.5747945205479452, + "grad_norm": 1.696982979774475, + "learning_rate": 2.6705930138099107e-06, + "loss": 2.1463, + "step": 1437 + }, + { + "epoch": 1.575890410958904, + "grad_norm": 1.5432531833648682, + "learning_rate": 2.668562144597888e-06, + "loss": 2.0515, + "step": 1438 + }, + { + "epoch": 1.576986301369863, + "grad_norm": 1.6274412870407104, + "learning_rate": 2.6665312753858653e-06, + "loss": 2.1281, + "step": 1439 + }, + { + "epoch": 1.5780821917808219, + "grad_norm": 1.5978962182998657, + "learning_rate": 2.6645004061738426e-06, + "loss": 2.123, + "step": 1440 + }, + { + "epoch": 1.5780821917808219, + "eval_loss": 2.2747466564178467, + "eval_runtime": 37.7905, + "eval_samples_per_second": 26.462, + "eval_steps_per_second": 1.111, + "step": 1440 + }, + { + "epoch": 1.5791780821917807, + "grad_norm": 1.684459924697876, + "learning_rate": 2.66246953696182e-06, + "loss": 2.09, + "step": 1441 + }, + { + "epoch": 1.5802739726027397, + "grad_norm": 1.6446969509124756, + "learning_rate": 2.660438667749797e-06, + "loss": 2.0723, + "step": 1442 + }, + { + "epoch": 1.5813698630136988, + "grad_norm": 1.6567888259887695, + "learning_rate": 2.6584077985377744e-06, + "loss": 2.1052, + "step": 1443 + }, + { + "epoch": 1.5824657534246576, + "grad_norm": 1.6030794382095337, + "learning_rate": 2.6563769293257517e-06, + "loss": 2.11, + "step": 1444 + }, + { + "epoch": 1.5835616438356164, + "grad_norm": 1.5979795455932617, + "learning_rate": 2.654346060113729e-06, + "loss": 2.0903, + "step": 1445 + }, + { + "epoch": 1.5846575342465754, + "grad_norm": 1.5626153945922852, + "learning_rate": 2.6523151909017062e-06, + "loss": 2.0998, + "step": 1446 + }, + { + "epoch": 1.5857534246575342, + "grad_norm": 1.6497416496276855, + "learning_rate": 2.6502843216896835e-06, + "loss": 2.0863, + "step": 1447 + }, + { + "epoch": 1.586849315068493, + "grad_norm": 1.6650139093399048, + "learning_rate": 2.6482534524776603e-06, + "loss": 2.1615, + "step": 1448 + }, + { + "epoch": 1.587945205479452, + "grad_norm": 1.6197447776794434, + "learning_rate": 2.646222583265638e-06, + "loss": 2.1374, + "step": 1449 + }, + { + "epoch": 1.589041095890411, + "grad_norm": 1.6509665250778198, + "learning_rate": 2.644191714053615e-06, + "loss": 2.1242, + "step": 1450 + }, + { + "epoch": 1.589041095890411, + "eval_loss": 2.272801160812378, + "eval_runtime": 37.769, + "eval_samples_per_second": 26.477, + "eval_steps_per_second": 1.112, + "step": 1450 + }, + { + "epoch": 1.59013698630137, + "grad_norm": 1.570002555847168, + "learning_rate": 2.6421608448415926e-06, + "loss": 2.0899, + "step": 1451 + }, + { + "epoch": 1.5912328767123287, + "grad_norm": 1.653418779373169, + "learning_rate": 2.6401299756295694e-06, + "loss": 2.0823, + "step": 1452 + }, + { + "epoch": 1.5923287671232877, + "grad_norm": 1.7645854949951172, + "learning_rate": 2.638099106417547e-06, + "loss": 2.1162, + "step": 1453 + }, + { + "epoch": 1.5934246575342466, + "grad_norm": 1.6598302125930786, + "learning_rate": 2.636068237205524e-06, + "loss": 2.0887, + "step": 1454 + }, + { + "epoch": 1.5945205479452054, + "grad_norm": 1.7615699768066406, + "learning_rate": 2.6340373679935017e-06, + "loss": 2.153, + "step": 1455 + }, + { + "epoch": 1.5956164383561644, + "grad_norm": 1.5611566305160522, + "learning_rate": 2.6320064987814785e-06, + "loss": 2.0635, + "step": 1456 + }, + { + "epoch": 1.5967123287671234, + "grad_norm": 1.7776182889938354, + "learning_rate": 2.6299756295694562e-06, + "loss": 2.1248, + "step": 1457 + }, + { + "epoch": 1.5978082191780822, + "grad_norm": 1.563895344734192, + "learning_rate": 2.627944760357433e-06, + "loss": 2.1591, + "step": 1458 + }, + { + "epoch": 1.598904109589041, + "grad_norm": 1.5638104677200317, + "learning_rate": 2.6259138911454104e-06, + "loss": 2.1594, + "step": 1459 + }, + { + "epoch": 1.6, + "grad_norm": 1.6918673515319824, + "learning_rate": 2.6238830219333877e-06, + "loss": 2.143, + "step": 1460 + }, + { + "epoch": 1.6, + "eval_loss": 2.27238130569458, + "eval_runtime": 37.7964, + "eval_samples_per_second": 26.458, + "eval_steps_per_second": 1.111, + "step": 1460 + }, + { + "epoch": 1.601095890410959, + "grad_norm": 1.6937850713729858, + "learning_rate": 2.621852152721365e-06, + "loss": 2.0692, + "step": 1461 + }, + { + "epoch": 1.6021917808219177, + "grad_norm": 1.6500011682510376, + "learning_rate": 2.619821283509342e-06, + "loss": 2.1432, + "step": 1462 + }, + { + "epoch": 1.6032876712328767, + "grad_norm": 1.583245038986206, + "learning_rate": 2.6177904142973195e-06, + "loss": 2.0811, + "step": 1463 + }, + { + "epoch": 1.6043835616438358, + "grad_norm": 1.7283663749694824, + "learning_rate": 2.6157595450852968e-06, + "loss": 2.1268, + "step": 1464 + }, + { + "epoch": 1.6054794520547944, + "grad_norm": 1.6247549057006836, + "learning_rate": 2.613728675873274e-06, + "loss": 2.1464, + "step": 1465 + }, + { + "epoch": 1.6065753424657534, + "grad_norm": 1.6460528373718262, + "learning_rate": 2.6116978066612513e-06, + "loss": 2.1428, + "step": 1466 + }, + { + "epoch": 1.6076712328767124, + "grad_norm": 1.6332457065582275, + "learning_rate": 2.6096669374492286e-06, + "loss": 2.1216, + "step": 1467 + }, + { + "epoch": 1.6087671232876712, + "grad_norm": 1.7202653884887695, + "learning_rate": 2.607636068237206e-06, + "loss": 2.0877, + "step": 1468 + }, + { + "epoch": 1.60986301369863, + "grad_norm": 1.6361600160598755, + "learning_rate": 2.605605199025183e-06, + "loss": 2.129, + "step": 1469 + }, + { + "epoch": 1.610958904109589, + "grad_norm": 1.6851226091384888, + "learning_rate": 2.6035743298131604e-06, + "loss": 2.1181, + "step": 1470 + }, + { + "epoch": 1.610958904109589, + "eval_loss": 2.276170492172241, + "eval_runtime": 37.7741, + "eval_samples_per_second": 26.473, + "eval_steps_per_second": 1.112, + "step": 1470 + }, + { + "epoch": 1.612054794520548, + "grad_norm": 1.7388262748718262, + "learning_rate": 2.6015434606011377e-06, + "loss": 2.137, + "step": 1471 + }, + { + "epoch": 1.6131506849315067, + "grad_norm": 1.8428372144699097, + "learning_rate": 2.5995125913891145e-06, + "loss": 2.1324, + "step": 1472 + }, + { + "epoch": 1.6142465753424657, + "grad_norm": 1.6778063774108887, + "learning_rate": 2.5974817221770922e-06, + "loss": 2.1207, + "step": 1473 + }, + { + "epoch": 1.6153424657534248, + "grad_norm": 1.5915114879608154, + "learning_rate": 2.595450852965069e-06, + "loss": 2.1297, + "step": 1474 + }, + { + "epoch": 1.6164383561643836, + "grad_norm": 1.6911993026733398, + "learning_rate": 2.593419983753047e-06, + "loss": 2.1574, + "step": 1475 + }, + { + "epoch": 1.6175342465753424, + "grad_norm": 1.568377137184143, + "learning_rate": 2.5913891145410236e-06, + "loss": 2.1046, + "step": 1476 + }, + { + "epoch": 1.6186301369863014, + "grad_norm": 1.6314934492111206, + "learning_rate": 2.5893582453290013e-06, + "loss": 2.145, + "step": 1477 + }, + { + "epoch": 1.6197260273972602, + "grad_norm": 1.6405373811721802, + "learning_rate": 2.587327376116978e-06, + "loss": 2.1017, + "step": 1478 + }, + { + "epoch": 1.620821917808219, + "grad_norm": 1.5760985612869263, + "learning_rate": 2.585296506904956e-06, + "loss": 2.1106, + "step": 1479 + }, + { + "epoch": 1.621917808219178, + "grad_norm": 1.667673945426941, + "learning_rate": 2.5832656376929327e-06, + "loss": 2.0898, + "step": 1480 + }, + { + "epoch": 1.621917808219178, + "eval_loss": 2.272311210632324, + "eval_runtime": 37.6893, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 1480 + }, + { + "epoch": 1.6230136986301371, + "grad_norm": 1.6694347858428955, + "learning_rate": 2.5812347684809104e-06, + "loss": 2.0515, + "step": 1481 + }, + { + "epoch": 1.624109589041096, + "grad_norm": 1.6225323677062988, + "learning_rate": 2.5792038992688873e-06, + "loss": 2.1824, + "step": 1482 + }, + { + "epoch": 1.6252054794520547, + "grad_norm": 1.7553166151046753, + "learning_rate": 2.577173030056865e-06, + "loss": 2.0484, + "step": 1483 + }, + { + "epoch": 1.6263013698630138, + "grad_norm": 1.6127912998199463, + "learning_rate": 2.575142160844842e-06, + "loss": 2.1306, + "step": 1484 + }, + { + "epoch": 1.6273972602739726, + "grad_norm": 1.728456974029541, + "learning_rate": 2.5731112916328187e-06, + "loss": 2.1371, + "step": 1485 + }, + { + "epoch": 1.6284931506849314, + "grad_norm": 1.655109167098999, + "learning_rate": 2.5710804224207964e-06, + "loss": 2.0918, + "step": 1486 + }, + { + "epoch": 1.6295890410958904, + "grad_norm": 1.7091511487960815, + "learning_rate": 2.5690495532087733e-06, + "loss": 2.0534, + "step": 1487 + }, + { + "epoch": 1.6306849315068495, + "grad_norm": 1.663798213005066, + "learning_rate": 2.567018683996751e-06, + "loss": 2.1262, + "step": 1488 + }, + { + "epoch": 1.6317808219178083, + "grad_norm": 1.7468243837356567, + "learning_rate": 2.564987814784728e-06, + "loss": 2.088, + "step": 1489 + }, + { + "epoch": 1.632876712328767, + "grad_norm": 1.662075400352478, + "learning_rate": 2.5629569455727055e-06, + "loss": 2.1063, + "step": 1490 + }, + { + "epoch": 1.632876712328767, + "eval_loss": 2.268484354019165, + "eval_runtime": 37.698, + "eval_samples_per_second": 26.527, + "eval_steps_per_second": 1.114, + "step": 1490 + }, + { + "epoch": 1.633972602739726, + "grad_norm": 1.6704449653625488, + "learning_rate": 2.5609260763606824e-06, + "loss": 2.0996, + "step": 1491 + }, + { + "epoch": 1.635068493150685, + "grad_norm": 1.622410774230957, + "learning_rate": 2.55889520714866e-06, + "loss": 2.1792, + "step": 1492 + }, + { + "epoch": 1.6361643835616437, + "grad_norm": 1.6591641902923584, + "learning_rate": 2.556864337936637e-06, + "loss": 2.0999, + "step": 1493 + }, + { + "epoch": 1.6372602739726028, + "grad_norm": 1.7249618768692017, + "learning_rate": 2.5548334687246146e-06, + "loss": 2.1152, + "step": 1494 + }, + { + "epoch": 1.6383561643835618, + "grad_norm": 1.8214666843414307, + "learning_rate": 2.5528025995125915e-06, + "loss": 2.1512, + "step": 1495 + }, + { + "epoch": 1.6394520547945204, + "grad_norm": 1.6735785007476807, + "learning_rate": 2.550771730300569e-06, + "loss": 2.1375, + "step": 1496 + }, + { + "epoch": 1.6405479452054794, + "grad_norm": 1.8190255165100098, + "learning_rate": 2.548740861088546e-06, + "loss": 2.1425, + "step": 1497 + }, + { + "epoch": 1.6416438356164385, + "grad_norm": 1.6991790533065796, + "learning_rate": 2.5467099918765233e-06, + "loss": 2.0658, + "step": 1498 + }, + { + "epoch": 1.6427397260273973, + "grad_norm": 1.6228128671646118, + "learning_rate": 2.5446791226645006e-06, + "loss": 2.0505, + "step": 1499 + }, + { + "epoch": 1.643835616438356, + "grad_norm": 1.9889343976974487, + "learning_rate": 2.542648253452478e-06, + "loss": 2.1374, + "step": 1500 + }, + { + "epoch": 1.643835616438356, + "eval_loss": 2.2690248489379883, + "eval_runtime": 37.6989, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 1500 + }, + { + "epoch": 1.644931506849315, + "grad_norm": 1.6495356559753418, + "learning_rate": 2.540617384240455e-06, + "loss": 2.1171, + "step": 1501 + }, + { + "epoch": 1.6460273972602741, + "grad_norm": 1.7420153617858887, + "learning_rate": 2.5385865150284324e-06, + "loss": 2.092, + "step": 1502 + }, + { + "epoch": 1.6471232876712327, + "grad_norm": 1.7456740140914917, + "learning_rate": 2.5365556458164097e-06, + "loss": 2.1029, + "step": 1503 + }, + { + "epoch": 1.6482191780821918, + "grad_norm": 1.607245922088623, + "learning_rate": 2.534524776604387e-06, + "loss": 2.1173, + "step": 1504 + }, + { + "epoch": 1.6493150684931508, + "grad_norm": 1.648555040359497, + "learning_rate": 2.5324939073923642e-06, + "loss": 2.0275, + "step": 1505 + }, + { + "epoch": 1.6504109589041096, + "grad_norm": 1.753891944885254, + "learning_rate": 2.5304630381803415e-06, + "loss": 2.103, + "step": 1506 + }, + { + "epoch": 1.6515068493150684, + "grad_norm": 1.6390420198440552, + "learning_rate": 2.5284321689683188e-06, + "loss": 2.111, + "step": 1507 + }, + { + "epoch": 1.6526027397260274, + "grad_norm": 1.6427407264709473, + "learning_rate": 2.526401299756296e-06, + "loss": 2.0834, + "step": 1508 + }, + { + "epoch": 1.6536986301369863, + "grad_norm": 1.7193669080734253, + "learning_rate": 2.524370430544273e-06, + "loss": 2.1125, + "step": 1509 + }, + { + "epoch": 1.654794520547945, + "grad_norm": 1.6364102363586426, + "learning_rate": 2.5223395613322506e-06, + "loss": 2.0839, + "step": 1510 + }, + { + "epoch": 1.654794520547945, + "eval_loss": 2.271984577178955, + "eval_runtime": 37.4343, + "eval_samples_per_second": 26.713, + "eval_steps_per_second": 1.122, + "step": 1510 + }, + { + "epoch": 1.655890410958904, + "grad_norm": 1.6273912191390991, + "learning_rate": 2.5203086921202274e-06, + "loss": 2.1296, + "step": 1511 + }, + { + "epoch": 1.6569863013698631, + "grad_norm": 1.631685733795166, + "learning_rate": 2.518277822908205e-06, + "loss": 2.0956, + "step": 1512 + }, + { + "epoch": 1.658082191780822, + "grad_norm": 1.668209195137024, + "learning_rate": 2.516246953696182e-06, + "loss": 2.1216, + "step": 1513 + }, + { + "epoch": 1.6591780821917808, + "grad_norm": 1.7267968654632568, + "learning_rate": 2.5142160844841597e-06, + "loss": 2.1421, + "step": 1514 + }, + { + "epoch": 1.6602739726027398, + "grad_norm": 1.730349063873291, + "learning_rate": 2.5121852152721365e-06, + "loss": 2.1286, + "step": 1515 + }, + { + "epoch": 1.6613698630136986, + "grad_norm": 1.6914033889770508, + "learning_rate": 2.5101543460601142e-06, + "loss": 2.123, + "step": 1516 + }, + { + "epoch": 1.6624657534246574, + "grad_norm": 1.6512525081634521, + "learning_rate": 2.508123476848091e-06, + "loss": 2.0714, + "step": 1517 + }, + { + "epoch": 1.6635616438356164, + "grad_norm": 1.7400277853012085, + "learning_rate": 2.506092607636069e-06, + "loss": 2.1508, + "step": 1518 + }, + { + "epoch": 1.6646575342465755, + "grad_norm": 1.6652593612670898, + "learning_rate": 2.5040617384240457e-06, + "loss": 2.0897, + "step": 1519 + }, + { + "epoch": 1.6657534246575343, + "grad_norm": 1.8187719583511353, + "learning_rate": 2.5020308692120234e-06, + "loss": 2.0946, + "step": 1520 + }, + { + "epoch": 1.6657534246575343, + "eval_loss": 2.2724390029907227, + "eval_runtime": 37.6493, + "eval_samples_per_second": 26.561, + "eval_steps_per_second": 1.116, + "step": 1520 + }, + { + "epoch": 1.666849315068493, + "grad_norm": 1.6140000820159912, + "learning_rate": 2.5e-06, + "loss": 2.1534, + "step": 1521 + }, + { + "epoch": 1.6679452054794521, + "grad_norm": 1.7010711431503296, + "learning_rate": 2.4979691307879775e-06, + "loss": 2.1028, + "step": 1522 + }, + { + "epoch": 1.669041095890411, + "grad_norm": 1.661583423614502, + "learning_rate": 2.4959382615759548e-06, + "loss": 2.0763, + "step": 1523 + }, + { + "epoch": 1.6701369863013698, + "grad_norm": 1.6757241487503052, + "learning_rate": 2.493907392363932e-06, + "loss": 2.1602, + "step": 1524 + }, + { + "epoch": 1.6712328767123288, + "grad_norm": 1.5544378757476807, + "learning_rate": 2.4918765231519093e-06, + "loss": 2.1251, + "step": 1525 + }, + { + "epoch": 1.6723287671232878, + "grad_norm": 1.7000782489776611, + "learning_rate": 2.4898456539398866e-06, + "loss": 2.1044, + "step": 1526 + }, + { + "epoch": 1.6734246575342466, + "grad_norm": 1.584068775177002, + "learning_rate": 2.487814784727864e-06, + "loss": 2.166, + "step": 1527 + }, + { + "epoch": 1.6745205479452054, + "grad_norm": 1.6422476768493652, + "learning_rate": 2.4857839155158407e-06, + "loss": 2.1079, + "step": 1528 + }, + { + "epoch": 1.6756164383561645, + "grad_norm": 1.621401071548462, + "learning_rate": 2.483753046303818e-06, + "loss": 2.107, + "step": 1529 + }, + { + "epoch": 1.6767123287671233, + "grad_norm": 1.6382123231887817, + "learning_rate": 2.4817221770917953e-06, + "loss": 2.1161, + "step": 1530 + }, + { + "epoch": 1.6767123287671233, + "eval_loss": 2.2733490467071533, + "eval_runtime": 37.713, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 1530 + }, + { + "epoch": 1.677808219178082, + "grad_norm": 1.7815386056900024, + "learning_rate": 2.4796913078797725e-06, + "loss": 2.138, + "step": 1531 + }, + { + "epoch": 1.6789041095890411, + "grad_norm": 1.7218440771102905, + "learning_rate": 2.47766043866775e-06, + "loss": 2.1342, + "step": 1532 + }, + { + "epoch": 1.6800000000000002, + "grad_norm": 1.5806915760040283, + "learning_rate": 2.475629569455727e-06, + "loss": 2.1229, + "step": 1533 + }, + { + "epoch": 1.6810958904109587, + "grad_norm": 1.5785084962844849, + "learning_rate": 2.4735987002437044e-06, + "loss": 2.0883, + "step": 1534 + }, + { + "epoch": 1.6821917808219178, + "grad_norm": 1.587751030921936, + "learning_rate": 2.4715678310316816e-06, + "loss": 2.0956, + "step": 1535 + }, + { + "epoch": 1.6832876712328768, + "grad_norm": 1.604783296585083, + "learning_rate": 2.469536961819659e-06, + "loss": 2.1014, + "step": 1536 + }, + { + "epoch": 1.6843835616438356, + "grad_norm": 1.6625169515609741, + "learning_rate": 2.467506092607636e-06, + "loss": 2.1099, + "step": 1537 + }, + { + "epoch": 1.6854794520547944, + "grad_norm": 1.623159646987915, + "learning_rate": 2.4654752233956135e-06, + "loss": 2.1012, + "step": 1538 + }, + { + "epoch": 1.6865753424657535, + "grad_norm": 1.7528589963912964, + "learning_rate": 2.4634443541835907e-06, + "loss": 2.1635, + "step": 1539 + }, + { + "epoch": 1.6876712328767123, + "grad_norm": 1.6822079420089722, + "learning_rate": 2.461413484971568e-06, + "loss": 2.1057, + "step": 1540 + }, + { + "epoch": 1.6876712328767123, + "eval_loss": 2.269556999206543, + "eval_runtime": 37.6668, + "eval_samples_per_second": 26.549, + "eval_steps_per_second": 1.115, + "step": 1540 + }, + { + "epoch": 1.688767123287671, + "grad_norm": 1.5787808895111084, + "learning_rate": 2.4593826157595453e-06, + "loss": 2.1831, + "step": 1541 + }, + { + "epoch": 1.6898630136986301, + "grad_norm": 1.5939741134643555, + "learning_rate": 2.4573517465475226e-06, + "loss": 2.0706, + "step": 1542 + }, + { + "epoch": 1.6909589041095892, + "grad_norm": 1.796216607093811, + "learning_rate": 2.4553208773355e-06, + "loss": 2.0959, + "step": 1543 + }, + { + "epoch": 1.692054794520548, + "grad_norm": 1.6734933853149414, + "learning_rate": 2.453290008123477e-06, + "loss": 2.1098, + "step": 1544 + }, + { + "epoch": 1.6931506849315068, + "grad_norm": 1.5036938190460205, + "learning_rate": 2.4512591389114544e-06, + "loss": 2.12, + "step": 1545 + }, + { + "epoch": 1.6942465753424658, + "grad_norm": 1.631889820098877, + "learning_rate": 2.4492282696994317e-06, + "loss": 2.1207, + "step": 1546 + }, + { + "epoch": 1.6953424657534246, + "grad_norm": 1.6895856857299805, + "learning_rate": 2.447197400487409e-06, + "loss": 2.115, + "step": 1547 + }, + { + "epoch": 1.6964383561643834, + "grad_norm": 1.581718921661377, + "learning_rate": 2.4451665312753862e-06, + "loss": 2.1337, + "step": 1548 + }, + { + "epoch": 1.6975342465753425, + "grad_norm": 1.6116989850997925, + "learning_rate": 2.4431356620633635e-06, + "loss": 2.1269, + "step": 1549 + }, + { + "epoch": 1.6986301369863015, + "grad_norm": 1.5890625715255737, + "learning_rate": 2.4411047928513408e-06, + "loss": 2.1115, + "step": 1550 + }, + { + "epoch": 1.6986301369863015, + "eval_loss": 2.269603729248047, + "eval_runtime": 37.7242, + "eval_samples_per_second": 26.508, + "eval_steps_per_second": 1.113, + "step": 1550 + }, + { + "epoch": 1.6997260273972603, + "grad_norm": 1.6351535320281982, + "learning_rate": 2.439073923639318e-06, + "loss": 2.1243, + "step": 1551 + }, + { + "epoch": 1.7008219178082191, + "grad_norm": 1.628990650177002, + "learning_rate": 2.4370430544272953e-06, + "loss": 2.0874, + "step": 1552 + }, + { + "epoch": 1.7019178082191782, + "grad_norm": 1.6842089891433716, + "learning_rate": 2.435012185215272e-06, + "loss": 2.0774, + "step": 1553 + }, + { + "epoch": 1.703013698630137, + "grad_norm": 1.6843653917312622, + "learning_rate": 2.4329813160032495e-06, + "loss": 2.09, + "step": 1554 + }, + { + "epoch": 1.7041095890410958, + "grad_norm": 1.6409231424331665, + "learning_rate": 2.4309504467912267e-06, + "loss": 2.1029, + "step": 1555 + }, + { + "epoch": 1.7052054794520548, + "grad_norm": 1.6237034797668457, + "learning_rate": 2.428919577579204e-06, + "loss": 2.1014, + "step": 1556 + }, + { + "epoch": 1.7063013698630138, + "grad_norm": 1.7377818822860718, + "learning_rate": 2.4268887083671813e-06, + "loss": 2.1786, + "step": 1557 + }, + { + "epoch": 1.7073972602739727, + "grad_norm": 1.6430894136428833, + "learning_rate": 2.4248578391551586e-06, + "loss": 2.0985, + "step": 1558 + }, + { + "epoch": 1.7084931506849315, + "grad_norm": 1.5623244047164917, + "learning_rate": 2.422826969943136e-06, + "loss": 2.0913, + "step": 1559 + }, + { + "epoch": 1.7095890410958905, + "grad_norm": 1.6417165994644165, + "learning_rate": 2.420796100731113e-06, + "loss": 2.0673, + "step": 1560 + }, + { + "epoch": 1.7095890410958905, + "eval_loss": 2.268665075302124, + "eval_runtime": 37.7309, + "eval_samples_per_second": 26.504, + "eval_steps_per_second": 1.113, + "step": 1560 + }, + { + "epoch": 1.7106849315068493, + "grad_norm": 1.605925440788269, + "learning_rate": 2.4187652315190904e-06, + "loss": 2.1087, + "step": 1561 + }, + { + "epoch": 1.7117808219178081, + "grad_norm": 1.5908552408218384, + "learning_rate": 2.4167343623070677e-06, + "loss": 2.1254, + "step": 1562 + }, + { + "epoch": 1.7128767123287671, + "grad_norm": 1.6716837882995605, + "learning_rate": 2.414703493095045e-06, + "loss": 2.109, + "step": 1563 + }, + { + "epoch": 1.7139726027397262, + "grad_norm": 1.735393762588501, + "learning_rate": 2.4126726238830222e-06, + "loss": 2.1181, + "step": 1564 + }, + { + "epoch": 1.7150684931506848, + "grad_norm": 1.6103562116622925, + "learning_rate": 2.410641754670999e-06, + "loss": 2.1, + "step": 1565 + }, + { + "epoch": 1.7161643835616438, + "grad_norm": 1.6393123865127563, + "learning_rate": 2.4086108854589763e-06, + "loss": 2.1363, + "step": 1566 + }, + { + "epoch": 1.7172602739726028, + "grad_norm": 1.6986010074615479, + "learning_rate": 2.4065800162469536e-06, + "loss": 2.0909, + "step": 1567 + }, + { + "epoch": 1.7183561643835616, + "grad_norm": 1.5851256847381592, + "learning_rate": 2.404549147034931e-06, + "loss": 2.1046, + "step": 1568 + }, + { + "epoch": 1.7194520547945205, + "grad_norm": 1.7347966432571411, + "learning_rate": 2.402518277822908e-06, + "loss": 2.1517, + "step": 1569 + }, + { + "epoch": 1.7205479452054795, + "grad_norm": 1.6968820095062256, + "learning_rate": 2.4004874086108854e-06, + "loss": 2.0702, + "step": 1570 + }, + { + "epoch": 1.7205479452054795, + "eval_loss": 2.26865816116333, + "eval_runtime": 37.6876, + "eval_samples_per_second": 26.534, + "eval_steps_per_second": 1.114, + "step": 1570 + }, + { + "epoch": 1.7216438356164385, + "grad_norm": 1.5828689336776733, + "learning_rate": 2.3984565393988627e-06, + "loss": 2.1416, + "step": 1571 + }, + { + "epoch": 1.7227397260273971, + "grad_norm": 1.672504186630249, + "learning_rate": 2.39642567018684e-06, + "loss": 2.09, + "step": 1572 + }, + { + "epoch": 1.7238356164383561, + "grad_norm": 1.6303755044937134, + "learning_rate": 2.3943948009748173e-06, + "loss": 2.1042, + "step": 1573 + }, + { + "epoch": 1.7249315068493152, + "grad_norm": 1.6052359342575073, + "learning_rate": 2.3923639317627945e-06, + "loss": 2.1246, + "step": 1574 + }, + { + "epoch": 1.726027397260274, + "grad_norm": 1.5766305923461914, + "learning_rate": 2.390333062550772e-06, + "loss": 2.0823, + "step": 1575 + }, + { + "epoch": 1.7271232876712328, + "grad_norm": 1.6897045373916626, + "learning_rate": 2.388302193338749e-06, + "loss": 2.0112, + "step": 1576 + }, + { + "epoch": 1.7282191780821918, + "grad_norm": 1.732053518295288, + "learning_rate": 2.3862713241267264e-06, + "loss": 2.1817, + "step": 1577 + }, + { + "epoch": 1.7293150684931506, + "grad_norm": 1.7444037199020386, + "learning_rate": 2.3842404549147037e-06, + "loss": 2.0971, + "step": 1578 + }, + { + "epoch": 1.7304109589041095, + "grad_norm": 1.6277426481246948, + "learning_rate": 2.382209585702681e-06, + "loss": 2.0825, + "step": 1579 + }, + { + "epoch": 1.7315068493150685, + "grad_norm": 1.765645146369934, + "learning_rate": 2.380178716490658e-06, + "loss": 2.146, + "step": 1580 + }, + { + "epoch": 1.7315068493150685, + "eval_loss": 2.2666616439819336, + "eval_runtime": 37.7024, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 1580 + }, + { + "epoch": 1.7326027397260275, + "grad_norm": 1.6550240516662598, + "learning_rate": 2.3781478472786355e-06, + "loss": 2.1107, + "step": 1581 + }, + { + "epoch": 1.7336986301369863, + "grad_norm": 1.5257171392440796, + "learning_rate": 2.3761169780666128e-06, + "loss": 2.1346, + "step": 1582 + }, + { + "epoch": 1.7347945205479451, + "grad_norm": 1.6080381870269775, + "learning_rate": 2.37408610885459e-06, + "loss": 2.1402, + "step": 1583 + }, + { + "epoch": 1.7358904109589042, + "grad_norm": 1.6197988986968994, + "learning_rate": 2.3720552396425673e-06, + "loss": 2.0894, + "step": 1584 + }, + { + "epoch": 1.736986301369863, + "grad_norm": 1.5313374996185303, + "learning_rate": 2.3700243704305446e-06, + "loss": 2.1612, + "step": 1585 + }, + { + "epoch": 1.7380821917808218, + "grad_norm": 1.5613369941711426, + "learning_rate": 2.367993501218522e-06, + "loss": 2.1149, + "step": 1586 + }, + { + "epoch": 1.7391780821917808, + "grad_norm": 1.616064429283142, + "learning_rate": 2.365962632006499e-06, + "loss": 2.1186, + "step": 1587 + }, + { + "epoch": 1.7402739726027399, + "grad_norm": 1.6820755004882812, + "learning_rate": 2.3639317627944764e-06, + "loss": 2.1373, + "step": 1588 + }, + { + "epoch": 1.7413698630136987, + "grad_norm": 1.6518917083740234, + "learning_rate": 2.3619008935824537e-06, + "loss": 2.1316, + "step": 1589 + }, + { + "epoch": 1.7424657534246575, + "grad_norm": 1.6021987199783325, + "learning_rate": 2.3598700243704305e-06, + "loss": 2.1281, + "step": 1590 + }, + { + "epoch": 1.7424657534246575, + "eval_loss": 2.2650322914123535, + "eval_runtime": 37.6736, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 1590 + }, + { + "epoch": 1.7435616438356165, + "grad_norm": 1.571864366531372, + "learning_rate": 2.357839155158408e-06, + "loss": 2.1237, + "step": 1591 + }, + { + "epoch": 1.7446575342465753, + "grad_norm": 1.5786241292953491, + "learning_rate": 2.355808285946385e-06, + "loss": 2.116, + "step": 1592 + }, + { + "epoch": 1.7457534246575341, + "grad_norm": 1.8110829591751099, + "learning_rate": 2.3537774167343624e-06, + "loss": 2.1398, + "step": 1593 + }, + { + "epoch": 1.7468493150684932, + "grad_norm": 1.648918628692627, + "learning_rate": 2.3517465475223396e-06, + "loss": 2.1183, + "step": 1594 + }, + { + "epoch": 1.7479452054794522, + "grad_norm": 1.6972789764404297, + "learning_rate": 2.349715678310317e-06, + "loss": 2.133, + "step": 1595 + }, + { + "epoch": 1.749041095890411, + "grad_norm": 1.6375590562820435, + "learning_rate": 2.347684809098294e-06, + "loss": 2.0891, + "step": 1596 + }, + { + "epoch": 1.7501369863013698, + "grad_norm": 1.7189046144485474, + "learning_rate": 2.3456539398862715e-06, + "loss": 2.1456, + "step": 1597 + }, + { + "epoch": 1.7512328767123289, + "grad_norm": 1.729980230331421, + "learning_rate": 2.3436230706742487e-06, + "loss": 2.0876, + "step": 1598 + }, + { + "epoch": 1.7523287671232877, + "grad_norm": 1.8501031398773193, + "learning_rate": 2.341592201462226e-06, + "loss": 2.1071, + "step": 1599 + }, + { + "epoch": 1.7534246575342465, + "grad_norm": 1.6985840797424316, + "learning_rate": 2.3395613322502033e-06, + "loss": 2.1006, + "step": 1600 + }, + { + "epoch": 1.7534246575342465, + "eval_loss": 2.2667760848999023, + "eval_runtime": 37.7374, + "eval_samples_per_second": 26.499, + "eval_steps_per_second": 1.113, + "step": 1600 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 4.987695054979072e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/latest new file mode 100644 index 0000000000000000000000000000000000000000..cb24338af451be6d3ccddc18a950b58a70589761 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/latest @@ -0,0 +1 @@ +global_step1700 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..e56f444a187cbde00a193842f977eebfb56a9340 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/trainer_state.json @@ -0,0 +1,13293 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.8630136986301369, + "eval_steps": 10, + "global_step": 1700, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9794018268585205, + "learning_rate": 1.551094890510949e-06, + "loss": 2.3047, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.396454334259033, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.3671, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 2.0826210975646973, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.4105, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.3590970039367676, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.3405, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 2.3945844173431396, + "learning_rate": 1.624087591240876e-06, + "loss": 2.2922, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 2.185126781463623, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.3399, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.2687461376190186, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.3547, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.920499324798584, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.28, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.127774715423584, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.3409, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 1.8677122592926025, + "learning_rate": 1.715328467153285e-06, + "loss": 2.3406, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.4506173133850098, + "eval_runtime": 37.5008, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9803308248519897, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.3642, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 1.7733794450759888, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.3666, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 1.7639747858047485, + "learning_rate": 1.77007299270073e-06, + "loss": 2.3926, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 1.9172815084457397, + "learning_rate": 1.788321167883212e-06, + "loss": 2.3431, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 1.580621361732483, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.2967, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.0889124870300293, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.331, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.6659903526306152, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.3181, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.028651237487793, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.3237, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.681212306022644, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.32, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 1.9986696243286133, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.3425, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.4408469200134277, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 1.832582950592041, + "learning_rate": 1.916058394160584e-06, + "loss": 2.2548, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 1.8609144687652588, + "learning_rate": 1.934306569343066e-06, + "loss": 2.3024, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 1.8699817657470703, + "learning_rate": 1.952554744525548e-06, + "loss": 2.3044, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 1.81354558467865, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.3098, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 1.7789452075958252, + "learning_rate": 1.989051094890511e-06, + "loss": 2.3623, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 1.6322330236434937, + "learning_rate": 2.007299270072993e-06, + "loss": 2.303, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 1.8742916584014893, + "learning_rate": 2.025547445255475e-06, + "loss": 2.3183, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.7790557146072388, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.3168, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 1.7624462842941284, + "learning_rate": 2.062043795620438e-06, + "loss": 2.3629, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.781114101409912, + "learning_rate": 2.08029197080292e-06, + "loss": 2.2813, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.426335334777832, + "eval_runtime": 37.7146, + "eval_samples_per_second": 26.515, + "eval_steps_per_second": 1.114, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.769988775253296, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.3396, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.0640275478363037, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.3688, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.7884094715118408, + "learning_rate": 2.135036496350365e-06, + "loss": 2.2775, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.831847071647644, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.2966, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9044525623321533, + "learning_rate": 2.171532846715329e-06, + "loss": 2.3359, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.9899096488952637, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.3099, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8570183515548706, + "learning_rate": 2.208029197080292e-06, + "loss": 2.3016, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 1.8255733251571655, + "learning_rate": 2.226277372262774e-06, + "loss": 2.335, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.8759729862213135, + "learning_rate": 2.244525547445256e-06, + "loss": 2.3357, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.9031389951705933, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.3014, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.4213266372680664, + "eval_runtime": 37.6757, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.1044118404388428, + "learning_rate": 2.281021897810219e-06, + "loss": 2.2946, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.005735158920288, + "learning_rate": 2.299270072992701e-06, + "loss": 2.3132, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.8821312189102173, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.2816, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7866829633712769, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.309, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.8452801704406738, + "learning_rate": 2.354014598540146e-06, + "loss": 2.3222, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.5781638622283936, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.2688, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 1.9746729135513306, + "learning_rate": 2.39051094890511e-06, + "loss": 2.2864, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 2.0312447547912598, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.2914, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.675674557685852, + "learning_rate": 2.427007299270073e-06, + "loss": 2.3441, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 1.8987798690795898, + "learning_rate": 2.445255474452555e-06, + "loss": 2.2589, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.4149818420410156, + "eval_runtime": 37.6627, + "eval_samples_per_second": 26.551, + "eval_steps_per_second": 1.115, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 2.041752338409424, + "learning_rate": 2.463503649635037e-06, + "loss": 2.2457, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.7845569849014282, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.3376, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 2.489330530166626, + "learning_rate": 2.5e-06, + "loss": 2.258, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 2.3793296813964844, + "learning_rate": 2.518248175182482e-06, + "loss": 2.2561, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 2.0474460124969482, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.2992, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.7181849479675293, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.3003, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 2.3854947090148926, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.2866, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.7228457927703857, + "learning_rate": 2.591240875912409e-06, + "loss": 2.3027, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.3928253650665283, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.3233, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.136934757232666, + "learning_rate": 2.627737226277373e-06, + "loss": 2.2949, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.410449266433716, + "eval_runtime": 37.6837, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.8906159400939941, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.2693, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 1.8668863773345947, + "learning_rate": 2.664233576642336e-06, + "loss": 2.2944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.319363594055176, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.2791, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0869688987731934, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.32, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 1.8910208940505981, + "learning_rate": 2.718978102189781e-06, + "loss": 2.3059, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.157984495162964, + "learning_rate": 2.737226277372263e-06, + "loss": 2.2756, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.387314558029175, + "learning_rate": 2.755474452554745e-06, + "loss": 2.3597, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.183333158493042, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.322, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.158015489578247, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.2663, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 1.9594391584396362, + "learning_rate": 2.81021897810219e-06, + "loss": 2.3, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.411637306213379, + "eval_runtime": 37.6428, + "eval_samples_per_second": 26.565, + "eval_steps_per_second": 1.116, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.28957462310791, + "learning_rate": 2.828467153284672e-06, + "loss": 2.2258, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.8085683584213257, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.2739, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 1.9662787914276123, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.2398, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 2.337609052658081, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.3216, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 1.6008479595184326, + "learning_rate": 2.901459854014599e-06, + "loss": 2.2997, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 2.177690029144287, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.2893, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8746743202209473, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.3303, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 1.7862998247146606, + "learning_rate": 2.956204379562044e-06, + "loss": 2.3007, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.7578285932540894, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.2591, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.1532702445983887, + "learning_rate": 2.992700729927008e-06, + "loss": 2.2544, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.3992533683776855, + "eval_runtime": 37.6571, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.0168659687042236, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.3108, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 1.8336135149002075, + "learning_rate": 3.029197080291971e-06, + "loss": 2.3044, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.6675384044647217, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.2978, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 1.8866558074951172, + "learning_rate": 3.065693430656935e-06, + "loss": 2.3169, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0588760375976562, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.3041, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.9300280809402466, + "learning_rate": 3.102189781021898e-06, + "loss": 2.2494, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 1.837496042251587, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.2339, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.155080795288086, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.324, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 1.9562510251998901, + "learning_rate": 3.156934306569343e-06, + "loss": 2.2937, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 2.522388219833374, + "learning_rate": 3.175182481751825e-06, + "loss": 2.2683, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.3936216831207275, + "eval_runtime": 37.6499, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 1.7409396171569824, + "learning_rate": 3.193430656934307e-06, + "loss": 2.2833, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.176687002182007, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.2748, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 1.8576542139053345, + "learning_rate": 3.2299270072992707e-06, + "loss": 2.2144, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.9564870595932007, + "learning_rate": 3.248175182481752e-06, + "loss": 2.2898, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.896238088607788, + "learning_rate": 3.266423357664234e-06, + "loss": 2.2793, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.735960841178894, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.2869, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.018303394317627, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.2775, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 1.7448145151138306, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.2736, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8245689868927002, + "learning_rate": 3.339416058394161e-06, + "loss": 2.2799, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 1.821118950843811, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.2823, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.3893656730651855, + "eval_runtime": 37.676, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 1.9112446308135986, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.287, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 1.6978867053985596, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.2865, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9376300573349, + "learning_rate": 3.412408759124088e-06, + "loss": 2.2326, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.8067563772201538, + "learning_rate": 3.43065693430657e-06, + "loss": 2.2739, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.8476370573043823, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.2328, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.927923321723938, + "learning_rate": 3.467153284671533e-06, + "loss": 2.2679, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7956453561782837, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.2607, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 1.857413649559021, + "learning_rate": 3.503649635036497e-06, + "loss": 2.2854, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 1.8384093046188354, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.3021, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.7346371412277222, + "learning_rate": 3.54014598540146e-06, + "loss": 2.3402, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.390045642852783, + "eval_runtime": 37.4358, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.7546677589416504, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.3214, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 1.7437881231307983, + "learning_rate": 3.576642335766424e-06, + "loss": 2.2004, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9509227275848389, + "learning_rate": 3.5948905109489054e-06, + "loss": 2.2044, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 1.6918139457702637, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.2369, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9664885997772217, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.254, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 1.7942848205566406, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.2296, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 1.8610197305679321, + "learning_rate": 3.667883211678833e-06, + "loss": 2.2273, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 1.8405956029891968, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.3034, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 1.7151877880096436, + "learning_rate": 3.704379562043796e-06, + "loss": 2.2689, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 1.8795795440673828, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.2363, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.384481430053711, + "eval_runtime": 37.6512, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 1.7494690418243408, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.257, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 1.7700639963150024, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.2457, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 1.846092700958252, + "learning_rate": 3.777372262773723e-06, + "loss": 2.2425, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 1.6699751615524292, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.2734, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 1.8978508710861206, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.2435, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.6387466192245483, + "learning_rate": 3.832116788321168e-06, + "loss": 2.2241, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 1.6999541521072388, + "learning_rate": 3.85036496350365e-06, + "loss": 2.2468, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 1.6322710514068604, + "learning_rate": 3.868613138686132e-06, + "loss": 2.2642, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.5633572340011597, + "learning_rate": 3.886861313868613e-06, + "loss": 2.2862, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 1.7642648220062256, + "learning_rate": 3.905109489051096e-06, + "loss": 2.2414, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.3775289058685303, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.76724374294281, + "learning_rate": 3.923357664233577e-06, + "loss": 2.2976, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.6049190759658813, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.3034, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.6591981649398804, + "learning_rate": 3.95985401459854e-06, + "loss": 2.1957, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.650676965713501, + "learning_rate": 3.978102189781022e-06, + "loss": 2.275, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.6665960550308228, + "learning_rate": 3.996350364963504e-06, + "loss": 2.1853, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.5785901546478271, + "learning_rate": 4.014598540145986e-06, + "loss": 2.2399, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.976162075996399, + "learning_rate": 4.032846715328467e-06, + "loss": 2.2499, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.642327904701233, + "learning_rate": 4.05109489051095e-06, + "loss": 2.235, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 2.0126261711120605, + "learning_rate": 4.069343065693431e-06, + "loss": 2.2466, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.6737161874771118, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.2984, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.375901222229004, + "eval_runtime": 37.7072, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 2.094257354736328, + "learning_rate": 4.105839416058395e-06, + "loss": 2.251, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.655393123626709, + "learning_rate": 4.124087591240876e-06, + "loss": 2.2366, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 2.5909082889556885, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.2986, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.8278532028198242, + "learning_rate": 4.16058394160584e-06, + "loss": 2.1977, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9404895305633545, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.2578, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.6525582075119019, + "learning_rate": 4.197080291970803e-06, + "loss": 2.2263, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 2.021822929382324, + "learning_rate": 4.215328467153285e-06, + "loss": 2.2326, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 1.9136033058166504, + "learning_rate": 4.233576642335767e-06, + "loss": 2.2457, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.389714002609253, + "learning_rate": 4.251824817518249e-06, + "loss": 2.2256, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.7886695861816406, + "learning_rate": 4.27007299270073e-06, + "loss": 2.2222, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.3729844093322754, + "eval_runtime": 37.7638, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.9499074220657349, + "learning_rate": 4.288321167883212e-06, + "loss": 2.2026, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 1.834797978401184, + "learning_rate": 4.306569343065693e-06, + "loss": 2.2564, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 2.0415048599243164, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.2462, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 1.6527886390686035, + "learning_rate": 4.343065693430658e-06, + "loss": 2.2493, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.0504889488220215, + "learning_rate": 4.361313868613139e-06, + "loss": 2.2734, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.004772424697876, + "learning_rate": 4.379562043795621e-06, + "loss": 2.2469, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.1973347663879395, + "learning_rate": 4.397810218978102e-06, + "loss": 2.2479, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.8222625255584717, + "learning_rate": 4.416058394160584e-06, + "loss": 2.2206, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 2.1759912967681885, + "learning_rate": 4.434306569343066e-06, + "loss": 2.261, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.8344624042510986, + "learning_rate": 4.452554744525548e-06, + "loss": 2.2605, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.3706893920898438, + "eval_runtime": 37.734, + "eval_samples_per_second": 26.501, + "eval_steps_per_second": 1.113, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.7783704996109009, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.284, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 2.101193428039551, + "learning_rate": 4.489051094890512e-06, + "loss": 2.2524, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 1.7923569679260254, + "learning_rate": 4.507299270072993e-06, + "loss": 2.2434, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 2.170883893966675, + "learning_rate": 4.525547445255475e-06, + "loss": 2.2173, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.6749192476272583, + "learning_rate": 4.543795620437956e-06, + "loss": 2.283, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.0484230518341064, + "learning_rate": 4.562043795620438e-06, + "loss": 2.2931, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.9212749004364014, + "learning_rate": 4.580291970802921e-06, + "loss": 2.2661, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.6729028224945068, + "learning_rate": 4.598540145985402e-06, + "loss": 2.2269, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 2.1190619468688965, + "learning_rate": 4.6167883211678835e-06, + "loss": 2.2278, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.7444697618484497, + "learning_rate": 4.635036496350365e-06, + "loss": 2.2634, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.3706092834472656, + "eval_runtime": 37.7391, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 1.8921526670455933, + "learning_rate": 4.653284671532847e-06, + "loss": 2.2697, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7347339391708374, + "learning_rate": 4.671532846715329e-06, + "loss": 2.2406, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 2.1617953777313232, + "learning_rate": 4.689781021897811e-06, + "loss": 2.2891, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.873384714126587, + "learning_rate": 4.708029197080292e-06, + "loss": 2.2554, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.801356554031372, + "learning_rate": 4.726277372262774e-06, + "loss": 2.2073, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.9250633716583252, + "learning_rate": 4.744525547445255e-06, + "loss": 2.2501, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 1.8360860347747803, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.2561, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 1.8711498975753784, + "learning_rate": 4.78102189781022e-06, + "loss": 2.2247, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.6604810953140259, + "learning_rate": 4.799270072992701e-06, + "loss": 2.2448, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 1.689201831817627, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.2494, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.3682427406311035, + "eval_runtime": 37.7233, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.720961093902588, + "learning_rate": 4.835766423357665e-06, + "loss": 2.2013, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.7863153219223022, + "learning_rate": 4.854014598540146e-06, + "loss": 2.2183, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 1.8583060503005981, + "learning_rate": 4.872262773722628e-06, + "loss": 2.2013, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 1.722428560256958, + "learning_rate": 4.89051094890511e-06, + "loss": 2.2399, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.8408993482589722, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.2489, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.887604832649231, + "learning_rate": 4.927007299270074e-06, + "loss": 2.2145, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 1.7112189531326294, + "learning_rate": 4.945255474452555e-06, + "loss": 2.2915, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 1.8314064741134644, + "learning_rate": 4.963503649635037e-06, + "loss": 2.2786, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.7001186609268188, + "learning_rate": 4.981751824817518e-06, + "loss": 2.249, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.5820505619049072, + "learning_rate": 5e-06, + "loss": 2.254, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.3624653816223145, + "eval_runtime": 37.7283, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 1.8516303300857544, + "learning_rate": 4.997969130787978e-06, + "loss": 2.1894, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 1.8591808080673218, + "learning_rate": 4.995938261575955e-06, + "loss": 2.2537, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.6822415590286255, + "learning_rate": 4.993907392363932e-06, + "loss": 2.1978, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 2.0779759883880615, + "learning_rate": 4.9918765231519095e-06, + "loss": 2.2286, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.6937508583068848, + "learning_rate": 4.989845653939887e-06, + "loss": 2.2339, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.9342437982559204, + "learning_rate": 4.987814784727864e-06, + "loss": 2.2145, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.773258924484253, + "learning_rate": 4.985783915515841e-06, + "loss": 2.2368, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.8196488618850708, + "learning_rate": 4.983753046303819e-06, + "loss": 2.2224, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.6786775588989258, + "learning_rate": 4.981722177091796e-06, + "loss": 2.2439, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.7761427164077759, + "learning_rate": 4.979691307879773e-06, + "loss": 2.236, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.3614282608032227, + "eval_runtime": 37.7314, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.8389159440994263, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.2745, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.7774744033813477, + "learning_rate": 4.975629569455728e-06, + "loss": 2.2122, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.7862017154693604, + "learning_rate": 4.973598700243705e-06, + "loss": 2.2097, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.8851722478866577, + "learning_rate": 4.971567831031681e-06, + "loss": 2.118, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.068608045578003, + "learning_rate": 4.9695369618196595e-06, + "loss": 2.1884, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.8885847330093384, + "learning_rate": 4.967506092607636e-06, + "loss": 2.2097, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.9441415071487427, + "learning_rate": 4.965475223395614e-06, + "loss": 2.2068, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 2.313577890396118, + "learning_rate": 4.9634443541835905e-06, + "loss": 2.2355, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 1.8542472124099731, + "learning_rate": 4.961413484971569e-06, + "loss": 2.2299, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9382838010787964, + "learning_rate": 4.959382615759545e-06, + "loss": 2.2072, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.361767053604126, + "eval_runtime": 37.4603, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.081242322921753, + "learning_rate": 4.957351746547523e-06, + "loss": 2.3167, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 1.9109065532684326, + "learning_rate": 4.9553208773355e-06, + "loss": 2.2336, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.124897003173828, + "learning_rate": 4.953290008123478e-06, + "loss": 2.184, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.8820433616638184, + "learning_rate": 4.951259138911454e-06, + "loss": 2.2836, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.016204595565796, + "learning_rate": 4.9492282696994315e-06, + "loss": 2.2221, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.0679776668548584, + "learning_rate": 4.947197400487409e-06, + "loss": 2.2113, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 1.8040891885757446, + "learning_rate": 4.945166531275386e-06, + "loss": 2.2292, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.0315871238708496, + "learning_rate": 4.943135662063363e-06, + "loss": 2.2657, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.71120023727417, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.2517, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0443193912506104, + "learning_rate": 4.939073923639318e-06, + "loss": 2.2105, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.3582265377044678, + "eval_runtime": 37.6749, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.9363837242126465, + "learning_rate": 4.937043054427295e-06, + "loss": 2.1944, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 2.0878167152404785, + "learning_rate": 4.935012185215272e-06, + "loss": 2.2477, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.9059010744094849, + "learning_rate": 4.93298131600325e-06, + "loss": 2.2278, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 2.220851421356201, + "learning_rate": 4.930950446791227e-06, + "loss": 2.2391, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.7555570602416992, + "learning_rate": 4.928919577579204e-06, + "loss": 2.2206, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 2.1546502113342285, + "learning_rate": 4.9268887083671815e-06, + "loss": 2.2216, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 1.9688911437988281, + "learning_rate": 4.924857839155159e-06, + "loss": 2.2423, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 1.983587622642517, + "learning_rate": 4.922826969943136e-06, + "loss": 2.1738, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 2.1990034580230713, + "learning_rate": 4.920796100731113e-06, + "loss": 2.2341, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 1.9690128564834595, + "learning_rate": 4.918765231519091e-06, + "loss": 2.1914, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.36171817779541, + "eval_runtime": 37.727, + "eval_samples_per_second": 26.506, + "eval_steps_per_second": 1.113, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 2.5950212478637695, + "learning_rate": 4.916734362307068e-06, + "loss": 2.1921, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 2.023130416870117, + "learning_rate": 4.914703493095045e-06, + "loss": 2.1932, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2486252784729004, + "learning_rate": 4.912672623883022e-06, + "loss": 2.2174, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 1.723677635192871, + "learning_rate": 4.910641754671e-06, + "loss": 2.2173, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.94509756565094, + "learning_rate": 4.908610885458977e-06, + "loss": 2.2539, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 1.6934322118759155, + "learning_rate": 4.906580016246954e-06, + "loss": 2.2345, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 1.6882652044296265, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.231, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.6730620861053467, + "learning_rate": 4.902518277822909e-06, + "loss": 2.2394, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.7261624336242676, + "learning_rate": 4.900487408610886e-06, + "loss": 2.2524, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.8310920000076294, + "learning_rate": 4.898456539398863e-06, + "loss": 2.3004, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.3536295890808105, + "eval_runtime": 37.7461, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.7025809288024902, + "learning_rate": 4.89642567018684e-06, + "loss": 2.1923, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.7285082340240479, + "learning_rate": 4.894394800974818e-06, + "loss": 2.241, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 1.7589807510375977, + "learning_rate": 4.892363931762794e-06, + "loss": 2.2019, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.6664931774139404, + "learning_rate": 4.8903330625507725e-06, + "loss": 2.1822, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 1.8308027982711792, + "learning_rate": 4.888302193338749e-06, + "loss": 2.2651, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.6426894664764404, + "learning_rate": 4.886271324126727e-06, + "loss": 2.2514, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.9765205383300781, + "learning_rate": 4.8842404549147034e-06, + "loss": 2.226, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.7872203588485718, + "learning_rate": 4.8822095857026816e-06, + "loss": 2.18, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 2.0645813941955566, + "learning_rate": 4.880178716490658e-06, + "loss": 2.2023, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 1.9823166131973267, + "learning_rate": 4.878147847278636e-06, + "loss": 2.1801, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.353776454925537, + "eval_runtime": 37.771, + "eval_samples_per_second": 26.475, + "eval_steps_per_second": 1.112, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.831152319908142, + "learning_rate": 4.8761169780666125e-06, + "loss": 2.1858, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 1.9321790933609009, + "learning_rate": 4.874086108854591e-06, + "loss": 2.1638, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.9097596406936646, + "learning_rate": 4.872055239642567e-06, + "loss": 2.2206, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8821815252304077, + "learning_rate": 4.870024370430544e-06, + "loss": 2.2338, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.7959749698638916, + "learning_rate": 4.867993501218522e-06, + "loss": 2.1882, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 1.625320315361023, + "learning_rate": 4.865962632006499e-06, + "loss": 2.2137, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 1.6914536952972412, + "learning_rate": 4.863931762794476e-06, + "loss": 2.2374, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0801584720611572, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.2517, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 1.6874804496765137, + "learning_rate": 4.859870024370431e-06, + "loss": 2.2621, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 1.9347832202911377, + "learning_rate": 4.857839155158408e-06, + "loss": 2.2326, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.3468236923217773, + "eval_runtime": 37.7184, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.6798334121704102, + "learning_rate": 4.855808285946385e-06, + "loss": 2.2222, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 1.8993884325027466, + "learning_rate": 4.8537774167343626e-06, + "loss": 2.2178, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 1.697892189025879, + "learning_rate": 4.85174654752234e-06, + "loss": 2.2217, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8484156131744385, + "learning_rate": 4.849715678310317e-06, + "loss": 2.2069, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 1.8125075101852417, + "learning_rate": 4.847684809098294e-06, + "loss": 2.2285, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 1.8215030431747437, + "learning_rate": 4.845653939886272e-06, + "loss": 2.2029, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 1.777129054069519, + "learning_rate": 4.843623070674249e-06, + "loss": 2.2512, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.7327351570129395, + "learning_rate": 4.841592201462226e-06, + "loss": 2.1767, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.7363202571868896, + "learning_rate": 4.8395613322502035e-06, + "loss": 2.2174, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9183942079544067, + "learning_rate": 4.837530463038181e-06, + "loss": 2.2383, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.345961332321167, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.7290767431259155, + "learning_rate": 4.835499593826158e-06, + "loss": 2.22, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 1.7892694473266602, + "learning_rate": 4.833468724614135e-06, + "loss": 2.1403, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.6671900749206543, + "learning_rate": 4.831437855402113e-06, + "loss": 2.1989, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 1.848751425743103, + "learning_rate": 4.82940698619009e-06, + "loss": 2.245, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.6957710981369019, + "learning_rate": 4.827376116978067e-06, + "loss": 2.2239, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.650583267211914, + "learning_rate": 4.8253452477660444e-06, + "loss": 2.2101, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.6510974168777466, + "learning_rate": 4.823314378554022e-06, + "loss": 2.204, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 1.9469026327133179, + "learning_rate": 4.821283509341998e-06, + "loss": 2.221, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 1.6031194925308228, + "learning_rate": 4.819252640129976e-06, + "loss": 2.1926, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.120325803756714, + "learning_rate": 4.817221770917953e-06, + "loss": 2.2166, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.3419017791748047, + "eval_runtime": 37.6854, + "eval_samples_per_second": 26.535, + "eval_steps_per_second": 1.114, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.5533722639083862, + "learning_rate": 4.815190901705931e-06, + "loss": 2.1981, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9650007486343384, + "learning_rate": 4.813160032493907e-06, + "loss": 2.2319, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 1.6874849796295166, + "learning_rate": 4.811129163281885e-06, + "loss": 2.2263, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.787996768951416, + "learning_rate": 4.809098294069862e-06, + "loss": 2.2546, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 1.7853440046310425, + "learning_rate": 4.80706742485784e-06, + "loss": 2.1994, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.7882641553878784, + "learning_rate": 4.805036555645816e-06, + "loss": 2.2449, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 1.8206003904342651, + "learning_rate": 4.8030056864337945e-06, + "loss": 2.2144, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.7012749910354614, + "learning_rate": 4.800974817221771e-06, + "loss": 2.2122, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 1.8711670637130737, + "learning_rate": 4.798943948009749e-06, + "loss": 2.2132, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 1.6880611181259155, + "learning_rate": 4.7969130787977254e-06, + "loss": 2.2336, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.340393543243408, + "eval_runtime": 37.6885, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 2.0543065071105957, + "learning_rate": 4.794882209585703e-06, + "loss": 2.1899, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 1.5774480104446411, + "learning_rate": 4.79285134037368e-06, + "loss": 2.1982, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 1.7154620885849, + "learning_rate": 4.790820471161657e-06, + "loss": 2.2096, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.8200151920318604, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.2391, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 1.6368298530578613, + "learning_rate": 4.786758732737612e-06, + "loss": 2.1855, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 1.6851340532302856, + "learning_rate": 4.784727863525589e-06, + "loss": 2.2114, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 1.8293818235397339, + "learning_rate": 4.782696994313566e-06, + "loss": 2.1636, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 1.6213513612747192, + "learning_rate": 4.780666125101544e-06, + "loss": 2.1884, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.7791825532913208, + "learning_rate": 4.778635255889521e-06, + "loss": 2.1977, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 1.6627994775772095, + "learning_rate": 4.776604386677498e-06, + "loss": 2.2444, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.3396637439727783, + "eval_runtime": 37.7069, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 1.7597277164459229, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.2369, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.7754322290420532, + "learning_rate": 4.772542648253453e-06, + "loss": 2.1859, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 1.7897601127624512, + "learning_rate": 4.77051177904143e-06, + "loss": 2.2245, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 1.7294282913208008, + "learning_rate": 4.768480909829407e-06, + "loss": 2.1998, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 1.8406974077224731, + "learning_rate": 4.766450040617385e-06, + "loss": 2.2353, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 1.849493384361267, + "learning_rate": 4.764419171405362e-06, + "loss": 2.2165, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.5568877458572388, + "learning_rate": 4.762388302193339e-06, + "loss": 2.1989, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 1.8816689252853394, + "learning_rate": 4.760357432981316e-06, + "loss": 2.2062, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 1.7648332118988037, + "learning_rate": 4.758326563769294e-06, + "loss": 2.2204, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.753944754600525, + "learning_rate": 4.756295694557271e-06, + "loss": 2.1748, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.339653253555298, + "eval_runtime": 37.4438, + "eval_samples_per_second": 26.707, + "eval_steps_per_second": 1.122, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.7343329191207886, + "learning_rate": 4.754264825345248e-06, + "loss": 2.1843, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 1.8125407695770264, + "learning_rate": 4.7522339561332255e-06, + "loss": 2.1492, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.5255184173583984, + "learning_rate": 4.750203086921203e-06, + "loss": 2.2178, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 1.6692545413970947, + "learning_rate": 4.74817221770918e-06, + "loss": 2.2516, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 1.7804819345474243, + "learning_rate": 4.7461413484971565e-06, + "loss": 2.1592, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8048933744430542, + "learning_rate": 4.744110479285135e-06, + "loss": 2.1869, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 1.6439207792282104, + "learning_rate": 4.742079610073111e-06, + "loss": 2.1995, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.7375173568725586, + "learning_rate": 4.740048740861089e-06, + "loss": 2.2067, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 1.9471659660339355, + "learning_rate": 4.738017871649066e-06, + "loss": 2.2513, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 1.6524564027786255, + "learning_rate": 4.735987002437044e-06, + "loss": 2.2643, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.337153434753418, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 1.8574563264846802, + "learning_rate": 4.73395613322502e-06, + "loss": 2.2373, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 1.836464524269104, + "learning_rate": 4.731925264012998e-06, + "loss": 2.1769, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 1.7012423276901245, + "learning_rate": 4.729894394800975e-06, + "loss": 2.1494, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 1.953180193901062, + "learning_rate": 4.727863525588953e-06, + "loss": 2.186, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 1.8468470573425293, + "learning_rate": 4.725832656376929e-06, + "loss": 2.2279, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 2.005077362060547, + "learning_rate": 4.723801787164907e-06, + "loss": 2.1798, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 1.8246952295303345, + "learning_rate": 4.721770917952885e-06, + "loss": 2.1916, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 1.8558839559555054, + "learning_rate": 4.719740048740861e-06, + "loss": 2.2237, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 1.7670999765396118, + "learning_rate": 4.717709179528839e-06, + "loss": 2.2179, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.8544219732284546, + "learning_rate": 4.715678310316816e-06, + "loss": 2.1811, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.333791732788086, + "eval_runtime": 37.7254, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 1.5792925357818604, + "learning_rate": 4.713647441104794e-06, + "loss": 2.1749, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 1.5678222179412842, + "learning_rate": 4.71161657189277e-06, + "loss": 2.2294, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 1.9467812776565552, + "learning_rate": 4.709585702680748e-06, + "loss": 2.2684, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 1.6647964715957642, + "learning_rate": 4.707554833468725e-06, + "loss": 2.1722, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 1.7037622928619385, + "learning_rate": 4.705523964256703e-06, + "loss": 2.1915, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.5857148170471191, + "learning_rate": 4.703493095044679e-06, + "loss": 2.1466, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 1.5792173147201538, + "learning_rate": 4.701462225832657e-06, + "loss": 2.2472, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 1.5998480319976807, + "learning_rate": 4.699431356620634e-06, + "loss": 2.2129, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.6728194952011108, + "learning_rate": 4.697400487408612e-06, + "loss": 2.2057, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 1.5402852296829224, + "learning_rate": 4.695369618196588e-06, + "loss": 2.1903, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.3325934410095215, + "eval_runtime": 37.7463, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 1.5890164375305176, + "learning_rate": 4.693338748984566e-06, + "loss": 2.1584, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 1.805923581123352, + "learning_rate": 4.691307879772543e-06, + "loss": 2.2263, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.5932700634002686, + "learning_rate": 4.68927701056052e-06, + "loss": 2.2109, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 1.5611685514450073, + "learning_rate": 4.6872461413484975e-06, + "loss": 2.192, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 1.6976194381713867, + "learning_rate": 4.685215272136475e-06, + "loss": 2.1998, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.7149630784988403, + "learning_rate": 4.683184402924452e-06, + "loss": 2.2599, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 1.7597423791885376, + "learning_rate": 4.681153533712429e-06, + "loss": 2.1765, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 1.539652943611145, + "learning_rate": 4.679122664500407e-06, + "loss": 2.1635, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.690402865409851, + "learning_rate": 4.677091795288384e-06, + "loss": 2.2417, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.7267673015594482, + "learning_rate": 4.675060926076361e-06, + "loss": 2.166, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.331728935241699, + "eval_runtime": 37.7133, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.6383073329925537, + "learning_rate": 4.673030056864338e-06, + "loss": 2.1624, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.5364140272140503, + "learning_rate": 4.670999187652316e-06, + "loss": 2.1696, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.591477870941162, + "learning_rate": 4.668968318440293e-06, + "loss": 2.2306, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 1.8121449947357178, + "learning_rate": 4.66693744922827e-06, + "loss": 2.1449, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.515539288520813, + "learning_rate": 4.6649065800162475e-06, + "loss": 2.2108, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 1.6521453857421875, + "learning_rate": 4.662875710804225e-06, + "loss": 2.1398, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 1.7074251174926758, + "learning_rate": 4.660844841592202e-06, + "loss": 2.2342, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.7688119411468506, + "learning_rate": 4.658813972380179e-06, + "loss": 2.1287, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.7170679569244385, + "learning_rate": 4.656783103168157e-06, + "loss": 2.2143, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 1.7401350736618042, + "learning_rate": 4.654752233956134e-06, + "loss": 2.1855, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.335416555404663, + "eval_runtime": 37.7115, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.6344956159591675, + "learning_rate": 4.652721364744111e-06, + "loss": 2.2205, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.629184603691101, + "learning_rate": 4.6506904955320885e-06, + "loss": 2.2271, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.769491195678711, + "learning_rate": 4.648659626320066e-06, + "loss": 2.1945, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.6513251066207886, + "learning_rate": 4.646628757108043e-06, + "loss": 2.1705, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9357609748840332, + "learning_rate": 4.6445978878960194e-06, + "loss": 2.1968, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 1.75972318649292, + "learning_rate": 4.6425670186839976e-06, + "loss": 2.2229, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 1.686574935913086, + "learning_rate": 4.640536149471974e-06, + "loss": 2.2361, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 1.655661702156067, + "learning_rate": 4.638505280259952e-06, + "loss": 2.2064, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.7583094835281372, + "learning_rate": 4.6364744110479285e-06, + "loss": 2.2352, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.5906383991241455, + "learning_rate": 4.634443541835907e-06, + "loss": 2.187, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.332923412322998, + "eval_runtime": 37.7496, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 1.708019733428955, + "learning_rate": 4.632412672623883e-06, + "loss": 2.1032, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 1.6350252628326416, + "learning_rate": 4.630381803411861e-06, + "loss": 2.1976, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.9037389755249023, + "learning_rate": 4.628350934199838e-06, + "loss": 2.224, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 1.7126816511154175, + "learning_rate": 4.626320064987816e-06, + "loss": 2.1925, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 1.5828406810760498, + "learning_rate": 4.624289195775792e-06, + "loss": 2.1928, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.6797102689743042, + "learning_rate": 4.62225832656377e-06, + "loss": 2.1902, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 1.6871609687805176, + "learning_rate": 4.620227457351747e-06, + "loss": 2.2024, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.8785804510116577, + "learning_rate": 4.618196588139724e-06, + "loss": 2.1778, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.6175488233566284, + "learning_rate": 4.616165718927701e-06, + "loss": 2.2438, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.7724019289016724, + "learning_rate": 4.6141348497156786e-06, + "loss": 2.2272, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.328923225402832, + "eval_runtime": 37.7443, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.7230926752090454, + "learning_rate": 4.612103980503656e-06, + "loss": 2.2021, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.7899917364120483, + "learning_rate": 4.610073111291633e-06, + "loss": 2.1854, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.5864417552947998, + "learning_rate": 4.60804224207961e-06, + "loss": 2.2042, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.875375509262085, + "learning_rate": 4.606011372867588e-06, + "loss": 2.1844, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 1.830597162246704, + "learning_rate": 4.603980503655565e-06, + "loss": 2.1474, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 1.8624461889266968, + "learning_rate": 4.601949634443542e-06, + "loss": 2.158, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.7161756753921509, + "learning_rate": 4.5999187652315195e-06, + "loss": 2.1838, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.6212890148162842, + "learning_rate": 4.597887896019497e-06, + "loss": 2.1911, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.8586416244506836, + "learning_rate": 4.595857026807474e-06, + "loss": 2.1794, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.653976559638977, + "learning_rate": 4.593826157595451e-06, + "loss": 2.2428, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.3279106616973877, + "eval_runtime": 37.739, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 2.054537057876587, + "learning_rate": 4.591795288383429e-06, + "loss": 2.2003, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.730942964553833, + "learning_rate": 4.589764419171406e-06, + "loss": 2.1816, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.7264728546142578, + "learning_rate": 4.587733549959383e-06, + "loss": 2.1615, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 1.8276281356811523, + "learning_rate": 4.5857026807473604e-06, + "loss": 2.1442, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 1.9947556257247925, + "learning_rate": 4.583671811535338e-06, + "loss": 2.2192, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 1.770601511001587, + "learning_rate": 4.581640942323315e-06, + "loss": 2.233, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 1.8946491479873657, + "learning_rate": 4.579610073111292e-06, + "loss": 2.2249, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 1.7839634418487549, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.2469, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.6298807859420776, + "learning_rate": 4.575548334687247e-06, + "loss": 2.1723, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 1.9310041666030884, + "learning_rate": 4.573517465475224e-06, + "loss": 2.2156, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.3249237537384033, + "eval_runtime": 37.7505, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 1.7901647090911865, + "learning_rate": 4.571486596263201e-06, + "loss": 2.2169, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.7739574909210205, + "learning_rate": 4.569455727051178e-06, + "loss": 2.2005, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 1.7310121059417725, + "learning_rate": 4.567424857839156e-06, + "loss": 2.2369, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.004469156265259, + "learning_rate": 4.565393988627132e-06, + "loss": 2.2051, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 1.7906666994094849, + "learning_rate": 4.5633631194151105e-06, + "loss": 2.1908, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 1.7224774360656738, + "learning_rate": 4.561332250203087e-06, + "loss": 2.1529, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 1.9950156211853027, + "learning_rate": 4.559301380991065e-06, + "loss": 2.1776, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 1.8257590532302856, + "learning_rate": 4.5572705117790414e-06, + "loss": 2.2267, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.038490056991577, + "learning_rate": 4.5552396425670196e-06, + "loss": 2.1703, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 1.7314448356628418, + "learning_rate": 4.553208773354996e-06, + "loss": 2.2409, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.327078104019165, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 2.0013606548309326, + "learning_rate": 4.551177904142974e-06, + "loss": 2.1732, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.7352690696716309, + "learning_rate": 4.5491470349309505e-06, + "loss": 2.2027, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9024865627288818, + "learning_rate": 4.547116165718929e-06, + "loss": 2.1806, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8870717287063599, + "learning_rate": 4.545085296506905e-06, + "loss": 2.1664, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.781226634979248, + "learning_rate": 4.543054427294882e-06, + "loss": 2.166, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.8994221687316895, + "learning_rate": 4.54102355808286e-06, + "loss": 2.1951, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.6999517679214478, + "learning_rate": 4.538992688870837e-06, + "loss": 2.1786, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.899590253829956, + "learning_rate": 4.536961819658814e-06, + "loss": 2.2051, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.8054970502853394, + "learning_rate": 4.5349309504467915e-06, + "loss": 2.1675, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.6732861995697021, + "learning_rate": 4.532900081234769e-06, + "loss": 2.1997, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.3242335319519043, + "eval_runtime": 37.6748, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 2.0063157081604004, + "learning_rate": 4.530869212022746e-06, + "loss": 2.1909, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.7750036716461182, + "learning_rate": 4.528838342810723e-06, + "loss": 2.1935, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9993029832839966, + "learning_rate": 4.526807473598701e-06, + "loss": 2.1753, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.6742994785308838, + "learning_rate": 4.524776604386678e-06, + "loss": 2.2205, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 1.9300428628921509, + "learning_rate": 4.522745735174655e-06, + "loss": 2.2081, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.8770017623901367, + "learning_rate": 4.520714865962632e-06, + "loss": 2.2074, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.847765326499939, + "learning_rate": 4.51868399675061e-06, + "loss": 2.1775, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.905117392539978, + "learning_rate": 4.516653127538587e-06, + "loss": 2.1511, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.7623670101165771, + "learning_rate": 4.514622258326564e-06, + "loss": 2.188, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.9232176542282104, + "learning_rate": 4.5125913891145415e-06, + "loss": 2.2109, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.3252475261688232, + "eval_runtime": 37.6923, + "eval_samples_per_second": 26.531, + "eval_steps_per_second": 1.114, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.6523535251617432, + "learning_rate": 4.510560519902519e-06, + "loss": 2.1499, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9934184551239014, + "learning_rate": 4.508529650690496e-06, + "loss": 2.1383, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.704522728919983, + "learning_rate": 4.506498781478473e-06, + "loss": 2.1892, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.9909989833831787, + "learning_rate": 4.504467912266451e-06, + "loss": 2.2422, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.7490800619125366, + "learning_rate": 4.502437043054428e-06, + "loss": 2.2008, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8620145320892334, + "learning_rate": 4.500406173842405e-06, + "loss": 2.1892, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.7069931030273438, + "learning_rate": 4.4983753046303824e-06, + "loss": 2.1816, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 2.152491807937622, + "learning_rate": 4.49634443541836e-06, + "loss": 2.2331, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.8526151180267334, + "learning_rate": 4.494313566206336e-06, + "loss": 2.1811, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.6865259408950806, + "learning_rate": 4.492282696994314e-06, + "loss": 2.1599, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.3265392780303955, + "eval_runtime": 37.7255, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.7710354328155518, + "learning_rate": 4.490251827782291e-06, + "loss": 2.1636, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.8444210290908813, + "learning_rate": 4.488220958570269e-06, + "loss": 2.1345, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.9095014333724976, + "learning_rate": 4.486190089358245e-06, + "loss": 2.2236, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.7406108379364014, + "learning_rate": 4.484159220146223e-06, + "loss": 2.2078, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.712534785270691, + "learning_rate": 4.4821283509342e-06, + "loss": 2.1809, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8921459913253784, + "learning_rate": 4.480097481722178e-06, + "loss": 2.158, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.805146336555481, + "learning_rate": 4.478066612510154e-06, + "loss": 2.1765, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 2.1620516777038574, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.2832, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.661618947982788, + "learning_rate": 4.474004874086109e-06, + "loss": 2.1397, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.732483983039856, + "learning_rate": 4.471974004874087e-06, + "loss": 2.225, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.3276846408843994, + "eval_runtime": 37.7156, + "eval_samples_per_second": 26.514, + "eval_steps_per_second": 1.114, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.8693495988845825, + "learning_rate": 4.4699431356620635e-06, + "loss": 2.1959, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8201690912246704, + "learning_rate": 4.467912266450041e-06, + "loss": 2.1882, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.9229968786239624, + "learning_rate": 4.465881397238018e-06, + "loss": 2.2006, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.6330770254135132, + "learning_rate": 4.463850528025995e-06, + "loss": 2.209, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9654744863510132, + "learning_rate": 4.4618196588139726e-06, + "loss": 2.1963, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.8808507919311523, + "learning_rate": 4.45978878960195e-06, + "loss": 2.1538, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 1.8481040000915527, + "learning_rate": 4.457757920389927e-06, + "loss": 2.1769, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 2.114439010620117, + "learning_rate": 4.455727051177904e-06, + "loss": 2.2241, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.5492331981658936, + "learning_rate": 4.453696181965882e-06, + "loss": 2.206, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 2.11787486076355, + "learning_rate": 4.451665312753859e-06, + "loss": 2.2328, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.3215034008026123, + "eval_runtime": 37.7471, + "eval_samples_per_second": 26.492, + "eval_steps_per_second": 1.113, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.6561832427978516, + "learning_rate": 4.449634443541836e-06, + "loss": 2.1607, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.7510966062545776, + "learning_rate": 4.4476035743298135e-06, + "loss": 2.1699, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.9424657821655273, + "learning_rate": 4.445572705117791e-06, + "loss": 2.2298, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 1.5836161375045776, + "learning_rate": 4.443541835905768e-06, + "loss": 2.1711, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.895216464996338, + "learning_rate": 4.441510966693745e-06, + "loss": 2.2252, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.3365776538848877, + "learning_rate": 4.439480097481723e-06, + "loss": 2.1309, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.6988815069198608, + "learning_rate": 4.4374492282697e-06, + "loss": 2.2023, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 2.228900671005249, + "learning_rate": 4.435418359057677e-06, + "loss": 2.2021, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.5536632537841797, + "learning_rate": 4.433387489845654e-06, + "loss": 2.1706, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.8463878631591797, + "learning_rate": 4.431356620633632e-06, + "loss": 2.1715, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.320664167404175, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.924741506576538, + "learning_rate": 4.429325751421609e-06, + "loss": 2.1894, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 1.5921828746795654, + "learning_rate": 4.427294882209586e-06, + "loss": 2.1358, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 2.4617693424224854, + "learning_rate": 4.4252640129975635e-06, + "loss": 2.1918, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.6239196062088013, + "learning_rate": 4.423233143785541e-06, + "loss": 2.1996, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.952112078666687, + "learning_rate": 4.421202274573518e-06, + "loss": 2.112, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.663409948348999, + "learning_rate": 4.4191714053614945e-06, + "loss": 2.149, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.831711769104004, + "learning_rate": 4.417140536149473e-06, + "loss": 2.209, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 1.9281738996505737, + "learning_rate": 4.415109666937449e-06, + "loss": 2.1637, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.5661600828170776, + "learning_rate": 4.413078797725427e-06, + "loss": 2.1832, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.6930522918701172, + "learning_rate": 4.411047928513404e-06, + "loss": 2.196, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.318861246109009, + "eval_runtime": 37.6994, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.74558424949646, + "learning_rate": 4.409017059301382e-06, + "loss": 2.2097, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.664880633354187, + "learning_rate": 4.406986190089358e-06, + "loss": 2.2076, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.5981497764587402, + "learning_rate": 4.404955320877336e-06, + "loss": 2.1706, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.6035112142562866, + "learning_rate": 4.402924451665313e-06, + "loss": 2.1538, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 1.6819255352020264, + "learning_rate": 4.400893582453291e-06, + "loss": 2.1816, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.7082457542419434, + "learning_rate": 4.398862713241267e-06, + "loss": 2.1804, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.6893151998519897, + "learning_rate": 4.396831844029245e-06, + "loss": 2.1783, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 1.8639923334121704, + "learning_rate": 4.394800974817222e-06, + "loss": 2.1989, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.6407119035720825, + "learning_rate": 4.392770105605199e-06, + "loss": 2.2013, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.6435848474502563, + "learning_rate": 4.390739236393176e-06, + "loss": 2.2049, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.318084478378296, + "eval_runtime": 37.6999, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.7974674701690674, + "learning_rate": 4.388708367181154e-06, + "loss": 2.188, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 1.8340431451797485, + "learning_rate": 4.386677497969131e-06, + "loss": 2.2305, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.55428946018219, + "learning_rate": 4.384646628757108e-06, + "loss": 2.1837, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.7258585691452026, + "learning_rate": 4.3826157595450855e-06, + "loss": 2.1426, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.7347694635391235, + "learning_rate": 4.380584890333063e-06, + "loss": 2.1161, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.6082992553710938, + "learning_rate": 4.37855402112104e-06, + "loss": 2.2284, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.9446425437927246, + "learning_rate": 4.376523151909017e-06, + "loss": 2.1571, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 1.8597826957702637, + "learning_rate": 4.3744922826969946e-06, + "loss": 2.1982, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.8589327335357666, + "learning_rate": 4.372461413484972e-06, + "loss": 2.2097, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.6825448274612427, + "learning_rate": 4.370430544272949e-06, + "loss": 2.2077, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.315889596939087, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 600 + }, + { + "epoch": 0.6586301369863014, + "grad_norm": 1.8007057905197144, + "learning_rate": 4.368399675060926e-06, + "loss": 2.1843, + "step": 601 + }, + { + "epoch": 0.6597260273972603, + "grad_norm": 2.057697057723999, + "learning_rate": 4.366368805848904e-06, + "loss": 2.13, + "step": 602 + }, + { + "epoch": 0.6608219178082192, + "grad_norm": 1.5714081525802612, + "learning_rate": 4.364337936636881e-06, + "loss": 2.1879, + "step": 603 + }, + { + "epoch": 0.6619178082191781, + "grad_norm": 1.855185866355896, + "learning_rate": 4.362307067424858e-06, + "loss": 2.1659, + "step": 604 + }, + { + "epoch": 0.663013698630137, + "grad_norm": 1.826479196548462, + "learning_rate": 4.3602761982128355e-06, + "loss": 2.1716, + "step": 605 + }, + { + "epoch": 0.6641095890410958, + "grad_norm": 1.6905351877212524, + "learning_rate": 4.358245329000813e-06, + "loss": 2.1825, + "step": 606 + }, + { + "epoch": 0.6652054794520548, + "grad_norm": 1.8157447576522827, + "learning_rate": 4.35621445978879e-06, + "loss": 2.1811, + "step": 607 + }, + { + "epoch": 0.6663013698630137, + "grad_norm": 1.6311850547790527, + "learning_rate": 4.354183590576767e-06, + "loss": 2.1507, + "step": 608 + }, + { + "epoch": 0.6673972602739726, + "grad_norm": 1.579039454460144, + "learning_rate": 4.352152721364745e-06, + "loss": 2.2031, + "step": 609 + }, + { + "epoch": 0.6684931506849315, + "grad_norm": 1.7773518562316895, + "learning_rate": 4.350121852152722e-06, + "loss": 2.1978, + "step": 610 + }, + { + "epoch": 0.6684931506849315, + "eval_loss": 2.314854383468628, + "eval_runtime": 37.4501, + "eval_samples_per_second": 26.702, + "eval_steps_per_second": 1.121, + "step": 610 + }, + { + "epoch": 0.6695890410958905, + "grad_norm": 1.7606587409973145, + "learning_rate": 4.348090982940699e-06, + "loss": 2.1891, + "step": 611 + }, + { + "epoch": 0.6706849315068493, + "grad_norm": 1.6145644187927246, + "learning_rate": 4.3460601137286764e-06, + "loss": 2.1799, + "step": 612 + }, + { + "epoch": 0.6717808219178082, + "grad_norm": 1.8070226907730103, + "learning_rate": 4.344029244516654e-06, + "loss": 2.241, + "step": 613 + }, + { + "epoch": 0.6728767123287671, + "grad_norm": 1.6185903549194336, + "learning_rate": 4.341998375304631e-06, + "loss": 2.1908, + "step": 614 + }, + { + "epoch": 0.673972602739726, + "grad_norm": 1.6690049171447754, + "learning_rate": 4.339967506092607e-06, + "loss": 2.1804, + "step": 615 + }, + { + "epoch": 0.675068493150685, + "grad_norm": 1.740003228187561, + "learning_rate": 4.3379366368805855e-06, + "loss": 2.1734, + "step": 616 + }, + { + "epoch": 0.6761643835616439, + "grad_norm": 1.7311010360717773, + "learning_rate": 4.335905767668562e-06, + "loss": 2.2015, + "step": 617 + }, + { + "epoch": 0.6772602739726027, + "grad_norm": 1.7491734027862549, + "learning_rate": 4.33387489845654e-06, + "loss": 2.1894, + "step": 618 + }, + { + "epoch": 0.6783561643835616, + "grad_norm": 1.6343051195144653, + "learning_rate": 4.3318440292445165e-06, + "loss": 2.1583, + "step": 619 + }, + { + "epoch": 0.6794520547945205, + "grad_norm": 1.584243893623352, + "learning_rate": 4.329813160032495e-06, + "loss": 2.1396, + "step": 620 + }, + { + "epoch": 0.6794520547945205, + "eval_loss": 2.3157904148101807, + "eval_runtime": 37.6734, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 620 + }, + { + "epoch": 0.6805479452054795, + "grad_norm": 1.6275745630264282, + "learning_rate": 4.327782290820471e-06, + "loss": 2.2064, + "step": 621 + }, + { + "epoch": 0.6816438356164384, + "grad_norm": 1.5353672504425049, + "learning_rate": 4.325751421608449e-06, + "loss": 2.1686, + "step": 622 + }, + { + "epoch": 0.6827397260273973, + "grad_norm": 1.6155133247375488, + "learning_rate": 4.323720552396426e-06, + "loss": 2.1652, + "step": 623 + }, + { + "epoch": 0.6838356164383562, + "grad_norm": 1.638383150100708, + "learning_rate": 4.321689683184404e-06, + "loss": 2.1747, + "step": 624 + }, + { + "epoch": 0.684931506849315, + "grad_norm": 1.6153279542922974, + "learning_rate": 4.31965881397238e-06, + "loss": 2.1975, + "step": 625 + }, + { + "epoch": 0.686027397260274, + "grad_norm": 2.023699998855591, + "learning_rate": 4.3176279447603574e-06, + "loss": 2.1904, + "step": 626 + }, + { + "epoch": 0.6871232876712329, + "grad_norm": 1.5969340801239014, + "learning_rate": 4.315597075548335e-06, + "loss": 2.2133, + "step": 627 + }, + { + "epoch": 0.6882191780821918, + "grad_norm": 2.138498067855835, + "learning_rate": 4.313566206336312e-06, + "loss": 2.1685, + "step": 628 + }, + { + "epoch": 0.6893150684931507, + "grad_norm": 1.7141013145446777, + "learning_rate": 4.311535337124289e-06, + "loss": 2.1818, + "step": 629 + }, + { + "epoch": 0.6904109589041096, + "grad_norm": 1.9214462041854858, + "learning_rate": 4.3095044679122665e-06, + "loss": 2.191, + "step": 630 + }, + { + "epoch": 0.6904109589041096, + "eval_loss": 2.3150644302368164, + "eval_runtime": 37.723, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 630 + }, + { + "epoch": 0.6915068493150685, + "grad_norm": 2.03241229057312, + "learning_rate": 4.307473598700244e-06, + "loss": 2.1398, + "step": 631 + }, + { + "epoch": 0.6926027397260274, + "grad_norm": 1.608492374420166, + "learning_rate": 4.305442729488221e-06, + "loss": 2.2113, + "step": 632 + }, + { + "epoch": 0.6936986301369863, + "grad_norm": 2.1223549842834473, + "learning_rate": 4.303411860276198e-06, + "loss": 2.2012, + "step": 633 + }, + { + "epoch": 0.6947945205479452, + "grad_norm": 1.5212339162826538, + "learning_rate": 4.301380991064176e-06, + "loss": 2.2201, + "step": 634 + }, + { + "epoch": 0.6958904109589041, + "grad_norm": 1.9565715789794922, + "learning_rate": 4.299350121852153e-06, + "loss": 2.2155, + "step": 635 + }, + { + "epoch": 0.6969863013698631, + "grad_norm": 1.700822114944458, + "learning_rate": 4.29731925264013e-06, + "loss": 2.1764, + "step": 636 + }, + { + "epoch": 0.6980821917808219, + "grad_norm": 1.5625321865081787, + "learning_rate": 4.2952883834281075e-06, + "loss": 2.1407, + "step": 637 + }, + { + "epoch": 0.6991780821917808, + "grad_norm": 2.3319756984710693, + "learning_rate": 4.293257514216085e-06, + "loss": 2.2058, + "step": 638 + }, + { + "epoch": 0.7002739726027397, + "grad_norm": 1.617644190788269, + "learning_rate": 4.291226645004062e-06, + "loss": 2.2157, + "step": 639 + }, + { + "epoch": 0.7013698630136986, + "grad_norm": 1.9357377290725708, + "learning_rate": 4.289195775792039e-06, + "loss": 2.1736, + "step": 640 + }, + { + "epoch": 0.7013698630136986, + "eval_loss": 2.310267925262451, + "eval_runtime": 37.731, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 640 + }, + { + "epoch": 0.7024657534246576, + "grad_norm": 1.6908962726593018, + "learning_rate": 4.287164906580017e-06, + "loss": 2.1523, + "step": 641 + }, + { + "epoch": 0.7035616438356165, + "grad_norm": 1.6051634550094604, + "learning_rate": 4.285134037367994e-06, + "loss": 2.1499, + "step": 642 + }, + { + "epoch": 0.7046575342465753, + "grad_norm": 1.760164499282837, + "learning_rate": 4.283103168155971e-06, + "loss": 2.1493, + "step": 643 + }, + { + "epoch": 0.7057534246575342, + "grad_norm": 1.6225790977478027, + "learning_rate": 4.281072298943948e-06, + "loss": 2.1737, + "step": 644 + }, + { + "epoch": 0.7068493150684931, + "grad_norm": 1.8128079175949097, + "learning_rate": 4.279041429731926e-06, + "loss": 2.1492, + "step": 645 + }, + { + "epoch": 0.707945205479452, + "grad_norm": 1.993295431137085, + "learning_rate": 4.277010560519903e-06, + "loss": 2.1702, + "step": 646 + }, + { + "epoch": 0.709041095890411, + "grad_norm": 1.7715190649032593, + "learning_rate": 4.27497969130788e-06, + "loss": 2.1939, + "step": 647 + }, + { + "epoch": 0.7101369863013699, + "grad_norm": 1.921738624572754, + "learning_rate": 4.2729488220958575e-06, + "loss": 2.201, + "step": 648 + }, + { + "epoch": 0.7112328767123288, + "grad_norm": 1.5670042037963867, + "learning_rate": 4.270917952883835e-06, + "loss": 2.152, + "step": 649 + }, + { + "epoch": 0.7123287671232876, + "grad_norm": 1.6405668258666992, + "learning_rate": 4.268887083671812e-06, + "loss": 2.2488, + "step": 650 + }, + { + "epoch": 0.7123287671232876, + "eval_loss": 2.3100948333740234, + "eval_runtime": 37.7393, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 650 + }, + { + "epoch": 0.7134246575342466, + "grad_norm": 1.79966402053833, + "learning_rate": 4.266856214459789e-06, + "loss": 2.1506, + "step": 651 + }, + { + "epoch": 0.7145205479452055, + "grad_norm": 1.6062074899673462, + "learning_rate": 4.264825345247766e-06, + "loss": 2.1818, + "step": 652 + }, + { + "epoch": 0.7156164383561644, + "grad_norm": 1.61439049243927, + "learning_rate": 4.262794476035744e-06, + "loss": 2.1269, + "step": 653 + }, + { + "epoch": 0.7167123287671233, + "grad_norm": 1.7649872303009033, + "learning_rate": 4.26076360682372e-06, + "loss": 2.1703, + "step": 654 + }, + { + "epoch": 0.7178082191780822, + "grad_norm": 1.6238528490066528, + "learning_rate": 4.2587327376116984e-06, + "loss": 2.1764, + "step": 655 + }, + { + "epoch": 0.718904109589041, + "grad_norm": 1.6847354173660278, + "learning_rate": 4.256701868399675e-06, + "loss": 2.1796, + "step": 656 + }, + { + "epoch": 0.72, + "grad_norm": 1.7208436727523804, + "learning_rate": 4.254670999187653e-06, + "loss": 2.1903, + "step": 657 + }, + { + "epoch": 0.7210958904109589, + "grad_norm": 1.5888099670410156, + "learning_rate": 4.252640129975629e-06, + "loss": 2.1686, + "step": 658 + }, + { + "epoch": 0.7221917808219178, + "grad_norm": 1.742118000984192, + "learning_rate": 4.2506092607636075e-06, + "loss": 2.1287, + "step": 659 + }, + { + "epoch": 0.7232876712328767, + "grad_norm": 1.8450883626937866, + "learning_rate": 4.248578391551584e-06, + "loss": 2.1761, + "step": 660 + }, + { + "epoch": 0.7232876712328767, + "eval_loss": 2.3134732246398926, + "eval_runtime": 37.7568, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 660 + }, + { + "epoch": 0.7243835616438357, + "grad_norm": 1.630985140800476, + "learning_rate": 4.246547522339562e-06, + "loss": 2.1743, + "step": 661 + }, + { + "epoch": 0.7254794520547945, + "grad_norm": 1.6984914541244507, + "learning_rate": 4.2445166531275385e-06, + "loss": 2.1356, + "step": 662 + }, + { + "epoch": 0.7265753424657534, + "grad_norm": 1.6732920408248901, + "learning_rate": 4.242485783915517e-06, + "loss": 2.1377, + "step": 663 + }, + { + "epoch": 0.7276712328767123, + "grad_norm": 1.6210153102874756, + "learning_rate": 4.240454914703493e-06, + "loss": 2.1455, + "step": 664 + }, + { + "epoch": 0.7287671232876712, + "grad_norm": 1.7643382549285889, + "learning_rate": 4.23842404549147e-06, + "loss": 2.1373, + "step": 665 + }, + { + "epoch": 0.7298630136986302, + "grad_norm": 1.6491565704345703, + "learning_rate": 4.236393176279448e-06, + "loss": 2.1897, + "step": 666 + }, + { + "epoch": 0.7309589041095891, + "grad_norm": 1.670323133468628, + "learning_rate": 4.234362307067425e-06, + "loss": 2.2081, + "step": 667 + }, + { + "epoch": 0.7320547945205479, + "grad_norm": 1.6350057125091553, + "learning_rate": 4.232331437855402e-06, + "loss": 2.2002, + "step": 668 + }, + { + "epoch": 0.7331506849315068, + "grad_norm": 1.6191266775131226, + "learning_rate": 4.2303005686433795e-06, + "loss": 2.1744, + "step": 669 + }, + { + "epoch": 0.7342465753424657, + "grad_norm": 1.5511432886123657, + "learning_rate": 4.228269699431357e-06, + "loss": 2.1797, + "step": 670 + }, + { + "epoch": 0.7342465753424657, + "eval_loss": 2.313626527786255, + "eval_runtime": 37.7763, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 670 + }, + { + "epoch": 0.7353424657534247, + "grad_norm": 1.5128833055496216, + "learning_rate": 4.226238830219334e-06, + "loss": 2.1729, + "step": 671 + }, + { + "epoch": 0.7364383561643836, + "grad_norm": 1.6752032041549683, + "learning_rate": 4.224207961007311e-06, + "loss": 2.1747, + "step": 672 + }, + { + "epoch": 0.7375342465753425, + "grad_norm": 1.8643760681152344, + "learning_rate": 4.2221770917952886e-06, + "loss": 2.174, + "step": 673 + }, + { + "epoch": 0.7386301369863014, + "grad_norm": 1.5659266710281372, + "learning_rate": 4.220146222583266e-06, + "loss": 2.1706, + "step": 674 + }, + { + "epoch": 0.7397260273972602, + "grad_norm": 1.9675986766815186, + "learning_rate": 4.218115353371243e-06, + "loss": 2.1558, + "step": 675 + }, + { + "epoch": 0.7408219178082192, + "grad_norm": 1.6091173887252808, + "learning_rate": 4.21608448415922e-06, + "loss": 2.2126, + "step": 676 + }, + { + "epoch": 0.7419178082191781, + "grad_norm": 1.6801071166992188, + "learning_rate": 4.214053614947198e-06, + "loss": 2.1352, + "step": 677 + }, + { + "epoch": 0.743013698630137, + "grad_norm": 1.6964598894119263, + "learning_rate": 4.212022745735175e-06, + "loss": 2.1863, + "step": 678 + }, + { + "epoch": 0.7441095890410959, + "grad_norm": 1.670597791671753, + "learning_rate": 4.209991876523152e-06, + "loss": 2.1478, + "step": 679 + }, + { + "epoch": 0.7452054794520548, + "grad_norm": 1.6591588258743286, + "learning_rate": 4.2079610073111295e-06, + "loss": 2.2198, + "step": 680 + }, + { + "epoch": 0.7452054794520548, + "eval_loss": 2.3091039657592773, + "eval_runtime": 37.7551, + "eval_samples_per_second": 26.486, + "eval_steps_per_second": 1.112, + "step": 680 + }, + { + "epoch": 0.7463013698630137, + "grad_norm": 1.5147838592529297, + "learning_rate": 4.205930138099107e-06, + "loss": 2.1836, + "step": 681 + }, + { + "epoch": 0.7473972602739726, + "grad_norm": 1.6278952360153198, + "learning_rate": 4.203899268887084e-06, + "loss": 2.196, + "step": 682 + }, + { + "epoch": 0.7484931506849315, + "grad_norm": 1.5964363813400269, + "learning_rate": 4.201868399675061e-06, + "loss": 2.1162, + "step": 683 + }, + { + "epoch": 0.7495890410958904, + "grad_norm": 1.6993104219436646, + "learning_rate": 4.199837530463039e-06, + "loss": 2.1731, + "step": 684 + }, + { + "epoch": 0.7506849315068493, + "grad_norm": 1.6196811199188232, + "learning_rate": 4.197806661251016e-06, + "loss": 2.16, + "step": 685 + }, + { + "epoch": 0.7517808219178083, + "grad_norm": 1.6826317310333252, + "learning_rate": 4.195775792038993e-06, + "loss": 2.1123, + "step": 686 + }, + { + "epoch": 0.7528767123287671, + "grad_norm": 1.7604600191116333, + "learning_rate": 4.19374492282697e-06, + "loss": 2.1962, + "step": 687 + }, + { + "epoch": 0.753972602739726, + "grad_norm": 1.5533616542816162, + "learning_rate": 4.191714053614948e-06, + "loss": 2.1593, + "step": 688 + }, + { + "epoch": 0.7550684931506849, + "grad_norm": 1.8652232885360718, + "learning_rate": 4.189683184402924e-06, + "loss": 2.2067, + "step": 689 + }, + { + "epoch": 0.7561643835616438, + "grad_norm": 1.6330097913742065, + "learning_rate": 4.187652315190902e-06, + "loss": 2.1483, + "step": 690 + }, + { + "epoch": 0.7561643835616438, + "eval_loss": 2.309835195541382, + "eval_runtime": 37.746, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 690 + }, + { + "epoch": 0.7572602739726028, + "grad_norm": 1.5989009141921997, + "learning_rate": 4.185621445978879e-06, + "loss": 2.1686, + "step": 691 + }, + { + "epoch": 0.7583561643835617, + "grad_norm": 1.7416167259216309, + "learning_rate": 4.183590576766857e-06, + "loss": 2.2111, + "step": 692 + }, + { + "epoch": 0.7594520547945205, + "grad_norm": 1.5840280055999756, + "learning_rate": 4.181559707554833e-06, + "loss": 2.1499, + "step": 693 + }, + { + "epoch": 0.7605479452054794, + "grad_norm": 1.839306354522705, + "learning_rate": 4.179528838342811e-06, + "loss": 2.1492, + "step": 694 + }, + { + "epoch": 0.7616438356164383, + "grad_norm": 1.54163658618927, + "learning_rate": 4.177497969130788e-06, + "loss": 2.2029, + "step": 695 + }, + { + "epoch": 0.7627397260273973, + "grad_norm": 1.8050168752670288, + "learning_rate": 4.175467099918766e-06, + "loss": 2.1615, + "step": 696 + }, + { + "epoch": 0.7638356164383562, + "grad_norm": 1.6971197128295898, + "learning_rate": 4.173436230706742e-06, + "loss": 2.1476, + "step": 697 + }, + { + "epoch": 0.7649315068493151, + "grad_norm": 1.6200546026229858, + "learning_rate": 4.1714053614947205e-06, + "loss": 2.1602, + "step": 698 + }, + { + "epoch": 0.766027397260274, + "grad_norm": 1.7440853118896484, + "learning_rate": 4.169374492282697e-06, + "loss": 2.1949, + "step": 699 + }, + { + "epoch": 0.7671232876712328, + "grad_norm": 1.5590331554412842, + "learning_rate": 4.167343623070675e-06, + "loss": 2.1656, + "step": 700 + }, + { + "epoch": 0.7671232876712328, + "eval_loss": 2.309752941131592, + "eval_runtime": 37.7663, + "eval_samples_per_second": 26.479, + "eval_steps_per_second": 1.112, + "step": 700 + }, + { + "epoch": 0.7682191780821918, + "grad_norm": 1.7314256429672241, + "learning_rate": 4.165312753858652e-06, + "loss": 2.1516, + "step": 701 + }, + { + "epoch": 0.7693150684931507, + "grad_norm": 1.5145938396453857, + "learning_rate": 4.163281884646629e-06, + "loss": 2.1795, + "step": 702 + }, + { + "epoch": 0.7704109589041096, + "grad_norm": 1.5300369262695312, + "learning_rate": 4.161251015434607e-06, + "loss": 2.1356, + "step": 703 + }, + { + "epoch": 0.7715068493150685, + "grad_norm": 1.6113280057907104, + "learning_rate": 4.159220146222583e-06, + "loss": 2.157, + "step": 704 + }, + { + "epoch": 0.7726027397260274, + "grad_norm": 1.5458502769470215, + "learning_rate": 4.157189277010561e-06, + "loss": 2.2017, + "step": 705 + }, + { + "epoch": 0.7736986301369863, + "grad_norm": 1.5456187725067139, + "learning_rate": 4.155158407798538e-06, + "loss": 2.1758, + "step": 706 + }, + { + "epoch": 0.7747945205479452, + "grad_norm": 1.6051454544067383, + "learning_rate": 4.153127538586516e-06, + "loss": 2.183, + "step": 707 + }, + { + "epoch": 0.7758904109589041, + "grad_norm": 1.565788745880127, + "learning_rate": 4.151096669374492e-06, + "loss": 2.1538, + "step": 708 + }, + { + "epoch": 0.776986301369863, + "grad_norm": 1.7484439611434937, + "learning_rate": 4.1490658001624705e-06, + "loss": 2.169, + "step": 709 + }, + { + "epoch": 0.7780821917808219, + "grad_norm": 1.7206509113311768, + "learning_rate": 4.147034930950447e-06, + "loss": 2.1661, + "step": 710 + }, + { + "epoch": 0.7780821917808219, + "eval_loss": 2.307927131652832, + "eval_runtime": 37.5145, + "eval_samples_per_second": 26.656, + "eval_steps_per_second": 1.12, + "step": 710 + }, + { + "epoch": 0.7791780821917809, + "grad_norm": 1.8865350484848022, + "learning_rate": 4.145004061738425e-06, + "loss": 2.093, + "step": 711 + }, + { + "epoch": 0.7802739726027397, + "grad_norm": 1.935110092163086, + "learning_rate": 4.1429731925264015e-06, + "loss": 2.1298, + "step": 712 + }, + { + "epoch": 0.7813698630136986, + "grad_norm": 1.6938892602920532, + "learning_rate": 4.140942323314379e-06, + "loss": 2.1508, + "step": 713 + }, + { + "epoch": 0.7824657534246575, + "grad_norm": 1.5774824619293213, + "learning_rate": 4.138911454102356e-06, + "loss": 2.1808, + "step": 714 + }, + { + "epoch": 0.7835616438356164, + "grad_norm": 1.8637802600860596, + "learning_rate": 4.136880584890333e-06, + "loss": 2.1822, + "step": 715 + }, + { + "epoch": 0.7846575342465754, + "grad_norm": 1.6216785907745361, + "learning_rate": 4.1348497156783106e-06, + "loss": 2.1807, + "step": 716 + }, + { + "epoch": 0.7857534246575343, + "grad_norm": 1.7688151597976685, + "learning_rate": 4.132818846466288e-06, + "loss": 2.1967, + "step": 717 + }, + { + "epoch": 0.7868493150684932, + "grad_norm": 1.6824712753295898, + "learning_rate": 4.130787977254265e-06, + "loss": 2.1952, + "step": 718 + }, + { + "epoch": 0.787945205479452, + "grad_norm": 1.911048412322998, + "learning_rate": 4.128757108042242e-06, + "loss": 2.2408, + "step": 719 + }, + { + "epoch": 0.7890410958904109, + "grad_norm": 1.62872314453125, + "learning_rate": 4.12672623883022e-06, + "loss": 2.1647, + "step": 720 + }, + { + "epoch": 0.7890410958904109, + "eval_loss": 2.3071532249450684, + "eval_runtime": 37.6838, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 720 + }, + { + "epoch": 0.7901369863013699, + "grad_norm": 2.080674171447754, + "learning_rate": 4.124695369618197e-06, + "loss": 2.1614, + "step": 721 + }, + { + "epoch": 0.7912328767123288, + "grad_norm": 1.7225316762924194, + "learning_rate": 4.122664500406174e-06, + "loss": 2.1784, + "step": 722 + }, + { + "epoch": 0.7923287671232877, + "grad_norm": 1.838957667350769, + "learning_rate": 4.1206336311941515e-06, + "loss": 2.1281, + "step": 723 + }, + { + "epoch": 0.7934246575342466, + "grad_norm": 1.9360777139663696, + "learning_rate": 4.118602761982129e-06, + "loss": 2.1168, + "step": 724 + }, + { + "epoch": 0.7945205479452054, + "grad_norm": 1.7712862491607666, + "learning_rate": 4.116571892770106e-06, + "loss": 2.2004, + "step": 725 + }, + { + "epoch": 0.7956164383561644, + "grad_norm": 1.6759192943572998, + "learning_rate": 4.114541023558083e-06, + "loss": 2.1588, + "step": 726 + }, + { + "epoch": 0.7967123287671233, + "grad_norm": 1.8396296501159668, + "learning_rate": 4.112510154346061e-06, + "loss": 2.1777, + "step": 727 + }, + { + "epoch": 0.7978082191780822, + "grad_norm": 1.6775752305984497, + "learning_rate": 4.110479285134038e-06, + "loss": 2.1923, + "step": 728 + }, + { + "epoch": 0.7989041095890411, + "grad_norm": 1.7111576795578003, + "learning_rate": 4.108448415922015e-06, + "loss": 2.1854, + "step": 729 + }, + { + "epoch": 0.8, + "grad_norm": 1.7400974035263062, + "learning_rate": 4.1064175467099924e-06, + "loss": 2.0897, + "step": 730 + }, + { + "epoch": 0.8, + "eval_loss": 2.3070170879364014, + "eval_runtime": 37.7602, + "eval_samples_per_second": 26.483, + "eval_steps_per_second": 1.112, + "step": 730 + }, + { + "epoch": 0.8010958904109589, + "grad_norm": 1.9166353940963745, + "learning_rate": 4.10438667749797e-06, + "loss": 2.2417, + "step": 731 + }, + { + "epoch": 0.8021917808219178, + "grad_norm": 1.7914725542068481, + "learning_rate": 4.102355808285947e-06, + "loss": 2.1677, + "step": 732 + }, + { + "epoch": 0.8032876712328767, + "grad_norm": 1.7602828741073608, + "learning_rate": 4.100324939073924e-06, + "loss": 2.1623, + "step": 733 + }, + { + "epoch": 0.8043835616438356, + "grad_norm": 1.6864640712738037, + "learning_rate": 4.0982940698619015e-06, + "loss": 2.1748, + "step": 734 + }, + { + "epoch": 0.8054794520547945, + "grad_norm": 1.564070463180542, + "learning_rate": 4.096263200649879e-06, + "loss": 2.1458, + "step": 735 + }, + { + "epoch": 0.8065753424657535, + "grad_norm": 1.9004014730453491, + "learning_rate": 4.094232331437856e-06, + "loss": 2.1536, + "step": 736 + }, + { + "epoch": 0.8076712328767123, + "grad_norm": 1.6670420169830322, + "learning_rate": 4.092201462225833e-06, + "loss": 2.1809, + "step": 737 + }, + { + "epoch": 0.8087671232876712, + "grad_norm": 1.6553101539611816, + "learning_rate": 4.090170593013811e-06, + "loss": 2.1343, + "step": 738 + }, + { + "epoch": 0.8098630136986301, + "grad_norm": 1.9083986282348633, + "learning_rate": 4.088139723801787e-06, + "loss": 2.1455, + "step": 739 + }, + { + "epoch": 0.810958904109589, + "grad_norm": 1.6069375276565552, + "learning_rate": 4.086108854589765e-06, + "loss": 2.1621, + "step": 740 + }, + { + "epoch": 0.810958904109589, + "eval_loss": 2.3045499324798584, + "eval_runtime": 37.7804, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 740 + }, + { + "epoch": 0.812054794520548, + "grad_norm": 1.6774693727493286, + "learning_rate": 4.084077985377742e-06, + "loss": 2.2079, + "step": 741 + }, + { + "epoch": 0.8131506849315069, + "grad_norm": 1.772355556488037, + "learning_rate": 4.08204711616572e-06, + "loss": 2.1276, + "step": 742 + }, + { + "epoch": 0.8142465753424658, + "grad_norm": 1.6998035907745361, + "learning_rate": 4.080016246953696e-06, + "loss": 2.1903, + "step": 743 + }, + { + "epoch": 0.8153424657534246, + "grad_norm": 1.850844383239746, + "learning_rate": 4.077985377741674e-06, + "loss": 2.1965, + "step": 744 + }, + { + "epoch": 0.8164383561643835, + "grad_norm": 1.5916879177093506, + "learning_rate": 4.075954508529651e-06, + "loss": 2.1982, + "step": 745 + }, + { + "epoch": 0.8175342465753425, + "grad_norm": 1.8195812702178955, + "learning_rate": 4.073923639317629e-06, + "loss": 2.1507, + "step": 746 + }, + { + "epoch": 0.8186301369863014, + "grad_norm": 1.5145621299743652, + "learning_rate": 4.071892770105605e-06, + "loss": 2.1236, + "step": 747 + }, + { + "epoch": 0.8197260273972603, + "grad_norm": 1.5750877857208252, + "learning_rate": 4.069861900893583e-06, + "loss": 2.142, + "step": 748 + }, + { + "epoch": 0.8208219178082192, + "grad_norm": 1.708666443824768, + "learning_rate": 4.06783103168156e-06, + "loss": 2.1577, + "step": 749 + }, + { + "epoch": 0.821917808219178, + "grad_norm": 1.5988762378692627, + "learning_rate": 4.065800162469538e-06, + "loss": 2.1795, + "step": 750 + }, + { + "epoch": 0.821917808219178, + "eval_loss": 2.3052196502685547, + "eval_runtime": 37.7983, + "eval_samples_per_second": 26.456, + "eval_steps_per_second": 1.111, + "step": 750 + }, + { + "epoch": 0.823013698630137, + "grad_norm": 1.608738899230957, + "learning_rate": 4.063769293257514e-06, + "loss": 2.1578, + "step": 751 + }, + { + "epoch": 0.8241095890410959, + "grad_norm": 1.5578334331512451, + "learning_rate": 4.061738424045492e-06, + "loss": 2.1724, + "step": 752 + }, + { + "epoch": 0.8252054794520548, + "grad_norm": 1.6957333087921143, + "learning_rate": 4.059707554833469e-06, + "loss": 2.1975, + "step": 753 + }, + { + "epoch": 0.8263013698630137, + "grad_norm": 1.691512107849121, + "learning_rate": 4.057676685621446e-06, + "loss": 2.1705, + "step": 754 + }, + { + "epoch": 0.8273972602739726, + "grad_norm": 1.6061747074127197, + "learning_rate": 4.0556458164094235e-06, + "loss": 2.1031, + "step": 755 + }, + { + "epoch": 0.8284931506849315, + "grad_norm": 1.6074447631835938, + "learning_rate": 4.053614947197401e-06, + "loss": 2.1738, + "step": 756 + }, + { + "epoch": 0.8295890410958904, + "grad_norm": 1.9722917079925537, + "learning_rate": 4.051584077985378e-06, + "loss": 2.2043, + "step": 757 + }, + { + "epoch": 0.8306849315068493, + "grad_norm": 1.597692847251892, + "learning_rate": 4.049553208773355e-06, + "loss": 2.1621, + "step": 758 + }, + { + "epoch": 0.8317808219178082, + "grad_norm": 1.85458242893219, + "learning_rate": 4.047522339561333e-06, + "loss": 2.1504, + "step": 759 + }, + { + "epoch": 0.8328767123287671, + "grad_norm": 1.8328344821929932, + "learning_rate": 4.04549147034931e-06, + "loss": 2.2541, + "step": 760 + }, + { + "epoch": 0.8328767123287671, + "eval_loss": 2.3033809661865234, + "eval_runtime": 37.8088, + "eval_samples_per_second": 26.449, + "eval_steps_per_second": 1.111, + "step": 760 + }, + { + "epoch": 0.8339726027397261, + "grad_norm": 1.6544418334960938, + "learning_rate": 4.043460601137287e-06, + "loss": 2.1605, + "step": 761 + }, + { + "epoch": 0.8350684931506849, + "grad_norm": 1.9443641901016235, + "learning_rate": 4.041429731925264e-06, + "loss": 2.1621, + "step": 762 + }, + { + "epoch": 0.8361643835616438, + "grad_norm": 1.731277346611023, + "learning_rate": 4.039398862713242e-06, + "loss": 2.2168, + "step": 763 + }, + { + "epoch": 0.8372602739726027, + "grad_norm": 1.671072244644165, + "learning_rate": 4.037367993501219e-06, + "loss": 2.1612, + "step": 764 + }, + { + "epoch": 0.8383561643835616, + "grad_norm": 1.8777449131011963, + "learning_rate": 4.035337124289196e-06, + "loss": 2.1339, + "step": 765 + }, + { + "epoch": 0.8394520547945206, + "grad_norm": 1.5941543579101562, + "learning_rate": 4.0333062550771735e-06, + "loss": 2.1545, + "step": 766 + }, + { + "epoch": 0.8405479452054795, + "grad_norm": 1.9245301485061646, + "learning_rate": 4.031275385865151e-06, + "loss": 2.187, + "step": 767 + }, + { + "epoch": 0.8416438356164384, + "grad_norm": 1.69217848777771, + "learning_rate": 4.029244516653128e-06, + "loss": 2.1168, + "step": 768 + }, + { + "epoch": 0.8427397260273972, + "grad_norm": 1.7042242288589478, + "learning_rate": 4.027213647441105e-06, + "loss": 2.1353, + "step": 769 + }, + { + "epoch": 0.8438356164383561, + "grad_norm": 1.6391927003860474, + "learning_rate": 4.025182778229083e-06, + "loss": 2.1568, + "step": 770 + }, + { + "epoch": 0.8438356164383561, + "eval_loss": 2.310056686401367, + "eval_runtime": 37.7573, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 770 + }, + { + "epoch": 0.8449315068493151, + "grad_norm": 1.7833209037780762, + "learning_rate": 4.02315190901706e-06, + "loss": 2.1664, + "step": 771 + }, + { + "epoch": 0.846027397260274, + "grad_norm": 1.8203891515731812, + "learning_rate": 4.021121039805037e-06, + "loss": 2.2342, + "step": 772 + }, + { + "epoch": 0.8471232876712329, + "grad_norm": 1.572940707206726, + "learning_rate": 4.0190901705930144e-06, + "loss": 2.1418, + "step": 773 + }, + { + "epoch": 0.8482191780821918, + "grad_norm": 1.6871955394744873, + "learning_rate": 4.017059301380992e-06, + "loss": 2.2023, + "step": 774 + }, + { + "epoch": 0.8493150684931506, + "grad_norm": 1.6343133449554443, + "learning_rate": 4.015028432168969e-06, + "loss": 2.1302, + "step": 775 + }, + { + "epoch": 0.8504109589041096, + "grad_norm": 1.4492065906524658, + "learning_rate": 4.012997562956945e-06, + "loss": 2.1427, + "step": 776 + }, + { + "epoch": 0.8515068493150685, + "grad_norm": 1.6599146127700806, + "learning_rate": 4.0109666937449235e-06, + "loss": 2.1479, + "step": 777 + }, + { + "epoch": 0.8526027397260274, + "grad_norm": 1.6836353540420532, + "learning_rate": 4.0089358245329e-06, + "loss": 2.148, + "step": 778 + }, + { + "epoch": 0.8536986301369863, + "grad_norm": 1.5977990627288818, + "learning_rate": 4.006904955320878e-06, + "loss": 2.1946, + "step": 779 + }, + { + "epoch": 0.8547945205479452, + "grad_norm": 1.6259957551956177, + "learning_rate": 4.0048740861088545e-06, + "loss": 2.1761, + "step": 780 + }, + { + "epoch": 0.8547945205479452, + "eval_loss": 2.303910970687866, + "eval_runtime": 37.7728, + "eval_samples_per_second": 26.474, + "eval_steps_per_second": 1.112, + "step": 780 + }, + { + "epoch": 0.8558904109589041, + "grad_norm": 1.6990300416946411, + "learning_rate": 4.002843216896833e-06, + "loss": 2.1309, + "step": 781 + }, + { + "epoch": 0.856986301369863, + "grad_norm": 1.6556029319763184, + "learning_rate": 4.000812347684809e-06, + "loss": 2.151, + "step": 782 + }, + { + "epoch": 0.8580821917808219, + "grad_norm": 1.5659290552139282, + "learning_rate": 3.998781478472787e-06, + "loss": 2.1317, + "step": 783 + }, + { + "epoch": 0.8591780821917808, + "grad_norm": 1.6553704738616943, + "learning_rate": 3.996750609260764e-06, + "loss": 2.2344, + "step": 784 + }, + { + "epoch": 0.8602739726027397, + "grad_norm": 1.6608418226242065, + "learning_rate": 3.994719740048742e-06, + "loss": 2.1807, + "step": 785 + }, + { + "epoch": 0.8613698630136987, + "grad_norm": 1.5862030982971191, + "learning_rate": 3.992688870836718e-06, + "loss": 2.1351, + "step": 786 + }, + { + "epoch": 0.8624657534246576, + "grad_norm": 1.767526388168335, + "learning_rate": 3.990658001624696e-06, + "loss": 2.1856, + "step": 787 + }, + { + "epoch": 0.8635616438356164, + "grad_norm": 1.6902577877044678, + "learning_rate": 3.988627132412673e-06, + "loss": 2.1592, + "step": 788 + }, + { + "epoch": 0.8646575342465753, + "grad_norm": 1.7760623693466187, + "learning_rate": 3.98659626320065e-06, + "loss": 2.1645, + "step": 789 + }, + { + "epoch": 0.8657534246575342, + "grad_norm": 1.6791038513183594, + "learning_rate": 3.984565393988627e-06, + "loss": 2.1611, + "step": 790 + }, + { + "epoch": 0.8657534246575342, + "eval_loss": 2.3027942180633545, + "eval_runtime": 37.736, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 790 + }, + { + "epoch": 0.8668493150684932, + "grad_norm": 1.6695773601531982, + "learning_rate": 3.9825345247766046e-06, + "loss": 2.1747, + "step": 791 + }, + { + "epoch": 0.8679452054794521, + "grad_norm": 1.8736445903778076, + "learning_rate": 3.980503655564582e-06, + "loss": 2.1806, + "step": 792 + }, + { + "epoch": 0.869041095890411, + "grad_norm": 1.6695518493652344, + "learning_rate": 3.978472786352559e-06, + "loss": 2.1631, + "step": 793 + }, + { + "epoch": 0.8701369863013698, + "grad_norm": 1.6782366037368774, + "learning_rate": 3.976441917140536e-06, + "loss": 2.2175, + "step": 794 + }, + { + "epoch": 0.8712328767123287, + "grad_norm": 1.678619623184204, + "learning_rate": 3.974411047928514e-06, + "loss": 2.1883, + "step": 795 + }, + { + "epoch": 0.8723287671232877, + "grad_norm": 1.6274909973144531, + "learning_rate": 3.972380178716491e-06, + "loss": 2.1866, + "step": 796 + }, + { + "epoch": 0.8734246575342466, + "grad_norm": 1.5609477758407593, + "learning_rate": 3.970349309504468e-06, + "loss": 2.1257, + "step": 797 + }, + { + "epoch": 0.8745205479452055, + "grad_norm": 1.7419973611831665, + "learning_rate": 3.9683184402924455e-06, + "loss": 2.1253, + "step": 798 + }, + { + "epoch": 0.8756164383561644, + "grad_norm": 1.6296330690383911, + "learning_rate": 3.966287571080423e-06, + "loss": 2.1755, + "step": 799 + }, + { + "epoch": 0.8767123287671232, + "grad_norm": 1.5995864868164062, + "learning_rate": 3.9642567018684e-06, + "loss": 2.169, + "step": 800 + }, + { + "epoch": 0.8767123287671232, + "eval_loss": 2.301098108291626, + "eval_runtime": 37.7893, + "eval_samples_per_second": 26.463, + "eval_steps_per_second": 1.111, + "step": 800 + }, + { + "epoch": 0.8778082191780822, + "grad_norm": 1.653707504272461, + "learning_rate": 3.962225832656377e-06, + "loss": 2.115, + "step": 801 + }, + { + "epoch": 0.8789041095890411, + "grad_norm": 1.5908581018447876, + "learning_rate": 3.960194963444355e-06, + "loss": 2.1659, + "step": 802 + }, + { + "epoch": 0.88, + "grad_norm": 1.6865545511245728, + "learning_rate": 3.958164094232332e-06, + "loss": 2.1666, + "step": 803 + }, + { + "epoch": 0.8810958904109589, + "grad_norm": 1.6234707832336426, + "learning_rate": 3.956133225020309e-06, + "loss": 2.1594, + "step": 804 + }, + { + "epoch": 0.8821917808219178, + "grad_norm": 1.7375986576080322, + "learning_rate": 3.954102355808286e-06, + "loss": 2.1596, + "step": 805 + }, + { + "epoch": 0.8832876712328767, + "grad_norm": 1.6529247760772705, + "learning_rate": 3.952071486596264e-06, + "loss": 2.1495, + "step": 806 + }, + { + "epoch": 0.8843835616438356, + "grad_norm": 1.517357349395752, + "learning_rate": 3.950040617384241e-06, + "loss": 2.1394, + "step": 807 + }, + { + "epoch": 0.8854794520547945, + "grad_norm": 1.7367429733276367, + "learning_rate": 3.948009748172218e-06, + "loss": 2.206, + "step": 808 + }, + { + "epoch": 0.8865753424657534, + "grad_norm": 1.5467783212661743, + "learning_rate": 3.9459788789601955e-06, + "loss": 2.1672, + "step": 809 + }, + { + "epoch": 0.8876712328767123, + "grad_norm": 1.5647791624069214, + "learning_rate": 3.943948009748173e-06, + "loss": 2.1695, + "step": 810 + }, + { + "epoch": 0.8876712328767123, + "eval_loss": 2.2975995540618896, + "eval_runtime": 37.4912, + "eval_samples_per_second": 26.673, + "eval_steps_per_second": 1.12, + "step": 810 + }, + { + "epoch": 0.8887671232876713, + "grad_norm": 1.597876787185669, + "learning_rate": 3.94191714053615e-06, + "loss": 2.1808, + "step": 811 + }, + { + "epoch": 0.8898630136986302, + "grad_norm": 1.7645189762115479, + "learning_rate": 3.939886271324127e-06, + "loss": 2.1291, + "step": 812 + }, + { + "epoch": 0.890958904109589, + "grad_norm": 1.5209035873413086, + "learning_rate": 3.937855402112104e-06, + "loss": 2.1269, + "step": 813 + }, + { + "epoch": 0.8920547945205479, + "grad_norm": 1.5804905891418457, + "learning_rate": 3.935824532900082e-06, + "loss": 2.1751, + "step": 814 + }, + { + "epoch": 0.8931506849315068, + "grad_norm": 1.687865972518921, + "learning_rate": 3.933793663688058e-06, + "loss": 2.1917, + "step": 815 + }, + { + "epoch": 0.8942465753424658, + "grad_norm": 1.6767957210540771, + "learning_rate": 3.9317627944760365e-06, + "loss": 2.118, + "step": 816 + }, + { + "epoch": 0.8953424657534247, + "grad_norm": 1.657046914100647, + "learning_rate": 3.929731925264013e-06, + "loss": 2.1479, + "step": 817 + }, + { + "epoch": 0.8964383561643836, + "grad_norm": 1.5638110637664795, + "learning_rate": 3.927701056051991e-06, + "loss": 2.1565, + "step": 818 + }, + { + "epoch": 0.8975342465753424, + "grad_norm": 1.6865980625152588, + "learning_rate": 3.9256701868399674e-06, + "loss": 2.1382, + "step": 819 + }, + { + "epoch": 0.8986301369863013, + "grad_norm": 1.6738817691802979, + "learning_rate": 3.9236393176279456e-06, + "loss": 2.1257, + "step": 820 + }, + { + "epoch": 0.8986301369863013, + "eval_loss": 2.2953474521636963, + "eval_runtime": 37.7016, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 820 + }, + { + "epoch": 0.8997260273972603, + "grad_norm": 1.568678617477417, + "learning_rate": 3.921608448415922e-06, + "loss": 2.169, + "step": 821 + }, + { + "epoch": 0.9008219178082192, + "grad_norm": 1.6803818941116333, + "learning_rate": 3.9195775792039e-06, + "loss": 2.1489, + "step": 822 + }, + { + "epoch": 0.9019178082191781, + "grad_norm": 1.7010858058929443, + "learning_rate": 3.9175467099918765e-06, + "loss": 2.1846, + "step": 823 + }, + { + "epoch": 0.903013698630137, + "grad_norm": 1.5782372951507568, + "learning_rate": 3.915515840779855e-06, + "loss": 2.1788, + "step": 824 + }, + { + "epoch": 0.9041095890410958, + "grad_norm": 1.7804685831069946, + "learning_rate": 3.913484971567831e-06, + "loss": 2.1058, + "step": 825 + }, + { + "epoch": 0.9052054794520548, + "grad_norm": 1.5909292697906494, + "learning_rate": 3.911454102355808e-06, + "loss": 2.1716, + "step": 826 + }, + { + "epoch": 0.9063013698630137, + "grad_norm": 1.7001307010650635, + "learning_rate": 3.909423233143786e-06, + "loss": 2.1647, + "step": 827 + }, + { + "epoch": 0.9073972602739726, + "grad_norm": 1.5576646327972412, + "learning_rate": 3.907392363931763e-06, + "loss": 2.2174, + "step": 828 + }, + { + "epoch": 0.9084931506849315, + "grad_norm": 1.6621390581130981, + "learning_rate": 3.90536149471974e-06, + "loss": 2.2344, + "step": 829 + }, + { + "epoch": 0.9095890410958904, + "grad_norm": 1.592091679573059, + "learning_rate": 3.9033306255077175e-06, + "loss": 2.2501, + "step": 830 + }, + { + "epoch": 0.9095890410958904, + "eval_loss": 2.294229507446289, + "eval_runtime": 37.7195, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.113, + "step": 830 + }, + { + "epoch": 0.9106849315068493, + "grad_norm": 1.5698535442352295, + "learning_rate": 3.901299756295695e-06, + "loss": 2.1534, + "step": 831 + }, + { + "epoch": 0.9117808219178082, + "grad_norm": 1.7159922122955322, + "learning_rate": 3.899268887083672e-06, + "loss": 2.1601, + "step": 832 + }, + { + "epoch": 0.9128767123287671, + "grad_norm": 1.5601210594177246, + "learning_rate": 3.897238017871649e-06, + "loss": 2.1598, + "step": 833 + }, + { + "epoch": 0.913972602739726, + "grad_norm": 1.5987435579299927, + "learning_rate": 3.8952071486596266e-06, + "loss": 2.1307, + "step": 834 + }, + { + "epoch": 0.915068493150685, + "grad_norm": 1.7568409442901611, + "learning_rate": 3.893176279447604e-06, + "loss": 2.2209, + "step": 835 + }, + { + "epoch": 0.9161643835616439, + "grad_norm": 1.5044634342193604, + "learning_rate": 3.891145410235581e-06, + "loss": 2.1699, + "step": 836 + }, + { + "epoch": 0.9172602739726028, + "grad_norm": 1.7287421226501465, + "learning_rate": 3.889114541023558e-06, + "loss": 2.1489, + "step": 837 + }, + { + "epoch": 0.9183561643835616, + "grad_norm": 1.6595203876495361, + "learning_rate": 3.887083671811536e-06, + "loss": 2.1167, + "step": 838 + }, + { + "epoch": 0.9194520547945205, + "grad_norm": 1.5620248317718506, + "learning_rate": 3.885052802599513e-06, + "loss": 2.2119, + "step": 839 + }, + { + "epoch": 0.9205479452054794, + "grad_norm": 1.6492353677749634, + "learning_rate": 3.88302193338749e-06, + "loss": 2.1725, + "step": 840 + }, + { + "epoch": 0.9205479452054794, + "eval_loss": 2.2956349849700928, + "eval_runtime": 37.779, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 840 + }, + { + "epoch": 0.9216438356164384, + "grad_norm": 1.6601520776748657, + "learning_rate": 3.8809910641754675e-06, + "loss": 2.1532, + "step": 841 + }, + { + "epoch": 0.9227397260273973, + "grad_norm": 1.6723994016647339, + "learning_rate": 3.878960194963445e-06, + "loss": 2.1931, + "step": 842 + }, + { + "epoch": 0.9238356164383562, + "grad_norm": 1.7594473361968994, + "learning_rate": 3.876929325751422e-06, + "loss": 2.1219, + "step": 843 + }, + { + "epoch": 0.924931506849315, + "grad_norm": 1.546398401260376, + "learning_rate": 3.874898456539399e-06, + "loss": 2.1679, + "step": 844 + }, + { + "epoch": 0.9260273972602739, + "grad_norm": 1.8171741962432861, + "learning_rate": 3.872867587327377e-06, + "loss": 2.1434, + "step": 845 + }, + { + "epoch": 0.9271232876712329, + "grad_norm": 1.703196406364441, + "learning_rate": 3.870836718115354e-06, + "loss": 2.1574, + "step": 846 + }, + { + "epoch": 0.9282191780821918, + "grad_norm": 1.6290628910064697, + "learning_rate": 3.868805848903331e-06, + "loss": 2.1657, + "step": 847 + }, + { + "epoch": 0.9293150684931507, + "grad_norm": 1.5868806838989258, + "learning_rate": 3.8667749796913084e-06, + "loss": 2.1659, + "step": 848 + }, + { + "epoch": 0.9304109589041096, + "grad_norm": 1.551581621170044, + "learning_rate": 3.864744110479286e-06, + "loss": 2.1916, + "step": 849 + }, + { + "epoch": 0.9315068493150684, + "grad_norm": 1.6634963750839233, + "learning_rate": 3.862713241267262e-06, + "loss": 2.2027, + "step": 850 + }, + { + "epoch": 0.9315068493150684, + "eval_loss": 2.2988946437835693, + "eval_runtime": 37.7841, + "eval_samples_per_second": 26.466, + "eval_steps_per_second": 1.112, + "step": 850 + }, + { + "epoch": 0.9326027397260274, + "grad_norm": 1.6138510704040527, + "learning_rate": 3.86068237205524e-06, + "loss": 2.1634, + "step": 851 + }, + { + "epoch": 0.9336986301369863, + "grad_norm": 1.6294615268707275, + "learning_rate": 3.858651502843217e-06, + "loss": 2.1175, + "step": 852 + }, + { + "epoch": 0.9347945205479452, + "grad_norm": 1.7515549659729004, + "learning_rate": 3.856620633631195e-06, + "loss": 2.1562, + "step": 853 + }, + { + "epoch": 0.9358904109589041, + "grad_norm": 1.6304186582565308, + "learning_rate": 3.854589764419171e-06, + "loss": 2.2256, + "step": 854 + }, + { + "epoch": 0.936986301369863, + "grad_norm": 1.6189597845077515, + "learning_rate": 3.852558895207149e-06, + "loss": 2.1742, + "step": 855 + }, + { + "epoch": 0.938082191780822, + "grad_norm": 1.665515661239624, + "learning_rate": 3.850528025995126e-06, + "loss": 2.1481, + "step": 856 + }, + { + "epoch": 0.9391780821917808, + "grad_norm": 1.7066419124603271, + "learning_rate": 3.848497156783104e-06, + "loss": 2.1522, + "step": 857 + }, + { + "epoch": 0.9402739726027397, + "grad_norm": 1.884599208831787, + "learning_rate": 3.84646628757108e-06, + "loss": 2.1493, + "step": 858 + }, + { + "epoch": 0.9413698630136986, + "grad_norm": 1.6224873065948486, + "learning_rate": 3.8444354183590585e-06, + "loss": 2.1461, + "step": 859 + }, + { + "epoch": 0.9424657534246575, + "grad_norm": 1.916416049003601, + "learning_rate": 3.842404549147035e-06, + "loss": 2.1265, + "step": 860 + }, + { + "epoch": 0.9424657534246575, + "eval_loss": 2.297154426574707, + "eval_runtime": 37.77, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 860 + }, + { + "epoch": 0.9435616438356165, + "grad_norm": 1.5281857252120972, + "learning_rate": 3.840373679935013e-06, + "loss": 2.1476, + "step": 861 + }, + { + "epoch": 0.9446575342465754, + "grad_norm": 1.7905207872390747, + "learning_rate": 3.8383428107229894e-06, + "loss": 2.1421, + "step": 862 + }, + { + "epoch": 0.9457534246575342, + "grad_norm": 1.8282849788665771, + "learning_rate": 3.836311941510967e-06, + "loss": 2.1807, + "step": 863 + }, + { + "epoch": 0.9468493150684931, + "grad_norm": 1.7861522436141968, + "learning_rate": 3.834281072298944e-06, + "loss": 2.1337, + "step": 864 + }, + { + "epoch": 0.947945205479452, + "grad_norm": 1.6107250452041626, + "learning_rate": 3.832250203086921e-06, + "loss": 2.1911, + "step": 865 + }, + { + "epoch": 0.949041095890411, + "grad_norm": 1.657909631729126, + "learning_rate": 3.8302193338748985e-06, + "loss": 2.2257, + "step": 866 + }, + { + "epoch": 0.9501369863013699, + "grad_norm": 1.8969722986221313, + "learning_rate": 3.828188464662876e-06, + "loss": 2.1778, + "step": 867 + }, + { + "epoch": 0.9512328767123288, + "grad_norm": 1.6498360633850098, + "learning_rate": 3.826157595450853e-06, + "loss": 2.127, + "step": 868 + }, + { + "epoch": 0.9523287671232876, + "grad_norm": 1.5750585794448853, + "learning_rate": 3.82412672623883e-06, + "loss": 2.1716, + "step": 869 + }, + { + "epoch": 0.9534246575342465, + "grad_norm": 1.6047886610031128, + "learning_rate": 3.822095857026808e-06, + "loss": 2.1354, + "step": 870 + }, + { + "epoch": 0.9534246575342465, + "eval_loss": 2.295785665512085, + "eval_runtime": 37.7419, + "eval_samples_per_second": 26.496, + "eval_steps_per_second": 1.113, + "step": 870 + }, + { + "epoch": 0.9545205479452055, + "grad_norm": 1.6426187753677368, + "learning_rate": 3.820064987814785e-06, + "loss": 2.1375, + "step": 871 + }, + { + "epoch": 0.9556164383561644, + "grad_norm": 1.5118632316589355, + "learning_rate": 3.818034118602762e-06, + "loss": 2.1491, + "step": 872 + }, + { + "epoch": 0.9567123287671233, + "grad_norm": 1.630476474761963, + "learning_rate": 3.8160032493907395e-06, + "loss": 2.1497, + "step": 873 + }, + { + "epoch": 0.9578082191780822, + "grad_norm": 1.6820838451385498, + "learning_rate": 3.813972380178717e-06, + "loss": 2.1746, + "step": 874 + }, + { + "epoch": 0.958904109589041, + "grad_norm": 1.5480482578277588, + "learning_rate": 3.811941510966694e-06, + "loss": 2.1226, + "step": 875 + }, + { + "epoch": 0.96, + "grad_norm": 1.601493239402771, + "learning_rate": 3.809910641754671e-06, + "loss": 2.0934, + "step": 876 + }, + { + "epoch": 0.9610958904109589, + "grad_norm": 1.5756622552871704, + "learning_rate": 3.8078797725426486e-06, + "loss": 2.1679, + "step": 877 + }, + { + "epoch": 0.9621917808219178, + "grad_norm": 1.6811070442199707, + "learning_rate": 3.8058489033306254e-06, + "loss": 2.1407, + "step": 878 + }, + { + "epoch": 0.9632876712328767, + "grad_norm": 1.5748940706253052, + "learning_rate": 3.803818034118603e-06, + "loss": 2.1985, + "step": 879 + }, + { + "epoch": 0.9643835616438357, + "grad_norm": 1.5412342548370361, + "learning_rate": 3.80178716490658e-06, + "loss": 2.1876, + "step": 880 + }, + { + "epoch": 0.9643835616438357, + "eval_loss": 2.291630268096924, + "eval_runtime": 37.7702, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 880 + }, + { + "epoch": 0.9654794520547946, + "grad_norm": 1.7080684900283813, + "learning_rate": 3.7997562956945577e-06, + "loss": 2.1453, + "step": 881 + }, + { + "epoch": 0.9665753424657534, + "grad_norm": 1.8705110549926758, + "learning_rate": 3.7977254264825345e-06, + "loss": 2.1298, + "step": 882 + }, + { + "epoch": 0.9676712328767123, + "grad_norm": 1.7332537174224854, + "learning_rate": 3.7956945572705122e-06, + "loss": 2.2117, + "step": 883 + }, + { + "epoch": 0.9687671232876712, + "grad_norm": 1.88144052028656, + "learning_rate": 3.793663688058489e-06, + "loss": 2.2005, + "step": 884 + }, + { + "epoch": 0.9698630136986301, + "grad_norm": 1.6263943910598755, + "learning_rate": 3.791632818846467e-06, + "loss": 2.1853, + "step": 885 + }, + { + "epoch": 0.9709589041095891, + "grad_norm": 1.5968067646026611, + "learning_rate": 3.7896019496344436e-06, + "loss": 2.1998, + "step": 886 + }, + { + "epoch": 0.972054794520548, + "grad_norm": 1.7149827480316162, + "learning_rate": 3.787571080422421e-06, + "loss": 2.1855, + "step": 887 + }, + { + "epoch": 0.9731506849315068, + "grad_norm": 1.7375026941299438, + "learning_rate": 3.785540211210398e-06, + "loss": 2.1664, + "step": 888 + }, + { + "epoch": 0.9742465753424657, + "grad_norm": 1.6411423683166504, + "learning_rate": 3.7835093419983755e-06, + "loss": 2.1492, + "step": 889 + }, + { + "epoch": 0.9753424657534246, + "grad_norm": 1.6208266019821167, + "learning_rate": 3.7814784727863527e-06, + "loss": 2.2341, + "step": 890 + }, + { + "epoch": 0.9753424657534246, + "eval_loss": 2.2929985523223877, + "eval_runtime": 37.6693, + "eval_samples_per_second": 26.547, + "eval_steps_per_second": 1.115, + "step": 890 + }, + { + "epoch": 0.9764383561643836, + "grad_norm": 1.6871188879013062, + "learning_rate": 3.77944760357433e-06, + "loss": 2.1985, + "step": 891 + }, + { + "epoch": 0.9775342465753425, + "grad_norm": 1.8475587368011475, + "learning_rate": 3.7774167343623073e-06, + "loss": 2.1423, + "step": 892 + }, + { + "epoch": 0.9786301369863014, + "grad_norm": 1.6115065813064575, + "learning_rate": 3.7753858651502846e-06, + "loss": 2.1944, + "step": 893 + }, + { + "epoch": 0.9797260273972602, + "grad_norm": 1.6232231855392456, + "learning_rate": 3.773354995938262e-06, + "loss": 2.1751, + "step": 894 + }, + { + "epoch": 0.9808219178082191, + "grad_norm": 1.8564099073410034, + "learning_rate": 3.771324126726239e-06, + "loss": 2.2021, + "step": 895 + }, + { + "epoch": 0.9819178082191781, + "grad_norm": 1.7270197868347168, + "learning_rate": 3.7692932575142164e-06, + "loss": 2.1351, + "step": 896 + }, + { + "epoch": 0.983013698630137, + "grad_norm": 1.7187668085098267, + "learning_rate": 3.7672623883021937e-06, + "loss": 2.0856, + "step": 897 + }, + { + "epoch": 0.9841095890410959, + "grad_norm": 1.7766387462615967, + "learning_rate": 3.765231519090171e-06, + "loss": 2.168, + "step": 898 + }, + { + "epoch": 0.9852054794520548, + "grad_norm": 1.8063374757766724, + "learning_rate": 3.7632006498781482e-06, + "loss": 2.1317, + "step": 899 + }, + { + "epoch": 0.9863013698630136, + "grad_norm": 1.7601871490478516, + "learning_rate": 3.761169780666125e-06, + "loss": 2.1742, + "step": 900 + }, + { + "epoch": 0.9863013698630136, + "eval_loss": 2.292574167251587, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 900 + }, + { + "epoch": 0.9873972602739726, + "grad_norm": 1.8501280546188354, + "learning_rate": 3.7591389114541028e-06, + "loss": 2.1856, + "step": 901 + }, + { + "epoch": 0.9884931506849315, + "grad_norm": 1.8786728382110596, + "learning_rate": 3.7571080422420796e-06, + "loss": 2.1432, + "step": 902 + }, + { + "epoch": 0.9895890410958904, + "grad_norm": 1.7662389278411865, + "learning_rate": 3.7550771730300573e-06, + "loss": 2.2011, + "step": 903 + }, + { + "epoch": 0.9906849315068493, + "grad_norm": 1.6810210943222046, + "learning_rate": 3.753046303818034e-06, + "loss": 2.1719, + "step": 904 + }, + { + "epoch": 0.9917808219178083, + "grad_norm": 2.034982681274414, + "learning_rate": 3.751015434606012e-06, + "loss": 2.2206, + "step": 905 + }, + { + "epoch": 0.9928767123287672, + "grad_norm": 1.804049015045166, + "learning_rate": 3.7489845653939887e-06, + "loss": 2.1248, + "step": 906 + }, + { + "epoch": 0.993972602739726, + "grad_norm": 1.6039106845855713, + "learning_rate": 3.7469536961819664e-06, + "loss": 2.1545, + "step": 907 + }, + { + "epoch": 0.9950684931506849, + "grad_norm": 1.9352415800094604, + "learning_rate": 3.7449228269699433e-06, + "loss": 2.1842, + "step": 908 + }, + { + "epoch": 0.9961643835616438, + "grad_norm": 1.5759649276733398, + "learning_rate": 3.742891957757921e-06, + "loss": 2.1588, + "step": 909 + }, + { + "epoch": 0.9972602739726028, + "grad_norm": 1.6169486045837402, + "learning_rate": 3.740861088545898e-06, + "loss": 2.1871, + "step": 910 + }, + { + "epoch": 0.9972602739726028, + "eval_loss": 2.2923426628112793, + "eval_runtime": 37.4584, + "eval_samples_per_second": 26.696, + "eval_steps_per_second": 1.121, + "step": 910 + }, + { + "epoch": 0.9983561643835617, + "grad_norm": 1.748138427734375, + "learning_rate": 3.7388302193338755e-06, + "loss": 2.144, + "step": 911 + }, + { + "epoch": 0.9994520547945206, + "grad_norm": 1.8182991743087769, + "learning_rate": 3.7367993501218524e-06, + "loss": 2.1474, + "step": 912 + }, + { + "epoch": 1.0005479452054795, + "grad_norm": 1.6532196998596191, + "learning_rate": 3.7347684809098297e-06, + "loss": 2.1146, + "step": 913 + }, + { + "epoch": 1.0016438356164383, + "grad_norm": 2.0338973999023438, + "learning_rate": 3.732737611697807e-06, + "loss": 2.0851, + "step": 914 + }, + { + "epoch": 1.0027397260273974, + "grad_norm": 1.71133553981781, + "learning_rate": 3.7307067424857842e-06, + "loss": 2.1197, + "step": 915 + }, + { + "epoch": 1.0038356164383562, + "grad_norm": 1.6081736087799072, + "learning_rate": 3.7286758732737615e-06, + "loss": 2.1106, + "step": 916 + }, + { + "epoch": 1.004931506849315, + "grad_norm": 2.100734233856201, + "learning_rate": 3.7266450040617388e-06, + "loss": 2.2125, + "step": 917 + }, + { + "epoch": 1.006027397260274, + "grad_norm": 1.758898377418518, + "learning_rate": 3.724614134849716e-06, + "loss": 2.0985, + "step": 918 + }, + { + "epoch": 1.0071232876712328, + "grad_norm": 1.7335586547851562, + "learning_rate": 3.7225832656376933e-06, + "loss": 2.1911, + "step": 919 + }, + { + "epoch": 1.0082191780821919, + "grad_norm": 1.7879023551940918, + "learning_rate": 3.7205523964256706e-06, + "loss": 2.1632, + "step": 920 + }, + { + "epoch": 1.0082191780821919, + "eval_loss": 2.294053554534912, + "eval_runtime": 37.6132, + "eval_samples_per_second": 26.586, + "eval_steps_per_second": 1.117, + "step": 920 + }, + { + "epoch": 1.0093150684931507, + "grad_norm": 1.7201987504959106, + "learning_rate": 3.718521527213648e-06, + "loss": 2.1501, + "step": 921 + }, + { + "epoch": 1.0104109589041095, + "grad_norm": 1.7638986110687256, + "learning_rate": 3.716490658001625e-06, + "loss": 2.0661, + "step": 922 + }, + { + "epoch": 1.0115068493150685, + "grad_norm": 1.7376288175582886, + "learning_rate": 3.7144597887896024e-06, + "loss": 2.1613, + "step": 923 + }, + { + "epoch": 1.0126027397260273, + "grad_norm": 1.9208155870437622, + "learning_rate": 3.7124289195775797e-06, + "loss": 2.1562, + "step": 924 + }, + { + "epoch": 1.0136986301369864, + "grad_norm": 1.6709166765213013, + "learning_rate": 3.710398050365557e-06, + "loss": 2.1309, + "step": 925 + }, + { + "epoch": 1.0147945205479452, + "grad_norm": 1.7070904970169067, + "learning_rate": 3.708367181153534e-06, + "loss": 2.1538, + "step": 926 + }, + { + "epoch": 1.0158904109589042, + "grad_norm": 1.7832821607589722, + "learning_rate": 3.7063363119415115e-06, + "loss": 2.1262, + "step": 927 + }, + { + "epoch": 1.016986301369863, + "grad_norm": 1.8703486919403076, + "learning_rate": 3.7043054427294884e-06, + "loss": 2.1142, + "step": 928 + }, + { + "epoch": 1.0180821917808218, + "grad_norm": 1.692858338356018, + "learning_rate": 3.702274573517466e-06, + "loss": 2.1182, + "step": 929 + }, + { + "epoch": 1.0191780821917809, + "grad_norm": 1.7156164646148682, + "learning_rate": 3.700243704305443e-06, + "loss": 2.1189, + "step": 930 + }, + { + "epoch": 1.0191780821917809, + "eval_loss": 2.2969048023223877, + "eval_runtime": 37.6673, + "eval_samples_per_second": 26.548, + "eval_steps_per_second": 1.115, + "step": 930 + }, + { + "epoch": 1.0202739726027397, + "grad_norm": 1.7443044185638428, + "learning_rate": 3.6982128350934206e-06, + "loss": 2.1058, + "step": 931 + }, + { + "epoch": 1.0213698630136987, + "grad_norm": 2.00412917137146, + "learning_rate": 3.6961819658813975e-06, + "loss": 2.2071, + "step": 932 + }, + { + "epoch": 1.0224657534246575, + "grad_norm": 1.7170531749725342, + "learning_rate": 3.694151096669375e-06, + "loss": 2.1254, + "step": 933 + }, + { + "epoch": 1.0235616438356165, + "grad_norm": 1.679018497467041, + "learning_rate": 3.692120227457352e-06, + "loss": 2.1487, + "step": 934 + }, + { + "epoch": 1.0246575342465754, + "grad_norm": 1.8477305173873901, + "learning_rate": 3.6900893582453297e-06, + "loss": 2.062, + "step": 935 + }, + { + "epoch": 1.0257534246575342, + "grad_norm": 1.664095401763916, + "learning_rate": 3.6880584890333066e-06, + "loss": 2.1418, + "step": 936 + }, + { + "epoch": 1.0268493150684932, + "grad_norm": 1.8831602334976196, + "learning_rate": 3.6860276198212834e-06, + "loss": 2.1483, + "step": 937 + }, + { + "epoch": 1.027945205479452, + "grad_norm": 1.7625365257263184, + "learning_rate": 3.683996750609261e-06, + "loss": 2.146, + "step": 938 + }, + { + "epoch": 1.029041095890411, + "grad_norm": 1.8168137073516846, + "learning_rate": 3.681965881397238e-06, + "loss": 2.1594, + "step": 939 + }, + { + "epoch": 1.0301369863013699, + "grad_norm": 1.6803343296051025, + "learning_rate": 3.6799350121852157e-06, + "loss": 2.1168, + "step": 940 + }, + { + "epoch": 1.0301369863013699, + "eval_loss": 2.2941982746124268, + "eval_runtime": 37.7112, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 940 + }, + { + "epoch": 1.0312328767123287, + "grad_norm": 2.0829122066497803, + "learning_rate": 3.6779041429731925e-06, + "loss": 2.1399, + "step": 941 + }, + { + "epoch": 1.0323287671232877, + "grad_norm": 1.6813585758209229, + "learning_rate": 3.6758732737611702e-06, + "loss": 2.1312, + "step": 942 + }, + { + "epoch": 1.0334246575342465, + "grad_norm": 1.7511613368988037, + "learning_rate": 3.673842404549147e-06, + "loss": 2.1273, + "step": 943 + }, + { + "epoch": 1.0345205479452055, + "grad_norm": 1.8264555931091309, + "learning_rate": 3.671811535337125e-06, + "loss": 2.1343, + "step": 944 + }, + { + "epoch": 1.0356164383561643, + "grad_norm": 1.7675082683563232, + "learning_rate": 3.6697806661251016e-06, + "loss": 2.2153, + "step": 945 + }, + { + "epoch": 1.0367123287671234, + "grad_norm": 1.6589823961257935, + "learning_rate": 3.6677497969130793e-06, + "loss": 2.1575, + "step": 946 + }, + { + "epoch": 1.0378082191780822, + "grad_norm": 1.6730010509490967, + "learning_rate": 3.665718927701056e-06, + "loss": 2.1773, + "step": 947 + }, + { + "epoch": 1.038904109589041, + "grad_norm": 1.627983570098877, + "learning_rate": 3.663688058489034e-06, + "loss": 2.1394, + "step": 948 + }, + { + "epoch": 1.04, + "grad_norm": 1.6363394260406494, + "learning_rate": 3.6616571892770107e-06, + "loss": 2.0828, + "step": 949 + }, + { + "epoch": 1.0410958904109588, + "grad_norm": 1.6270133256912231, + "learning_rate": 3.659626320064988e-06, + "loss": 2.1308, + "step": 950 + }, + { + "epoch": 1.0410958904109588, + "eval_loss": 2.293911933898926, + "eval_runtime": 37.7095, + "eval_samples_per_second": 26.519, + "eval_steps_per_second": 1.114, + "step": 950 + }, + { + "epoch": 1.0421917808219179, + "grad_norm": 1.784549593925476, + "learning_rate": 3.6575954508529653e-06, + "loss": 2.1239, + "step": 951 + }, + { + "epoch": 1.0432876712328767, + "grad_norm": 1.7265311479568481, + "learning_rate": 3.6555645816409426e-06, + "loss": 2.1573, + "step": 952 + }, + { + "epoch": 1.0443835616438357, + "grad_norm": 1.8133327960968018, + "learning_rate": 3.65353371242892e-06, + "loss": 2.046, + "step": 953 + }, + { + "epoch": 1.0454794520547945, + "grad_norm": 1.7085450887680054, + "learning_rate": 3.651502843216897e-06, + "loss": 2.1471, + "step": 954 + }, + { + "epoch": 1.0465753424657533, + "grad_norm": 2.0099172592163086, + "learning_rate": 3.6494719740048744e-06, + "loss": 2.1193, + "step": 955 + }, + { + "epoch": 1.0476712328767124, + "grad_norm": 1.582085132598877, + "learning_rate": 3.6474411047928517e-06, + "loss": 2.1522, + "step": 956 + }, + { + "epoch": 1.0487671232876712, + "grad_norm": 1.7923028469085693, + "learning_rate": 3.645410235580829e-06, + "loss": 2.1358, + "step": 957 + }, + { + "epoch": 1.0498630136986302, + "grad_norm": 1.6636080741882324, + "learning_rate": 3.6433793663688062e-06, + "loss": 2.1263, + "step": 958 + }, + { + "epoch": 1.050958904109589, + "grad_norm": 1.681675672531128, + "learning_rate": 3.6413484971567835e-06, + "loss": 2.1369, + "step": 959 + }, + { + "epoch": 1.0520547945205478, + "grad_norm": 1.7007646560668945, + "learning_rate": 3.6393176279447608e-06, + "loss": 2.1606, + "step": 960 + }, + { + "epoch": 1.0520547945205478, + "eval_loss": 2.291552782058716, + "eval_runtime": 37.704, + "eval_samples_per_second": 26.522, + "eval_steps_per_second": 1.114, + "step": 960 + }, + { + "epoch": 1.0531506849315069, + "grad_norm": 1.632997989654541, + "learning_rate": 3.637286758732738e-06, + "loss": 2.1742, + "step": 961 + }, + { + "epoch": 1.0542465753424657, + "grad_norm": 1.6881494522094727, + "learning_rate": 3.6352558895207153e-06, + "loss": 2.1555, + "step": 962 + }, + { + "epoch": 1.0553424657534247, + "grad_norm": 1.6624116897583008, + "learning_rate": 3.633225020308692e-06, + "loss": 2.1278, + "step": 963 + }, + { + "epoch": 1.0564383561643835, + "grad_norm": 1.7851325273513794, + "learning_rate": 3.63119415109667e-06, + "loss": 2.1599, + "step": 964 + }, + { + "epoch": 1.0575342465753426, + "grad_norm": 1.7592188119888306, + "learning_rate": 3.6291632818846467e-06, + "loss": 2.0898, + "step": 965 + }, + { + "epoch": 1.0586301369863014, + "grad_norm": 1.9573521614074707, + "learning_rate": 3.6271324126726244e-06, + "loss": 2.0876, + "step": 966 + }, + { + "epoch": 1.0597260273972602, + "grad_norm": 1.6684988737106323, + "learning_rate": 3.6251015434606013e-06, + "loss": 2.1288, + "step": 967 + }, + { + "epoch": 1.0608219178082192, + "grad_norm": 2.0011937618255615, + "learning_rate": 3.623070674248579e-06, + "loss": 2.1166, + "step": 968 + }, + { + "epoch": 1.061917808219178, + "grad_norm": 1.7490673065185547, + "learning_rate": 3.621039805036556e-06, + "loss": 2.1532, + "step": 969 + }, + { + "epoch": 1.063013698630137, + "grad_norm": 1.756422758102417, + "learning_rate": 3.6190089358245335e-06, + "loss": 2.1486, + "step": 970 + }, + { + "epoch": 1.063013698630137, + "eval_loss": 2.29402494430542, + "eval_runtime": 37.7078, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 970 + }, + { + "epoch": 1.0641095890410959, + "grad_norm": 2.014153480529785, + "learning_rate": 3.6169780666125104e-06, + "loss": 2.1512, + "step": 971 + }, + { + "epoch": 1.0652054794520547, + "grad_norm": 1.6026027202606201, + "learning_rate": 3.614947197400488e-06, + "loss": 2.0787, + "step": 972 + }, + { + "epoch": 1.0663013698630137, + "grad_norm": 1.8149386644363403, + "learning_rate": 3.612916328188465e-06, + "loss": 2.1675, + "step": 973 + }, + { + "epoch": 1.0673972602739725, + "grad_norm": 1.697019338607788, + "learning_rate": 3.610885458976442e-06, + "loss": 2.1584, + "step": 974 + }, + { + "epoch": 1.0684931506849316, + "grad_norm": 1.7330286502838135, + "learning_rate": 3.6088545897644195e-06, + "loss": 2.1048, + "step": 975 + }, + { + "epoch": 1.0695890410958904, + "grad_norm": 1.650336742401123, + "learning_rate": 3.6068237205523963e-06, + "loss": 2.1078, + "step": 976 + }, + { + "epoch": 1.0706849315068494, + "grad_norm": 1.7817063331604004, + "learning_rate": 3.604792851340374e-06, + "loss": 2.1656, + "step": 977 + }, + { + "epoch": 1.0717808219178082, + "grad_norm": 1.818617820739746, + "learning_rate": 3.602761982128351e-06, + "loss": 2.1388, + "step": 978 + }, + { + "epoch": 1.072876712328767, + "grad_norm": 1.7448735237121582, + "learning_rate": 3.6007311129163286e-06, + "loss": 2.1022, + "step": 979 + }, + { + "epoch": 1.073972602739726, + "grad_norm": 1.8401628732681274, + "learning_rate": 3.5987002437043054e-06, + "loss": 2.1181, + "step": 980 + }, + { + "epoch": 1.073972602739726, + "eval_loss": 2.2948012351989746, + "eval_runtime": 37.6902, + "eval_samples_per_second": 26.532, + "eval_steps_per_second": 1.114, + "step": 980 + }, + { + "epoch": 1.0750684931506849, + "grad_norm": 1.979794979095459, + "learning_rate": 3.596669374492283e-06, + "loss": 2.1766, + "step": 981 + }, + { + "epoch": 1.076164383561644, + "grad_norm": 1.7657036781311035, + "learning_rate": 3.59463850528026e-06, + "loss": 2.0889, + "step": 982 + }, + { + "epoch": 1.0772602739726027, + "grad_norm": 1.8293733596801758, + "learning_rate": 3.5926076360682377e-06, + "loss": 2.159, + "step": 983 + }, + { + "epoch": 1.0783561643835617, + "grad_norm": 1.7979159355163574, + "learning_rate": 3.5905767668562145e-06, + "loss": 2.1391, + "step": 984 + }, + { + "epoch": 1.0794520547945206, + "grad_norm": 1.847598910331726, + "learning_rate": 3.5885458976441922e-06, + "loss": 2.1695, + "step": 985 + }, + { + "epoch": 1.0805479452054794, + "grad_norm": 1.608178734779358, + "learning_rate": 3.586515028432169e-06, + "loss": 2.1745, + "step": 986 + }, + { + "epoch": 1.0816438356164384, + "grad_norm": 1.8166873455047607, + "learning_rate": 3.5844841592201464e-06, + "loss": 2.0882, + "step": 987 + }, + { + "epoch": 1.0827397260273972, + "grad_norm": 1.847463607788086, + "learning_rate": 3.5824532900081237e-06, + "loss": 2.1244, + "step": 988 + }, + { + "epoch": 1.0838356164383562, + "grad_norm": 1.6194696426391602, + "learning_rate": 3.580422420796101e-06, + "loss": 2.1398, + "step": 989 + }, + { + "epoch": 1.084931506849315, + "grad_norm": 1.9055372476577759, + "learning_rate": 3.578391551584078e-06, + "loss": 2.0774, + "step": 990 + }, + { + "epoch": 1.084931506849315, + "eval_loss": 2.2899444103240967, + "eval_runtime": 37.7237, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 990 + }, + { + "epoch": 1.086027397260274, + "grad_norm": 1.6451077461242676, + "learning_rate": 3.5763606823720555e-06, + "loss": 2.1351, + "step": 991 + }, + { + "epoch": 1.087123287671233, + "grad_norm": 1.8144199848175049, + "learning_rate": 3.5743298131600328e-06, + "loss": 2.1536, + "step": 992 + }, + { + "epoch": 1.0882191780821917, + "grad_norm": 1.760130763053894, + "learning_rate": 3.57229894394801e-06, + "loss": 2.1485, + "step": 993 + }, + { + "epoch": 1.0893150684931507, + "grad_norm": 1.7544556856155396, + "learning_rate": 3.5702680747359873e-06, + "loss": 2.1136, + "step": 994 + }, + { + "epoch": 1.0904109589041096, + "grad_norm": 1.9884154796600342, + "learning_rate": 3.5682372055239646e-06, + "loss": 2.1366, + "step": 995 + }, + { + "epoch": 1.0915068493150686, + "grad_norm": 1.72066330909729, + "learning_rate": 3.566206336311942e-06, + "loss": 2.1103, + "step": 996 + }, + { + "epoch": 1.0926027397260274, + "grad_norm": 1.807808756828308, + "learning_rate": 3.564175467099919e-06, + "loss": 2.0905, + "step": 997 + }, + { + "epoch": 1.0936986301369862, + "grad_norm": 2.1075172424316406, + "learning_rate": 3.5621445978878964e-06, + "loss": 2.1427, + "step": 998 + }, + { + "epoch": 1.0947945205479452, + "grad_norm": 1.7037009000778198, + "learning_rate": 3.5601137286758737e-06, + "loss": 2.1124, + "step": 999 + }, + { + "epoch": 1.095890410958904, + "grad_norm": 1.842721700668335, + "learning_rate": 3.5580828594638505e-06, + "loss": 2.1379, + "step": 1000 + }, + { + "epoch": 1.095890410958904, + "eval_loss": 2.291067361831665, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 1000 + }, + { + "epoch": 1.096986301369863, + "grad_norm": 1.802979588508606, + "learning_rate": 3.5560519902518282e-06, + "loss": 2.156, + "step": 1001 + }, + { + "epoch": 1.098082191780822, + "grad_norm": 1.699763298034668, + "learning_rate": 3.554021121039805e-06, + "loss": 2.1353, + "step": 1002 + }, + { + "epoch": 1.099178082191781, + "grad_norm": 1.695544958114624, + "learning_rate": 3.551990251827783e-06, + "loss": 2.1039, + "step": 1003 + }, + { + "epoch": 1.1002739726027397, + "grad_norm": 1.9509766101837158, + "learning_rate": 3.5499593826157596e-06, + "loss": 2.1201, + "step": 1004 + }, + { + "epoch": 1.1013698630136985, + "grad_norm": 1.7657231092453003, + "learning_rate": 3.5479285134037373e-06, + "loss": 2.1149, + "step": 1005 + }, + { + "epoch": 1.1024657534246576, + "grad_norm": 1.7416653633117676, + "learning_rate": 3.545897644191714e-06, + "loss": 2.1116, + "step": 1006 + }, + { + "epoch": 1.1035616438356164, + "grad_norm": 1.804185390472412, + "learning_rate": 3.543866774979692e-06, + "loss": 2.1426, + "step": 1007 + }, + { + "epoch": 1.1046575342465754, + "grad_norm": 1.8011984825134277, + "learning_rate": 3.5418359057676687e-06, + "loss": 2.1132, + "step": 1008 + }, + { + "epoch": 1.1057534246575342, + "grad_norm": 1.9039791822433472, + "learning_rate": 3.5398050365556464e-06, + "loss": 2.1479, + "step": 1009 + }, + { + "epoch": 1.106849315068493, + "grad_norm": 1.9513524770736694, + "learning_rate": 3.5377741673436233e-06, + "loss": 2.1093, + "step": 1010 + }, + { + "epoch": 1.106849315068493, + "eval_loss": 2.290522575378418, + "eval_runtime": 37.4617, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 1010 + }, + { + "epoch": 1.107945205479452, + "grad_norm": 1.7262980937957764, + "learning_rate": 3.535743298131601e-06, + "loss": 2.138, + "step": 1011 + }, + { + "epoch": 1.1090410958904109, + "grad_norm": 1.9514356851577759, + "learning_rate": 3.533712428919578e-06, + "loss": 2.1669, + "step": 1012 + }, + { + "epoch": 1.11013698630137, + "grad_norm": 1.7947099208831787, + "learning_rate": 3.5316815597075547e-06, + "loss": 2.105, + "step": 1013 + }, + { + "epoch": 1.1112328767123287, + "grad_norm": 1.8680298328399658, + "learning_rate": 3.5296506904955324e-06, + "loss": 2.1306, + "step": 1014 + }, + { + "epoch": 1.1123287671232878, + "grad_norm": 2.121732711791992, + "learning_rate": 3.5276198212835093e-06, + "loss": 2.1681, + "step": 1015 + }, + { + "epoch": 1.1134246575342466, + "grad_norm": 1.662564754486084, + "learning_rate": 3.525588952071487e-06, + "loss": 2.1681, + "step": 1016 + }, + { + "epoch": 1.1145205479452054, + "grad_norm": 1.8358488082885742, + "learning_rate": 3.523558082859464e-06, + "loss": 2.1363, + "step": 1017 + }, + { + "epoch": 1.1156164383561644, + "grad_norm": 1.8742910623550415, + "learning_rate": 3.5215272136474415e-06, + "loss": 2.1179, + "step": 1018 + }, + { + "epoch": 1.1167123287671232, + "grad_norm": 1.8780633211135864, + "learning_rate": 3.5194963444354184e-06, + "loss": 2.103, + "step": 1019 + }, + { + "epoch": 1.1178082191780823, + "grad_norm": 1.6979540586471558, + "learning_rate": 3.517465475223396e-06, + "loss": 2.0885, + "step": 1020 + }, + { + "epoch": 1.1178082191780823, + "eval_loss": 2.2939751148223877, + "eval_runtime": 37.66, + "eval_samples_per_second": 26.553, + "eval_steps_per_second": 1.115, + "step": 1020 + }, + { + "epoch": 1.118904109589041, + "grad_norm": 2.494971752166748, + "learning_rate": 3.515434606011373e-06, + "loss": 2.1582, + "step": 1021 + }, + { + "epoch": 1.12, + "grad_norm": 1.9183858633041382, + "learning_rate": 3.5134037367993506e-06, + "loss": 2.189, + "step": 1022 + }, + { + "epoch": 1.121095890410959, + "grad_norm": 1.9055982828140259, + "learning_rate": 3.5113728675873275e-06, + "loss": 2.1503, + "step": 1023 + }, + { + "epoch": 1.1221917808219177, + "grad_norm": 2.118046522140503, + "learning_rate": 3.5093419983753047e-06, + "loss": 2.1414, + "step": 1024 + }, + { + "epoch": 1.1232876712328768, + "grad_norm": 1.8413280248641968, + "learning_rate": 3.507311129163282e-06, + "loss": 2.09, + "step": 1025 + }, + { + "epoch": 1.1243835616438356, + "grad_norm": 2.0261356830596924, + "learning_rate": 3.5052802599512593e-06, + "loss": 2.132, + "step": 1026 + }, + { + "epoch": 1.1254794520547946, + "grad_norm": 1.8393104076385498, + "learning_rate": 3.5032493907392366e-06, + "loss": 2.1366, + "step": 1027 + }, + { + "epoch": 1.1265753424657534, + "grad_norm": 2.0282111167907715, + "learning_rate": 3.501218521527214e-06, + "loss": 2.0758, + "step": 1028 + }, + { + "epoch": 1.1276712328767124, + "grad_norm": 1.8438650369644165, + "learning_rate": 3.499187652315191e-06, + "loss": 2.1294, + "step": 1029 + }, + { + "epoch": 1.1287671232876713, + "grad_norm": 1.890739917755127, + "learning_rate": 3.4971567831031684e-06, + "loss": 2.1591, + "step": 1030 + }, + { + "epoch": 1.1287671232876713, + "eval_loss": 2.2894623279571533, + "eval_runtime": 37.6941, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1030 + }, + { + "epoch": 1.12986301369863, + "grad_norm": 1.877964973449707, + "learning_rate": 3.4951259138911457e-06, + "loss": 2.0823, + "step": 1031 + }, + { + "epoch": 1.130958904109589, + "grad_norm": 1.7460334300994873, + "learning_rate": 3.493095044679123e-06, + "loss": 2.1758, + "step": 1032 + }, + { + "epoch": 1.132054794520548, + "grad_norm": 1.668459415435791, + "learning_rate": 3.4910641754671002e-06, + "loss": 2.0461, + "step": 1033 + }, + { + "epoch": 1.133150684931507, + "grad_norm": 1.9560495615005493, + "learning_rate": 3.4890333062550775e-06, + "loss": 2.1394, + "step": 1034 + }, + { + "epoch": 1.1342465753424658, + "grad_norm": 1.820432424545288, + "learning_rate": 3.4870024370430548e-06, + "loss": 2.0735, + "step": 1035 + }, + { + "epoch": 1.1353424657534246, + "grad_norm": 1.7064971923828125, + "learning_rate": 3.484971567831032e-06, + "loss": 2.1196, + "step": 1036 + }, + { + "epoch": 1.1364383561643836, + "grad_norm": 1.837594747543335, + "learning_rate": 3.482940698619009e-06, + "loss": 2.1562, + "step": 1037 + }, + { + "epoch": 1.1375342465753424, + "grad_norm": 1.7570106983184814, + "learning_rate": 3.4809098294069866e-06, + "loss": 2.1253, + "step": 1038 + }, + { + "epoch": 1.1386301369863014, + "grad_norm": 1.7931634187698364, + "learning_rate": 3.4788789601949634e-06, + "loss": 2.1369, + "step": 1039 + }, + { + "epoch": 1.1397260273972603, + "grad_norm": 1.957241177558899, + "learning_rate": 3.476848090982941e-06, + "loss": 2.1204, + "step": 1040 + }, + { + "epoch": 1.1397260273972603, + "eval_loss": 2.2888705730438232, + "eval_runtime": 37.7002, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 1040 + }, + { + "epoch": 1.140821917808219, + "grad_norm": 1.7801663875579834, + "learning_rate": 3.474817221770918e-06, + "loss": 2.2008, + "step": 1041 + }, + { + "epoch": 1.141917808219178, + "grad_norm": 1.654398798942566, + "learning_rate": 3.4727863525588957e-06, + "loss": 2.1451, + "step": 1042 + }, + { + "epoch": 1.143013698630137, + "grad_norm": 1.8888044357299805, + "learning_rate": 3.4707554833468725e-06, + "loss": 2.1318, + "step": 1043 + }, + { + "epoch": 1.144109589041096, + "grad_norm": 1.8398077487945557, + "learning_rate": 3.4687246141348502e-06, + "loss": 2.1051, + "step": 1044 + }, + { + "epoch": 1.1452054794520548, + "grad_norm": 1.6971104145050049, + "learning_rate": 3.466693744922827e-06, + "loss": 2.119, + "step": 1045 + }, + { + "epoch": 1.1463013698630138, + "grad_norm": 1.605700969696045, + "learning_rate": 3.464662875710805e-06, + "loss": 2.1343, + "step": 1046 + }, + { + "epoch": 1.1473972602739726, + "grad_norm": 1.8513860702514648, + "learning_rate": 3.4626320064987817e-06, + "loss": 2.1331, + "step": 1047 + }, + { + "epoch": 1.1484931506849314, + "grad_norm": 1.7160199880599976, + "learning_rate": 3.4606011372867594e-06, + "loss": 2.1367, + "step": 1048 + }, + { + "epoch": 1.1495890410958904, + "grad_norm": 1.5926575660705566, + "learning_rate": 3.458570268074736e-06, + "loss": 2.1183, + "step": 1049 + }, + { + "epoch": 1.1506849315068493, + "grad_norm": 1.663843035697937, + "learning_rate": 3.4565393988627135e-06, + "loss": 2.1107, + "step": 1050 + }, + { + "epoch": 1.1506849315068493, + "eval_loss": 2.286076545715332, + "eval_runtime": 37.6944, + "eval_samples_per_second": 26.529, + "eval_steps_per_second": 1.114, + "step": 1050 + }, + { + "epoch": 1.1517808219178083, + "grad_norm": 1.7697075605392456, + "learning_rate": 3.4545085296506908e-06, + "loss": 2.0912, + "step": 1051 + }, + { + "epoch": 1.152876712328767, + "grad_norm": 1.626094937324524, + "learning_rate": 3.452477660438668e-06, + "loss": 2.1508, + "step": 1052 + }, + { + "epoch": 1.1539726027397261, + "grad_norm": 1.9503792524337769, + "learning_rate": 3.4504467912266453e-06, + "loss": 2.0819, + "step": 1053 + }, + { + "epoch": 1.155068493150685, + "grad_norm": 1.7748280763626099, + "learning_rate": 3.4484159220146226e-06, + "loss": 2.1233, + "step": 1054 + }, + { + "epoch": 1.1561643835616437, + "grad_norm": 2.099339008331299, + "learning_rate": 3.4463850528026e-06, + "loss": 2.1245, + "step": 1055 + }, + { + "epoch": 1.1572602739726028, + "grad_norm": 1.6200031042099, + "learning_rate": 3.444354183590577e-06, + "loss": 2.1437, + "step": 1056 + }, + { + "epoch": 1.1583561643835616, + "grad_norm": 1.7703361511230469, + "learning_rate": 3.4423233143785544e-06, + "loss": 2.0874, + "step": 1057 + }, + { + "epoch": 1.1594520547945206, + "grad_norm": 1.6505805253982544, + "learning_rate": 3.4402924451665317e-06, + "loss": 2.0655, + "step": 1058 + }, + { + "epoch": 1.1605479452054794, + "grad_norm": 1.7199327945709229, + "learning_rate": 3.438261575954509e-06, + "loss": 2.1383, + "step": 1059 + }, + { + "epoch": 1.1616438356164385, + "grad_norm": 1.772292137145996, + "learning_rate": 3.4362307067424862e-06, + "loss": 2.1673, + "step": 1060 + }, + { + "epoch": 1.1616438356164385, + "eval_loss": 2.2869999408721924, + "eval_runtime": 37.7398, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1060 + }, + { + "epoch": 1.1627397260273973, + "grad_norm": 1.5849649906158447, + "learning_rate": 3.434199837530463e-06, + "loss": 2.1331, + "step": 1061 + }, + { + "epoch": 1.163835616438356, + "grad_norm": 1.8389405012130737, + "learning_rate": 3.432168968318441e-06, + "loss": 2.0832, + "step": 1062 + }, + { + "epoch": 1.1649315068493151, + "grad_norm": 1.7142471075057983, + "learning_rate": 3.4301380991064176e-06, + "loss": 2.1447, + "step": 1063 + }, + { + "epoch": 1.166027397260274, + "grad_norm": 1.7853225469589233, + "learning_rate": 3.4281072298943953e-06, + "loss": 2.1591, + "step": 1064 + }, + { + "epoch": 1.167123287671233, + "grad_norm": 1.8228695392608643, + "learning_rate": 3.426076360682372e-06, + "loss": 2.1684, + "step": 1065 + }, + { + "epoch": 1.1682191780821918, + "grad_norm": 1.7090480327606201, + "learning_rate": 3.42404549147035e-06, + "loss": 2.1339, + "step": 1066 + }, + { + "epoch": 1.1693150684931506, + "grad_norm": 1.642682433128357, + "learning_rate": 3.4220146222583267e-06, + "loss": 2.075, + "step": 1067 + }, + { + "epoch": 1.1704109589041096, + "grad_norm": 1.8230397701263428, + "learning_rate": 3.4199837530463044e-06, + "loss": 2.1128, + "step": 1068 + }, + { + "epoch": 1.1715068493150684, + "grad_norm": 1.761135458946228, + "learning_rate": 3.4179528838342813e-06, + "loss": 2.1885, + "step": 1069 + }, + { + "epoch": 1.1726027397260275, + "grad_norm": 1.7829216718673706, + "learning_rate": 3.415922014622259e-06, + "loss": 2.1584, + "step": 1070 + }, + { + "epoch": 1.1726027397260275, + "eval_loss": 2.2878382205963135, + "eval_runtime": 37.7433, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1070 + }, + { + "epoch": 1.1736986301369863, + "grad_norm": 1.5480386018753052, + "learning_rate": 3.413891145410236e-06, + "loss": 2.1602, + "step": 1071 + }, + { + "epoch": 1.174794520547945, + "grad_norm": 1.8957836627960205, + "learning_rate": 3.4118602761982135e-06, + "loss": 2.1706, + "step": 1072 + }, + { + "epoch": 1.1758904109589041, + "grad_norm": 1.924424648284912, + "learning_rate": 3.4098294069861904e-06, + "loss": 2.1339, + "step": 1073 + }, + { + "epoch": 1.176986301369863, + "grad_norm": 1.660370111465454, + "learning_rate": 3.4077985377741673e-06, + "loss": 2.1175, + "step": 1074 + }, + { + "epoch": 1.178082191780822, + "grad_norm": 1.7255185842514038, + "learning_rate": 3.405767668562145e-06, + "loss": 2.1672, + "step": 1075 + }, + { + "epoch": 1.1791780821917808, + "grad_norm": 1.8028507232666016, + "learning_rate": 3.403736799350122e-06, + "loss": 2.1527, + "step": 1076 + }, + { + "epoch": 1.1802739726027398, + "grad_norm": 1.7100828886032104, + "learning_rate": 3.4017059301380995e-06, + "loss": 2.1031, + "step": 1077 + }, + { + "epoch": 1.1813698630136986, + "grad_norm": 1.785955548286438, + "learning_rate": 3.3996750609260764e-06, + "loss": 2.1829, + "step": 1078 + }, + { + "epoch": 1.1824657534246574, + "grad_norm": 1.8583980798721313, + "learning_rate": 3.397644191714054e-06, + "loss": 2.1198, + "step": 1079 + }, + { + "epoch": 1.1835616438356165, + "grad_norm": 1.8848942518234253, + "learning_rate": 3.395613322502031e-06, + "loss": 2.123, + "step": 1080 + }, + { + "epoch": 1.1835616438356165, + "eval_loss": 2.288412094116211, + "eval_runtime": 37.7637, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 1080 + }, + { + "epoch": 1.1846575342465753, + "grad_norm": 1.6747279167175293, + "learning_rate": 3.3935824532900086e-06, + "loss": 2.1087, + "step": 1081 + }, + { + "epoch": 1.1857534246575343, + "grad_norm": 1.8993490934371948, + "learning_rate": 3.3915515840779855e-06, + "loss": 2.1316, + "step": 1082 + }, + { + "epoch": 1.1868493150684931, + "grad_norm": 1.7047377824783325, + "learning_rate": 3.389520714865963e-06, + "loss": 2.1242, + "step": 1083 + }, + { + "epoch": 1.1879452054794521, + "grad_norm": 1.767811894416809, + "learning_rate": 3.38748984565394e-06, + "loss": 2.1556, + "step": 1084 + }, + { + "epoch": 1.189041095890411, + "grad_norm": 1.665313720703125, + "learning_rate": 3.3854589764419177e-06, + "loss": 2.101, + "step": 1085 + }, + { + "epoch": 1.1901369863013698, + "grad_norm": 1.7752090692520142, + "learning_rate": 3.3834281072298946e-06, + "loss": 2.1307, + "step": 1086 + }, + { + "epoch": 1.1912328767123288, + "grad_norm": 1.9759247303009033, + "learning_rate": 3.381397238017872e-06, + "loss": 2.1093, + "step": 1087 + }, + { + "epoch": 1.1923287671232876, + "grad_norm": 1.8772896528244019, + "learning_rate": 3.379366368805849e-06, + "loss": 2.1851, + "step": 1088 + }, + { + "epoch": 1.1934246575342466, + "grad_norm": 1.7442713975906372, + "learning_rate": 3.3773354995938264e-06, + "loss": 2.1433, + "step": 1089 + }, + { + "epoch": 1.1945205479452055, + "grad_norm": 1.8106931447982788, + "learning_rate": 3.3753046303818037e-06, + "loss": 2.1246, + "step": 1090 + }, + { + "epoch": 1.1945205479452055, + "eval_loss": 2.2858333587646484, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1090 + }, + { + "epoch": 1.1956164383561645, + "grad_norm": 1.7938610315322876, + "learning_rate": 3.373273761169781e-06, + "loss": 2.1325, + "step": 1091 + }, + { + "epoch": 1.1967123287671233, + "grad_norm": 1.5760858058929443, + "learning_rate": 3.3712428919577582e-06, + "loss": 2.1036, + "step": 1092 + }, + { + "epoch": 1.1978082191780821, + "grad_norm": 1.8173094987869263, + "learning_rate": 3.3692120227457355e-06, + "loss": 2.1509, + "step": 1093 + }, + { + "epoch": 1.1989041095890411, + "grad_norm": 2.0316812992095947, + "learning_rate": 3.3671811535337128e-06, + "loss": 2.0883, + "step": 1094 + }, + { + "epoch": 1.2, + "grad_norm": 1.7735658884048462, + "learning_rate": 3.36515028432169e-06, + "loss": 2.1038, + "step": 1095 + }, + { + "epoch": 1.201095890410959, + "grad_norm": 1.752718210220337, + "learning_rate": 3.3631194151096673e-06, + "loss": 2.1393, + "step": 1096 + }, + { + "epoch": 1.2021917808219178, + "grad_norm": 1.778838872909546, + "learning_rate": 3.3610885458976446e-06, + "loss": 2.1641, + "step": 1097 + }, + { + "epoch": 1.2032876712328768, + "grad_norm": 1.7422083616256714, + "learning_rate": 3.359057676685622e-06, + "loss": 2.1624, + "step": 1098 + }, + { + "epoch": 1.2043835616438356, + "grad_norm": 1.7196860313415527, + "learning_rate": 3.357026807473599e-06, + "loss": 2.0845, + "step": 1099 + }, + { + "epoch": 1.2054794520547945, + "grad_norm": 1.8373215198516846, + "learning_rate": 3.354995938261576e-06, + "loss": 2.1092, + "step": 1100 + }, + { + "epoch": 1.2054794520547945, + "eval_loss": 2.285470485687256, + "eval_runtime": 37.728, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 1100 + }, + { + "epoch": 1.2065753424657535, + "grad_norm": 1.626450538635254, + "learning_rate": 3.3529650690495537e-06, + "loss": 2.0789, + "step": 1101 + }, + { + "epoch": 1.2076712328767123, + "grad_norm": 1.736264705657959, + "learning_rate": 3.3509341998375305e-06, + "loss": 2.1227, + "step": 1102 + }, + { + "epoch": 1.208767123287671, + "grad_norm": 1.6195942163467407, + "learning_rate": 3.3489033306255082e-06, + "loss": 2.1301, + "step": 1103 + }, + { + "epoch": 1.2098630136986301, + "grad_norm": 1.6958918571472168, + "learning_rate": 3.346872461413485e-06, + "loss": 2.1059, + "step": 1104 + }, + { + "epoch": 1.210958904109589, + "grad_norm": 1.6385109424591064, + "learning_rate": 3.344841592201463e-06, + "loss": 2.0917, + "step": 1105 + }, + { + "epoch": 1.212054794520548, + "grad_norm": 1.6833516359329224, + "learning_rate": 3.3428107229894397e-06, + "loss": 2.1204, + "step": 1106 + }, + { + "epoch": 1.2131506849315068, + "grad_norm": 1.724780559539795, + "learning_rate": 3.3407798537774174e-06, + "loss": 2.138, + "step": 1107 + }, + { + "epoch": 1.2142465753424658, + "grad_norm": 1.7483834028244019, + "learning_rate": 3.338748984565394e-06, + "loss": 2.1628, + "step": 1108 + }, + { + "epoch": 1.2153424657534246, + "grad_norm": 1.6398268938064575, + "learning_rate": 3.336718115353372e-06, + "loss": 2.1738, + "step": 1109 + }, + { + "epoch": 1.2164383561643834, + "grad_norm": 1.6916258335113525, + "learning_rate": 3.3346872461413488e-06, + "loss": 2.1182, + "step": 1110 + }, + { + "epoch": 1.2164383561643834, + "eval_loss": 2.285295009613037, + "eval_runtime": 37.508, + "eval_samples_per_second": 26.661, + "eval_steps_per_second": 1.12, + "step": 1110 + }, + { + "epoch": 1.2175342465753425, + "grad_norm": 1.764147400856018, + "learning_rate": 3.3326563769293256e-06, + "loss": 2.1085, + "step": 1111 + }, + { + "epoch": 1.2186301369863013, + "grad_norm": 1.5984162092208862, + "learning_rate": 3.3306255077173033e-06, + "loss": 2.1208, + "step": 1112 + }, + { + "epoch": 1.2197260273972603, + "grad_norm": 1.6280403137207031, + "learning_rate": 3.32859463850528e-06, + "loss": 2.1226, + "step": 1113 + }, + { + "epoch": 1.2208219178082191, + "grad_norm": 1.5922425985336304, + "learning_rate": 3.326563769293258e-06, + "loss": 2.071, + "step": 1114 + }, + { + "epoch": 1.2219178082191782, + "grad_norm": 1.5941506624221802, + "learning_rate": 3.3245329000812347e-06, + "loss": 2.1059, + "step": 1115 + }, + { + "epoch": 1.223013698630137, + "grad_norm": 1.7812453508377075, + "learning_rate": 3.3225020308692124e-06, + "loss": 2.1192, + "step": 1116 + }, + { + "epoch": 1.2241095890410958, + "grad_norm": 1.6033459901809692, + "learning_rate": 3.3204711616571893e-06, + "loss": 2.0898, + "step": 1117 + }, + { + "epoch": 1.2252054794520548, + "grad_norm": 1.7053911685943604, + "learning_rate": 3.318440292445167e-06, + "loss": 2.1237, + "step": 1118 + }, + { + "epoch": 1.2263013698630136, + "grad_norm": 1.699777364730835, + "learning_rate": 3.316409423233144e-06, + "loss": 2.1014, + "step": 1119 + }, + { + "epoch": 1.2273972602739727, + "grad_norm": 1.7440428733825684, + "learning_rate": 3.3143785540211215e-06, + "loss": 2.1777, + "step": 1120 + }, + { + "epoch": 1.2273972602739727, + "eval_loss": 2.2857329845428467, + "eval_runtime": 37.7298, + "eval_samples_per_second": 26.504, + "eval_steps_per_second": 1.113, + "step": 1120 + }, + { + "epoch": 1.2284931506849315, + "grad_norm": 1.6901912689208984, + "learning_rate": 3.3123476848090984e-06, + "loss": 2.1373, + "step": 1121 + }, + { + "epoch": 1.2295890410958905, + "grad_norm": 1.5694825649261475, + "learning_rate": 3.310316815597076e-06, + "loss": 2.1285, + "step": 1122 + }, + { + "epoch": 1.2306849315068493, + "grad_norm": 1.6858313083648682, + "learning_rate": 3.308285946385053e-06, + "loss": 2.1858, + "step": 1123 + }, + { + "epoch": 1.2317808219178081, + "grad_norm": 1.6308691501617432, + "learning_rate": 3.30625507717303e-06, + "loss": 2.0793, + "step": 1124 + }, + { + "epoch": 1.2328767123287672, + "grad_norm": 1.6708226203918457, + "learning_rate": 3.3042242079610075e-06, + "loss": 2.1445, + "step": 1125 + }, + { + "epoch": 1.233972602739726, + "grad_norm": 1.7215708494186401, + "learning_rate": 3.3021933387489847e-06, + "loss": 2.159, + "step": 1126 + }, + { + "epoch": 1.235068493150685, + "grad_norm": 1.5764784812927246, + "learning_rate": 3.300162469536962e-06, + "loss": 2.0993, + "step": 1127 + }, + { + "epoch": 1.2361643835616438, + "grad_norm": 1.6322427988052368, + "learning_rate": 3.2981316003249393e-06, + "loss": 2.1643, + "step": 1128 + }, + { + "epoch": 1.2372602739726029, + "grad_norm": 1.6528997421264648, + "learning_rate": 3.2961007311129166e-06, + "loss": 2.1332, + "step": 1129 + }, + { + "epoch": 1.2383561643835617, + "grad_norm": 1.6434602737426758, + "learning_rate": 3.294069861900894e-06, + "loss": 2.1875, + "step": 1130 + }, + { + "epoch": 1.2383561643835617, + "eval_loss": 2.2840054035186768, + "eval_runtime": 37.7227, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 1130 + }, + { + "epoch": 1.2394520547945205, + "grad_norm": 1.6974412202835083, + "learning_rate": 3.292038992688871e-06, + "loss": 2.1396, + "step": 1131 + }, + { + "epoch": 1.2405479452054795, + "grad_norm": 1.7754731178283691, + "learning_rate": 3.2900081234768484e-06, + "loss": 2.1448, + "step": 1132 + }, + { + "epoch": 1.2416438356164383, + "grad_norm": 1.5928994417190552, + "learning_rate": 3.2879772542648257e-06, + "loss": 2.1665, + "step": 1133 + }, + { + "epoch": 1.2427397260273974, + "grad_norm": 1.609265685081482, + "learning_rate": 3.285946385052803e-06, + "loss": 2.0787, + "step": 1134 + }, + { + "epoch": 1.2438356164383562, + "grad_norm": 1.7366653680801392, + "learning_rate": 3.2839155158407802e-06, + "loss": 2.1453, + "step": 1135 + }, + { + "epoch": 1.244931506849315, + "grad_norm": 1.5945746898651123, + "learning_rate": 3.2818846466287575e-06, + "loss": 2.0705, + "step": 1136 + }, + { + "epoch": 1.246027397260274, + "grad_norm": 1.55001962184906, + "learning_rate": 3.2798537774167344e-06, + "loss": 2.1802, + "step": 1137 + }, + { + "epoch": 1.2471232876712328, + "grad_norm": 1.6575016975402832, + "learning_rate": 3.277822908204712e-06, + "loss": 2.1205, + "step": 1138 + }, + { + "epoch": 1.2482191780821918, + "grad_norm": 1.7479991912841797, + "learning_rate": 3.275792038992689e-06, + "loss": 2.0747, + "step": 1139 + }, + { + "epoch": 1.2493150684931507, + "grad_norm": 1.5351295471191406, + "learning_rate": 3.2737611697806666e-06, + "loss": 2.1263, + "step": 1140 + }, + { + "epoch": 1.2493150684931507, + "eval_loss": 2.286621570587158, + "eval_runtime": 37.744, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1140 + }, + { + "epoch": 1.2504109589041095, + "grad_norm": 1.6440327167510986, + "learning_rate": 3.2717303005686435e-06, + "loss": 2.0967, + "step": 1141 + }, + { + "epoch": 1.2515068493150685, + "grad_norm": 1.6527445316314697, + "learning_rate": 3.269699431356621e-06, + "loss": 2.1375, + "step": 1142 + }, + { + "epoch": 1.2526027397260273, + "grad_norm": 1.6786249876022339, + "learning_rate": 3.267668562144598e-06, + "loss": 2.1292, + "step": 1143 + }, + { + "epoch": 1.2536986301369863, + "grad_norm": 1.693476676940918, + "learning_rate": 3.2656376929325757e-06, + "loss": 2.1008, + "step": 1144 + }, + { + "epoch": 1.2547945205479452, + "grad_norm": 1.5634204149246216, + "learning_rate": 3.2636068237205526e-06, + "loss": 2.1504, + "step": 1145 + }, + { + "epoch": 1.2558904109589042, + "grad_norm": 1.6024060249328613, + "learning_rate": 3.2615759545085303e-06, + "loss": 2.1347, + "step": 1146 + }, + { + "epoch": 1.256986301369863, + "grad_norm": 1.7107738256454468, + "learning_rate": 3.259545085296507e-06, + "loss": 2.1722, + "step": 1147 + }, + { + "epoch": 1.2580821917808218, + "grad_norm": 1.6810829639434814, + "learning_rate": 3.257514216084484e-06, + "loss": 2.0948, + "step": 1148 + }, + { + "epoch": 1.2591780821917808, + "grad_norm": 1.6437987089157104, + "learning_rate": 3.2554833468724617e-06, + "loss": 2.1677, + "step": 1149 + }, + { + "epoch": 1.2602739726027397, + "grad_norm": 1.6707998514175415, + "learning_rate": 3.2534524776604385e-06, + "loss": 2.136, + "step": 1150 + }, + { + "epoch": 1.2602739726027397, + "eval_loss": 2.2854597568511963, + "eval_runtime": 37.72, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 1150 + }, + { + "epoch": 1.2613698630136987, + "grad_norm": 1.653172492980957, + "learning_rate": 3.2514216084484162e-06, + "loss": 2.1278, + "step": 1151 + }, + { + "epoch": 1.2624657534246575, + "grad_norm": 1.6651742458343506, + "learning_rate": 3.249390739236393e-06, + "loss": 2.1401, + "step": 1152 + }, + { + "epoch": 1.2635616438356165, + "grad_norm": 1.6661403179168701, + "learning_rate": 3.2473598700243708e-06, + "loss": 2.1067, + "step": 1153 + }, + { + "epoch": 1.2646575342465753, + "grad_norm": 1.6470211744308472, + "learning_rate": 3.2453290008123476e-06, + "loss": 2.1309, + "step": 1154 + }, + { + "epoch": 1.2657534246575342, + "grad_norm": 1.8089711666107178, + "learning_rate": 3.2432981316003253e-06, + "loss": 2.139, + "step": 1155 + }, + { + "epoch": 1.2668493150684932, + "grad_norm": 1.7904900312423706, + "learning_rate": 3.241267262388302e-06, + "loss": 2.0906, + "step": 1156 + }, + { + "epoch": 1.267945205479452, + "grad_norm": 1.6502841711044312, + "learning_rate": 3.23923639317628e-06, + "loss": 2.1263, + "step": 1157 + }, + { + "epoch": 1.269041095890411, + "grad_norm": 1.5825791358947754, + "learning_rate": 3.2372055239642567e-06, + "loss": 2.098, + "step": 1158 + }, + { + "epoch": 1.2701369863013698, + "grad_norm": 1.7173693180084229, + "learning_rate": 3.2351746547522344e-06, + "loss": 2.1367, + "step": 1159 + }, + { + "epoch": 1.2712328767123289, + "grad_norm": 1.766738772392273, + "learning_rate": 3.2331437855402113e-06, + "loss": 2.1914, + "step": 1160 + }, + { + "epoch": 1.2712328767123289, + "eval_loss": 2.282374858856201, + "eval_runtime": 37.7395, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1160 + }, + { + "epoch": 1.2723287671232877, + "grad_norm": 1.6669079065322876, + "learning_rate": 3.2311129163281885e-06, + "loss": 2.0959, + "step": 1161 + }, + { + "epoch": 1.2734246575342465, + "grad_norm": 1.7150713205337524, + "learning_rate": 3.229082047116166e-06, + "loss": 2.1431, + "step": 1162 + }, + { + "epoch": 1.2745205479452055, + "grad_norm": 1.5927042961120605, + "learning_rate": 3.227051177904143e-06, + "loss": 2.1239, + "step": 1163 + }, + { + "epoch": 1.2756164383561643, + "grad_norm": 1.8086377382278442, + "learning_rate": 3.2250203086921204e-06, + "loss": 2.1307, + "step": 1164 + }, + { + "epoch": 1.2767123287671232, + "grad_norm": 1.8078030347824097, + "learning_rate": 3.2229894394800977e-06, + "loss": 2.1632, + "step": 1165 + }, + { + "epoch": 1.2778082191780822, + "grad_norm": 1.7087128162384033, + "learning_rate": 3.220958570268075e-06, + "loss": 2.0663, + "step": 1166 + }, + { + "epoch": 1.2789041095890412, + "grad_norm": 1.8242524862289429, + "learning_rate": 3.218927701056052e-06, + "loss": 2.1225, + "step": 1167 + }, + { + "epoch": 1.28, + "grad_norm": 1.6094539165496826, + "learning_rate": 3.2168968318440295e-06, + "loss": 2.0586, + "step": 1168 + }, + { + "epoch": 1.2810958904109588, + "grad_norm": 1.8592051267623901, + "learning_rate": 3.2148659626320068e-06, + "loss": 2.1632, + "step": 1169 + }, + { + "epoch": 1.2821917808219179, + "grad_norm": 1.6238080263137817, + "learning_rate": 3.212835093419984e-06, + "loss": 2.1002, + "step": 1170 + }, + { + "epoch": 1.2821917808219179, + "eval_loss": 2.285175323486328, + "eval_runtime": 37.7569, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1170 + }, + { + "epoch": 1.2832876712328767, + "grad_norm": 1.631516933441162, + "learning_rate": 3.2108042242079613e-06, + "loss": 2.1094, + "step": 1171 + }, + { + "epoch": 1.2843835616438355, + "grad_norm": 1.6518285274505615, + "learning_rate": 3.2087733549959386e-06, + "loss": 2.1614, + "step": 1172 + }, + { + "epoch": 1.2854794520547945, + "grad_norm": 1.64463210105896, + "learning_rate": 3.206742485783916e-06, + "loss": 2.0863, + "step": 1173 + }, + { + "epoch": 1.2865753424657536, + "grad_norm": 1.5903335809707642, + "learning_rate": 3.2047116165718927e-06, + "loss": 2.0795, + "step": 1174 + }, + { + "epoch": 1.2876712328767124, + "grad_norm": 1.6504883766174316, + "learning_rate": 3.2026807473598704e-06, + "loss": 2.1352, + "step": 1175 + }, + { + "epoch": 1.2887671232876712, + "grad_norm": 1.7765742540359497, + "learning_rate": 3.2006498781478473e-06, + "loss": 2.1511, + "step": 1176 + }, + { + "epoch": 1.2898630136986302, + "grad_norm": 1.7253892421722412, + "learning_rate": 3.198619008935825e-06, + "loss": 2.0768, + "step": 1177 + }, + { + "epoch": 1.290958904109589, + "grad_norm": 1.7327924966812134, + "learning_rate": 3.196588139723802e-06, + "loss": 2.1196, + "step": 1178 + }, + { + "epoch": 1.2920547945205478, + "grad_norm": 1.6659244298934937, + "learning_rate": 3.1945572705117795e-06, + "loss": 2.1398, + "step": 1179 + }, + { + "epoch": 1.2931506849315069, + "grad_norm": 1.8803845643997192, + "learning_rate": 3.1925264012997564e-06, + "loss": 2.1382, + "step": 1180 + }, + { + "epoch": 1.2931506849315069, + "eval_loss": 2.284956455230713, + "eval_runtime": 37.7442, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1180 + }, + { + "epoch": 1.2942465753424657, + "grad_norm": 1.6562074422836304, + "learning_rate": 3.190495532087734e-06, + "loss": 2.1403, + "step": 1181 + }, + { + "epoch": 1.2953424657534247, + "grad_norm": 1.6407029628753662, + "learning_rate": 3.188464662875711e-06, + "loss": 2.1782, + "step": 1182 + }, + { + "epoch": 1.2964383561643835, + "grad_norm": 1.638464331626892, + "learning_rate": 3.1864337936636886e-06, + "loss": 2.1129, + "step": 1183 + }, + { + "epoch": 1.2975342465753426, + "grad_norm": 1.6085008382797241, + "learning_rate": 3.1844029244516655e-06, + "loss": 2.0967, + "step": 1184 + }, + { + "epoch": 1.2986301369863014, + "grad_norm": 1.9340208768844604, + "learning_rate": 3.182372055239643e-06, + "loss": 2.0906, + "step": 1185 + }, + { + "epoch": 1.2997260273972602, + "grad_norm": 1.6958277225494385, + "learning_rate": 3.18034118602762e-06, + "loss": 2.1519, + "step": 1186 + }, + { + "epoch": 1.3008219178082192, + "grad_norm": 1.6682950258255005, + "learning_rate": 3.1783103168155973e-06, + "loss": 2.1286, + "step": 1187 + }, + { + "epoch": 1.301917808219178, + "grad_norm": 1.7564821243286133, + "learning_rate": 3.1762794476035746e-06, + "loss": 2.1153, + "step": 1188 + }, + { + "epoch": 1.303013698630137, + "grad_norm": 1.765539526939392, + "learning_rate": 3.174248578391552e-06, + "loss": 2.112, + "step": 1189 + }, + { + "epoch": 1.3041095890410959, + "grad_norm": 1.8028367757797241, + "learning_rate": 3.172217709179529e-06, + "loss": 2.1169, + "step": 1190 + }, + { + "epoch": 1.3041095890410959, + "eval_loss": 2.2839744091033936, + "eval_runtime": 37.7782, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 1190 + }, + { + "epoch": 1.305205479452055, + "grad_norm": 1.6870243549346924, + "learning_rate": 3.1701868399675064e-06, + "loss": 2.1147, + "step": 1191 + }, + { + "epoch": 1.3063013698630137, + "grad_norm": 1.6161645650863647, + "learning_rate": 3.1681559707554837e-06, + "loss": 2.081, + "step": 1192 + }, + { + "epoch": 1.3073972602739725, + "grad_norm": 1.5438032150268555, + "learning_rate": 3.166125101543461e-06, + "loss": 2.139, + "step": 1193 + }, + { + "epoch": 1.3084931506849315, + "grad_norm": 1.6329416036605835, + "learning_rate": 3.1640942323314382e-06, + "loss": 2.1346, + "step": 1194 + }, + { + "epoch": 1.3095890410958904, + "grad_norm": 1.6933001279830933, + "learning_rate": 3.1620633631194155e-06, + "loss": 2.0906, + "step": 1195 + }, + { + "epoch": 1.3106849315068494, + "grad_norm": 1.5968612432479858, + "learning_rate": 3.1600324939073928e-06, + "loss": 2.1191, + "step": 1196 + }, + { + "epoch": 1.3117808219178082, + "grad_norm": 1.654739499092102, + "learning_rate": 3.15800162469537e-06, + "loss": 2.1253, + "step": 1197 + }, + { + "epoch": 1.3128767123287672, + "grad_norm": 1.6595433950424194, + "learning_rate": 3.155970755483347e-06, + "loss": 2.1326, + "step": 1198 + }, + { + "epoch": 1.313972602739726, + "grad_norm": 1.5709439516067505, + "learning_rate": 3.1539398862713246e-06, + "loss": 2.1427, + "step": 1199 + }, + { + "epoch": 1.3150684931506849, + "grad_norm": 1.7586764097213745, + "learning_rate": 3.1519090170593015e-06, + "loss": 2.1311, + "step": 1200 + }, + { + "epoch": 1.3150684931506849, + "eval_loss": 2.2803795337677, + "eval_runtime": 37.7752, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 1200 + }, + { + "epoch": 1.316164383561644, + "grad_norm": 1.675703763961792, + "learning_rate": 3.149878147847279e-06, + "loss": 2.1679, + "step": 1201 + }, + { + "epoch": 1.3172602739726027, + "grad_norm": 1.632204294204712, + "learning_rate": 3.147847278635256e-06, + "loss": 2.1889, + "step": 1202 + }, + { + "epoch": 1.3183561643835615, + "grad_norm": 1.6142032146453857, + "learning_rate": 3.1458164094232337e-06, + "loss": 2.1186, + "step": 1203 + }, + { + "epoch": 1.3194520547945205, + "grad_norm": 1.5145642757415771, + "learning_rate": 3.1437855402112106e-06, + "loss": 2.0699, + "step": 1204 + }, + { + "epoch": 1.3205479452054796, + "grad_norm": 1.6954951286315918, + "learning_rate": 3.1417546709991883e-06, + "loss": 2.1435, + "step": 1205 + }, + { + "epoch": 1.3216438356164384, + "grad_norm": 1.5725785493850708, + "learning_rate": 3.139723801787165e-06, + "loss": 2.1105, + "step": 1206 + }, + { + "epoch": 1.3227397260273972, + "grad_norm": 1.6296924352645874, + "learning_rate": 3.137692932575143e-06, + "loss": 2.1526, + "step": 1207 + }, + { + "epoch": 1.3238356164383562, + "grad_norm": 1.758583426475525, + "learning_rate": 3.1356620633631197e-06, + "loss": 2.0966, + "step": 1208 + }, + { + "epoch": 1.324931506849315, + "grad_norm": 1.6128432750701904, + "learning_rate": 3.1336311941510974e-06, + "loss": 2.1227, + "step": 1209 + }, + { + "epoch": 1.3260273972602739, + "grad_norm": 1.6251143217086792, + "learning_rate": 3.1316003249390742e-06, + "loss": 2.1557, + "step": 1210 + }, + { + "epoch": 1.3260273972602739, + "eval_loss": 2.277153253555298, + "eval_runtime": 37.4484, + "eval_samples_per_second": 26.703, + "eval_steps_per_second": 1.122, + "step": 1210 + }, + { + "epoch": 1.3271232876712329, + "grad_norm": 1.6224772930145264, + "learning_rate": 3.129569455727051e-06, + "loss": 2.1515, + "step": 1211 + }, + { + "epoch": 1.3282191780821917, + "grad_norm": 1.6408687829971313, + "learning_rate": 3.1275385865150288e-06, + "loss": 2.1415, + "step": 1212 + }, + { + "epoch": 1.3293150684931507, + "grad_norm": 1.5633060932159424, + "learning_rate": 3.1255077173030056e-06, + "loss": 2.1369, + "step": 1213 + }, + { + "epoch": 1.3304109589041095, + "grad_norm": 1.6440671682357788, + "learning_rate": 3.1234768480909833e-06, + "loss": 2.1302, + "step": 1214 + }, + { + "epoch": 1.3315068493150686, + "grad_norm": 1.6065458059310913, + "learning_rate": 3.12144597887896e-06, + "loss": 2.1359, + "step": 1215 + }, + { + "epoch": 1.3326027397260274, + "grad_norm": 1.6209193468093872, + "learning_rate": 3.119415109666938e-06, + "loss": 2.1669, + "step": 1216 + }, + { + "epoch": 1.3336986301369862, + "grad_norm": 1.745129942893982, + "learning_rate": 3.1173842404549147e-06, + "loss": 2.1318, + "step": 1217 + }, + { + "epoch": 1.3347945205479452, + "grad_norm": 1.6090255975723267, + "learning_rate": 3.1153533712428924e-06, + "loss": 2.1318, + "step": 1218 + }, + { + "epoch": 1.335890410958904, + "grad_norm": 1.5840481519699097, + "learning_rate": 3.1133225020308693e-06, + "loss": 2.152, + "step": 1219 + }, + { + "epoch": 1.336986301369863, + "grad_norm": 1.6378774642944336, + "learning_rate": 3.111291632818847e-06, + "loss": 2.1597, + "step": 1220 + }, + { + "epoch": 1.336986301369863, + "eval_loss": 2.2799103260040283, + "eval_runtime": 37.6598, + "eval_samples_per_second": 26.554, + "eval_steps_per_second": 1.115, + "step": 1220 + }, + { + "epoch": 1.3380821917808219, + "grad_norm": 1.7247775793075562, + "learning_rate": 3.109260763606824e-06, + "loss": 2.1328, + "step": 1221 + }, + { + "epoch": 1.339178082191781, + "grad_norm": 1.6469335556030273, + "learning_rate": 3.1072298943948015e-06, + "loss": 2.1434, + "step": 1222 + }, + { + "epoch": 1.3402739726027397, + "grad_norm": 1.5268198251724243, + "learning_rate": 3.1051990251827784e-06, + "loss": 2.1077, + "step": 1223 + }, + { + "epoch": 1.3413698630136985, + "grad_norm": 1.6413369178771973, + "learning_rate": 3.1031681559707557e-06, + "loss": 2.1583, + "step": 1224 + }, + { + "epoch": 1.3424657534246576, + "grad_norm": 1.7595359086990356, + "learning_rate": 3.101137286758733e-06, + "loss": 2.1625, + "step": 1225 + }, + { + "epoch": 1.3435616438356164, + "grad_norm": 1.6333140134811401, + "learning_rate": 3.09910641754671e-06, + "loss": 2.1555, + "step": 1226 + }, + { + "epoch": 1.3446575342465754, + "grad_norm": 1.8567262887954712, + "learning_rate": 3.0970755483346875e-06, + "loss": 2.1403, + "step": 1227 + }, + { + "epoch": 1.3457534246575342, + "grad_norm": 1.5342068672180176, + "learning_rate": 3.0950446791226648e-06, + "loss": 2.1686, + "step": 1228 + }, + { + "epoch": 1.3468493150684933, + "grad_norm": 1.6217644214630127, + "learning_rate": 3.093013809910642e-06, + "loss": 2.1345, + "step": 1229 + }, + { + "epoch": 1.347945205479452, + "grad_norm": 1.7755476236343384, + "learning_rate": 3.0909829406986193e-06, + "loss": 2.1769, + "step": 1230 + }, + { + "epoch": 1.347945205479452, + "eval_loss": 2.2792110443115234, + "eval_runtime": 37.6822, + "eval_samples_per_second": 26.538, + "eval_steps_per_second": 1.115, + "step": 1230 + }, + { + "epoch": 1.3490410958904109, + "grad_norm": 1.4750087261199951, + "learning_rate": 3.0889520714865966e-06, + "loss": 2.1501, + "step": 1231 + }, + { + "epoch": 1.35013698630137, + "grad_norm": 1.7657350301742554, + "learning_rate": 3.086921202274574e-06, + "loss": 2.1254, + "step": 1232 + }, + { + "epoch": 1.3512328767123287, + "grad_norm": 1.6447482109069824, + "learning_rate": 3.084890333062551e-06, + "loss": 2.1004, + "step": 1233 + }, + { + "epoch": 1.3523287671232875, + "grad_norm": 1.658786654472351, + "learning_rate": 3.0828594638505284e-06, + "loss": 2.1276, + "step": 1234 + }, + { + "epoch": 1.3534246575342466, + "grad_norm": 1.5626777410507202, + "learning_rate": 3.0808285946385057e-06, + "loss": 2.1067, + "step": 1235 + }, + { + "epoch": 1.3545205479452056, + "grad_norm": 1.7021020650863647, + "learning_rate": 3.078797725426483e-06, + "loss": 2.0656, + "step": 1236 + }, + { + "epoch": 1.3556164383561644, + "grad_norm": 1.5983797311782837, + "learning_rate": 3.07676685621446e-06, + "loss": 2.1319, + "step": 1237 + }, + { + "epoch": 1.3567123287671232, + "grad_norm": 1.5506508350372314, + "learning_rate": 3.0747359870024375e-06, + "loss": 2.1008, + "step": 1238 + }, + { + "epoch": 1.3578082191780823, + "grad_norm": 1.5998226404190063, + "learning_rate": 3.0727051177904144e-06, + "loss": 2.1102, + "step": 1239 + }, + { + "epoch": 1.358904109589041, + "grad_norm": 1.6222437620162964, + "learning_rate": 3.070674248578392e-06, + "loss": 2.1352, + "step": 1240 + }, + { + "epoch": 1.358904109589041, + "eval_loss": 2.278921127319336, + "eval_runtime": 37.6613, + "eval_samples_per_second": 26.552, + "eval_steps_per_second": 1.115, + "step": 1240 + }, + { + "epoch": 1.3599999999999999, + "grad_norm": 1.600468397140503, + "learning_rate": 3.068643379366369e-06, + "loss": 2.1115, + "step": 1241 + }, + { + "epoch": 1.361095890410959, + "grad_norm": 1.6324045658111572, + "learning_rate": 3.0666125101543466e-06, + "loss": 2.0887, + "step": 1242 + }, + { + "epoch": 1.362191780821918, + "grad_norm": 1.7005919218063354, + "learning_rate": 3.0645816409423235e-06, + "loss": 2.1344, + "step": 1243 + }, + { + "epoch": 1.3632876712328768, + "grad_norm": 1.6483793258666992, + "learning_rate": 3.062550771730301e-06, + "loss": 2.1165, + "step": 1244 + }, + { + "epoch": 1.3643835616438356, + "grad_norm": 1.774255633354187, + "learning_rate": 3.060519902518278e-06, + "loss": 2.1137, + "step": 1245 + }, + { + "epoch": 1.3654794520547946, + "grad_norm": 1.7157840728759766, + "learning_rate": 3.0584890333062557e-06, + "loss": 2.0656, + "step": 1246 + }, + { + "epoch": 1.3665753424657534, + "grad_norm": 1.6053105592727661, + "learning_rate": 3.0564581640942326e-06, + "loss": 2.128, + "step": 1247 + }, + { + "epoch": 1.3676712328767122, + "grad_norm": 1.8893406391143799, + "learning_rate": 3.0544272948822094e-06, + "loss": 2.106, + "step": 1248 + }, + { + "epoch": 1.3687671232876713, + "grad_norm": 1.6212185621261597, + "learning_rate": 3.052396425670187e-06, + "loss": 2.1422, + "step": 1249 + }, + { + "epoch": 1.36986301369863, + "grad_norm": 1.595241665840149, + "learning_rate": 3.050365556458164e-06, + "loss": 2.1132, + "step": 1250 + }, + { + "epoch": 1.36986301369863, + "eval_loss": 2.280195474624634, + "eval_runtime": 37.69, + "eval_samples_per_second": 26.532, + "eval_steps_per_second": 1.114, + "step": 1250 + }, + { + "epoch": 1.370958904109589, + "grad_norm": 1.7002182006835938, + "learning_rate": 3.0483346872461417e-06, + "loss": 2.1088, + "step": 1251 + }, + { + "epoch": 1.372054794520548, + "grad_norm": 1.7393778562545776, + "learning_rate": 3.0463038180341185e-06, + "loss": 2.1371, + "step": 1252 + }, + { + "epoch": 1.373150684931507, + "grad_norm": 1.5694235563278198, + "learning_rate": 3.0442729488220962e-06, + "loss": 2.1514, + "step": 1253 + }, + { + "epoch": 1.3742465753424657, + "grad_norm": 1.689393401145935, + "learning_rate": 3.042242079610073e-06, + "loss": 2.1512, + "step": 1254 + }, + { + "epoch": 1.3753424657534246, + "grad_norm": 1.6431982517242432, + "learning_rate": 3.0402112103980508e-06, + "loss": 2.0749, + "step": 1255 + }, + { + "epoch": 1.3764383561643836, + "grad_norm": 1.6122742891311646, + "learning_rate": 3.0381803411860276e-06, + "loss": 2.1024, + "step": 1256 + }, + { + "epoch": 1.3775342465753424, + "grad_norm": 1.5770806074142456, + "learning_rate": 3.0361494719740053e-06, + "loss": 2.1208, + "step": 1257 + }, + { + "epoch": 1.3786301369863014, + "grad_norm": 1.7665551900863647, + "learning_rate": 3.034118602761982e-06, + "loss": 2.1209, + "step": 1258 + }, + { + "epoch": 1.3797260273972602, + "grad_norm": 1.6896030902862549, + "learning_rate": 3.03208773354996e-06, + "loss": 2.0986, + "step": 1259 + }, + { + "epoch": 1.3808219178082193, + "grad_norm": 1.7461376190185547, + "learning_rate": 3.0300568643379367e-06, + "loss": 2.156, + "step": 1260 + }, + { + "epoch": 1.3808219178082193, + "eval_loss": 2.27716064453125, + "eval_runtime": 37.7183, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 1260 + }, + { + "epoch": 1.381917808219178, + "grad_norm": 1.6952084302902222, + "learning_rate": 3.028025995125914e-06, + "loss": 2.1449, + "step": 1261 + }, + { + "epoch": 1.383013698630137, + "grad_norm": 1.5834839344024658, + "learning_rate": 3.0259951259138913e-06, + "loss": 2.082, + "step": 1262 + }, + { + "epoch": 1.384109589041096, + "grad_norm": 1.7395371198654175, + "learning_rate": 3.0239642567018686e-06, + "loss": 2.126, + "step": 1263 + }, + { + "epoch": 1.3852054794520547, + "grad_norm": 1.6316404342651367, + "learning_rate": 3.021933387489846e-06, + "loss": 2.0833, + "step": 1264 + }, + { + "epoch": 1.3863013698630138, + "grad_norm": 1.6206353902816772, + "learning_rate": 3.019902518277823e-06, + "loss": 2.1311, + "step": 1265 + }, + { + "epoch": 1.3873972602739726, + "grad_norm": 1.632569432258606, + "learning_rate": 3.0178716490658004e-06, + "loss": 2.1468, + "step": 1266 + }, + { + "epoch": 1.3884931506849316, + "grad_norm": 1.64067804813385, + "learning_rate": 3.0158407798537777e-06, + "loss": 2.0859, + "step": 1267 + }, + { + "epoch": 1.3895890410958904, + "grad_norm": 1.54572331905365, + "learning_rate": 3.013809910641755e-06, + "loss": 2.0992, + "step": 1268 + }, + { + "epoch": 1.3906849315068492, + "grad_norm": 1.5950230360031128, + "learning_rate": 3.0117790414297322e-06, + "loss": 2.0895, + "step": 1269 + }, + { + "epoch": 1.3917808219178083, + "grad_norm": 1.8030891418457031, + "learning_rate": 3.0097481722177095e-06, + "loss": 2.1165, + "step": 1270 + }, + { + "epoch": 1.3917808219178083, + "eval_loss": 2.2781739234924316, + "eval_runtime": 37.7428, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 1270 + }, + { + "epoch": 1.392876712328767, + "grad_norm": 1.5848240852355957, + "learning_rate": 3.0077173030056868e-06, + "loss": 2.1015, + "step": 1271 + }, + { + "epoch": 1.393972602739726, + "grad_norm": 1.682953953742981, + "learning_rate": 3.005686433793664e-06, + "loss": 2.1341, + "step": 1272 + }, + { + "epoch": 1.395068493150685, + "grad_norm": 1.8385032415390015, + "learning_rate": 3.0036555645816413e-06, + "loss": 2.111, + "step": 1273 + }, + { + "epoch": 1.396164383561644, + "grad_norm": 1.5879863500595093, + "learning_rate": 3.001624695369618e-06, + "loss": 2.1418, + "step": 1274 + }, + { + "epoch": 1.3972602739726028, + "grad_norm": 1.8076952695846558, + "learning_rate": 2.999593826157596e-06, + "loss": 2.1076, + "step": 1275 + }, + { + "epoch": 1.3983561643835616, + "grad_norm": 1.6944698095321655, + "learning_rate": 2.9975629569455727e-06, + "loss": 2.1415, + "step": 1276 + }, + { + "epoch": 1.3994520547945206, + "grad_norm": 1.7351915836334229, + "learning_rate": 2.9955320877335504e-06, + "loss": 2.1186, + "step": 1277 + }, + { + "epoch": 1.4005479452054794, + "grad_norm": 1.6877539157867432, + "learning_rate": 2.9935012185215273e-06, + "loss": 2.1374, + "step": 1278 + }, + { + "epoch": 1.4016438356164382, + "grad_norm": 1.6253048181533813, + "learning_rate": 2.991470349309505e-06, + "loss": 2.114, + "step": 1279 + }, + { + "epoch": 1.4027397260273973, + "grad_norm": 1.6599102020263672, + "learning_rate": 2.989439480097482e-06, + "loss": 2.1231, + "step": 1280 + }, + { + "epoch": 1.4027397260273973, + "eval_loss": 2.2765634059906006, + "eval_runtime": 37.7438, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 1280 + }, + { + "epoch": 1.403835616438356, + "grad_norm": 1.6010860204696655, + "learning_rate": 2.9874086108854595e-06, + "loss": 2.108, + "step": 1281 + }, + { + "epoch": 1.4049315068493151, + "grad_norm": 1.7082704305648804, + "learning_rate": 2.9853777416734364e-06, + "loss": 2.143, + "step": 1282 + }, + { + "epoch": 1.406027397260274, + "grad_norm": 1.5962858200073242, + "learning_rate": 2.983346872461414e-06, + "loss": 2.1582, + "step": 1283 + }, + { + "epoch": 1.407123287671233, + "grad_norm": 1.678695559501648, + "learning_rate": 2.981316003249391e-06, + "loss": 2.1308, + "step": 1284 + }, + { + "epoch": 1.4082191780821918, + "grad_norm": 1.8608810901641846, + "learning_rate": 2.9792851340373678e-06, + "loss": 2.1192, + "step": 1285 + }, + { + "epoch": 1.4093150684931506, + "grad_norm": 1.6447912454605103, + "learning_rate": 2.9772542648253455e-06, + "loss": 2.149, + "step": 1286 + }, + { + "epoch": 1.4104109589041096, + "grad_norm": 1.8847064971923828, + "learning_rate": 2.9752233956133223e-06, + "loss": 2.1073, + "step": 1287 + }, + { + "epoch": 1.4115068493150684, + "grad_norm": 1.76127290725708, + "learning_rate": 2.9731925264013e-06, + "loss": 2.1123, + "step": 1288 + }, + { + "epoch": 1.4126027397260275, + "grad_norm": 1.8126407861709595, + "learning_rate": 2.971161657189277e-06, + "loss": 2.1123, + "step": 1289 + }, + { + "epoch": 1.4136986301369863, + "grad_norm": 1.8271949291229248, + "learning_rate": 2.9691307879772546e-06, + "loss": 2.1156, + "step": 1290 + }, + { + "epoch": 1.4136986301369863, + "eval_loss": 2.2737486362457275, + "eval_runtime": 37.6964, + "eval_samples_per_second": 26.528, + "eval_steps_per_second": 1.114, + "step": 1290 + }, + { + "epoch": 1.4147945205479453, + "grad_norm": 1.6088656187057495, + "learning_rate": 2.9670999187652314e-06, + "loss": 2.0904, + "step": 1291 + }, + { + "epoch": 1.415890410958904, + "grad_norm": 1.8934502601623535, + "learning_rate": 2.965069049553209e-06, + "loss": 2.1669, + "step": 1292 + }, + { + "epoch": 1.416986301369863, + "grad_norm": 1.770875334739685, + "learning_rate": 2.963038180341186e-06, + "loss": 2.0866, + "step": 1293 + }, + { + "epoch": 1.418082191780822, + "grad_norm": 1.6065635681152344, + "learning_rate": 2.9610073111291637e-06, + "loss": 2.1433, + "step": 1294 + }, + { + "epoch": 1.4191780821917808, + "grad_norm": 1.8257594108581543, + "learning_rate": 2.9589764419171405e-06, + "loss": 2.1611, + "step": 1295 + }, + { + "epoch": 1.4202739726027398, + "grad_norm": 1.7531297206878662, + "learning_rate": 2.9569455727051182e-06, + "loss": 2.0991, + "step": 1296 + }, + { + "epoch": 1.4213698630136986, + "grad_norm": 1.6781302690505981, + "learning_rate": 2.954914703493095e-06, + "loss": 2.121, + "step": 1297 + }, + { + "epoch": 1.4224657534246576, + "grad_norm": 1.7489228248596191, + "learning_rate": 2.9528838342810724e-06, + "loss": 2.1012, + "step": 1298 + }, + { + "epoch": 1.4235616438356165, + "grad_norm": 1.7776473760604858, + "learning_rate": 2.9508529650690496e-06, + "loss": 2.1066, + "step": 1299 + }, + { + "epoch": 1.4246575342465753, + "grad_norm": 1.7317523956298828, + "learning_rate": 2.948822095857027e-06, + "loss": 2.1278, + "step": 1300 + }, + { + "epoch": 1.4246575342465753, + "eval_loss": 2.2755467891693115, + "eval_runtime": 37.7024, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 1300 + }, + { + "epoch": 1.4257534246575343, + "grad_norm": 1.9439119100570679, + "learning_rate": 2.946791226645004e-06, + "loss": 2.0971, + "step": 1301 + }, + { + "epoch": 1.426849315068493, + "grad_norm": 1.683897614479065, + "learning_rate": 2.9447603574329815e-06, + "loss": 2.0895, + "step": 1302 + }, + { + "epoch": 1.427945205479452, + "grad_norm": 1.620665192604065, + "learning_rate": 2.9427294882209587e-06, + "loss": 2.1011, + "step": 1303 + }, + { + "epoch": 1.429041095890411, + "grad_norm": 1.9428532123565674, + "learning_rate": 2.940698619008936e-06, + "loss": 2.1253, + "step": 1304 + }, + { + "epoch": 1.43013698630137, + "grad_norm": 1.897630214691162, + "learning_rate": 2.9386677497969133e-06, + "loss": 2.1146, + "step": 1305 + }, + { + "epoch": 1.4312328767123288, + "grad_norm": 1.767519235610962, + "learning_rate": 2.9366368805848906e-06, + "loss": 2.1407, + "step": 1306 + }, + { + "epoch": 1.4323287671232876, + "grad_norm": 1.630615234375, + "learning_rate": 2.934606011372868e-06, + "loss": 2.1432, + "step": 1307 + }, + { + "epoch": 1.4334246575342466, + "grad_norm": 1.7037104368209839, + "learning_rate": 2.932575142160845e-06, + "loss": 2.1226, + "step": 1308 + }, + { + "epoch": 1.4345205479452054, + "grad_norm": 1.5496717691421509, + "learning_rate": 2.9305442729488224e-06, + "loss": 2.1235, + "step": 1309 + }, + { + "epoch": 1.4356164383561643, + "grad_norm": 1.6394745111465454, + "learning_rate": 2.9285134037367997e-06, + "loss": 2.1183, + "step": 1310 + }, + { + "epoch": 1.4356164383561643, + "eval_loss": 2.2730512619018555, + "eval_runtime": 37.4684, + "eval_samples_per_second": 26.689, + "eval_steps_per_second": 1.121, + "step": 1310 + }, + { + "epoch": 1.4367123287671233, + "grad_norm": 1.5624728202819824, + "learning_rate": 2.9264825345247765e-06, + "loss": 2.1297, + "step": 1311 + }, + { + "epoch": 1.4378082191780823, + "grad_norm": 1.7173740863800049, + "learning_rate": 2.9244516653127542e-06, + "loss": 2.1299, + "step": 1312 + }, + { + "epoch": 1.4389041095890411, + "grad_norm": 1.5682451725006104, + "learning_rate": 2.922420796100731e-06, + "loss": 2.0605, + "step": 1313 + }, + { + "epoch": 1.44, + "grad_norm": 1.6592832803726196, + "learning_rate": 2.9203899268887088e-06, + "loss": 2.1721, + "step": 1314 + }, + { + "epoch": 1.441095890410959, + "grad_norm": 1.7420763969421387, + "learning_rate": 2.9183590576766856e-06, + "loss": 2.0873, + "step": 1315 + }, + { + "epoch": 1.4421917808219178, + "grad_norm": 1.6949102878570557, + "learning_rate": 2.9163281884646633e-06, + "loss": 2.1213, + "step": 1316 + }, + { + "epoch": 1.4432876712328766, + "grad_norm": 1.7273104190826416, + "learning_rate": 2.91429731925264e-06, + "loss": 2.1501, + "step": 1317 + }, + { + "epoch": 1.4443835616438356, + "grad_norm": 1.684362530708313, + "learning_rate": 2.912266450040618e-06, + "loss": 2.1252, + "step": 1318 + }, + { + "epoch": 1.4454794520547944, + "grad_norm": 1.627016305923462, + "learning_rate": 2.9102355808285947e-06, + "loss": 2.1039, + "step": 1319 + }, + { + "epoch": 1.4465753424657535, + "grad_norm": 1.8114672899246216, + "learning_rate": 2.9082047116165724e-06, + "loss": 2.1049, + "step": 1320 + }, + { + "epoch": 1.4465753424657535, + "eval_loss": 2.273300886154175, + "eval_runtime": 37.6816, + "eval_samples_per_second": 26.538, + "eval_steps_per_second": 1.115, + "step": 1320 + }, + { + "epoch": 1.4476712328767123, + "grad_norm": 1.6053475141525269, + "learning_rate": 2.9061738424045493e-06, + "loss": 2.1026, + "step": 1321 + }, + { + "epoch": 1.4487671232876713, + "grad_norm": 1.757144570350647, + "learning_rate": 2.904142973192527e-06, + "loss": 2.1683, + "step": 1322 + }, + { + "epoch": 1.4498630136986301, + "grad_norm": 1.6230170726776123, + "learning_rate": 2.902112103980504e-06, + "loss": 2.1308, + "step": 1323 + }, + { + "epoch": 1.450958904109589, + "grad_norm": 1.5634756088256836, + "learning_rate": 2.900081234768481e-06, + "loss": 2.1194, + "step": 1324 + }, + { + "epoch": 1.452054794520548, + "grad_norm": 1.853056788444519, + "learning_rate": 2.8980503655564584e-06, + "loss": 2.1135, + "step": 1325 + }, + { + "epoch": 1.4531506849315068, + "grad_norm": 1.6028629541397095, + "learning_rate": 2.8960194963444357e-06, + "loss": 2.0872, + "step": 1326 + }, + { + "epoch": 1.4542465753424658, + "grad_norm": 1.657346487045288, + "learning_rate": 2.893988627132413e-06, + "loss": 2.1059, + "step": 1327 + }, + { + "epoch": 1.4553424657534246, + "grad_norm": 1.6872131824493408, + "learning_rate": 2.8919577579203902e-06, + "loss": 2.0726, + "step": 1328 + }, + { + "epoch": 1.4564383561643837, + "grad_norm": 1.677079439163208, + "learning_rate": 2.8899268887083675e-06, + "loss": 2.1199, + "step": 1329 + }, + { + "epoch": 1.4575342465753425, + "grad_norm": 1.6555675268173218, + "learning_rate": 2.8878960194963448e-06, + "loss": 2.1231, + "step": 1330 + }, + { + "epoch": 1.4575342465753425, + "eval_loss": 2.277087688446045, + "eval_runtime": 37.7399, + "eval_samples_per_second": 26.497, + "eval_steps_per_second": 1.113, + "step": 1330 + }, + { + "epoch": 1.4586301369863013, + "grad_norm": 1.6812067031860352, + "learning_rate": 2.885865150284322e-06, + "loss": 2.1397, + "step": 1331 + }, + { + "epoch": 1.4597260273972603, + "grad_norm": 1.5711122751235962, + "learning_rate": 2.8838342810722993e-06, + "loss": 2.1005, + "step": 1332 + }, + { + "epoch": 1.4608219178082191, + "grad_norm": 1.7067155838012695, + "learning_rate": 2.8818034118602766e-06, + "loss": 2.0874, + "step": 1333 + }, + { + "epoch": 1.4619178082191782, + "grad_norm": 1.7551555633544922, + "learning_rate": 2.879772542648254e-06, + "loss": 2.1376, + "step": 1334 + }, + { + "epoch": 1.463013698630137, + "grad_norm": 1.6309247016906738, + "learning_rate": 2.8777416734362307e-06, + "loss": 2.0595, + "step": 1335 + }, + { + "epoch": 1.464109589041096, + "grad_norm": 1.6252689361572266, + "learning_rate": 2.8757108042242084e-06, + "loss": 2.1778, + "step": 1336 + }, + { + "epoch": 1.4652054794520548, + "grad_norm": 1.5928912162780762, + "learning_rate": 2.8736799350121853e-06, + "loss": 2.1575, + "step": 1337 + }, + { + "epoch": 1.4663013698630136, + "grad_norm": 1.6079543828964233, + "learning_rate": 2.871649065800163e-06, + "loss": 2.1638, + "step": 1338 + }, + { + "epoch": 1.4673972602739727, + "grad_norm": 1.628999948501587, + "learning_rate": 2.86961819658814e-06, + "loss": 2.0999, + "step": 1339 + }, + { + "epoch": 1.4684931506849315, + "grad_norm": 1.5877981185913086, + "learning_rate": 2.8675873273761175e-06, + "loss": 2.1282, + "step": 1340 + }, + { + "epoch": 1.4684931506849315, + "eval_loss": 2.275458574295044, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 1340 + }, + { + "epoch": 1.4695890410958903, + "grad_norm": 1.6038731336593628, + "learning_rate": 2.8655564581640944e-06, + "loss": 2.1164, + "step": 1341 + }, + { + "epoch": 1.4706849315068493, + "grad_norm": 1.6492732763290405, + "learning_rate": 2.863525588952072e-06, + "loss": 2.1471, + "step": 1342 + }, + { + "epoch": 1.4717808219178083, + "grad_norm": 1.647605538368225, + "learning_rate": 2.861494719740049e-06, + "loss": 2.0737, + "step": 1343 + }, + { + "epoch": 1.4728767123287672, + "grad_norm": 1.5875589847564697, + "learning_rate": 2.8594638505280266e-06, + "loss": 2.0784, + "step": 1344 + }, + { + "epoch": 1.473972602739726, + "grad_norm": 1.660346508026123, + "learning_rate": 2.8574329813160035e-06, + "loss": 2.1166, + "step": 1345 + }, + { + "epoch": 1.475068493150685, + "grad_norm": 1.5677483081817627, + "learning_rate": 2.855402112103981e-06, + "loss": 2.1068, + "step": 1346 + }, + { + "epoch": 1.4761643835616438, + "grad_norm": 1.6814773082733154, + "learning_rate": 2.853371242891958e-06, + "loss": 2.1118, + "step": 1347 + }, + { + "epoch": 1.4772602739726026, + "grad_norm": 1.7251765727996826, + "learning_rate": 2.851340373679935e-06, + "loss": 2.1255, + "step": 1348 + }, + { + "epoch": 1.4783561643835617, + "grad_norm": 1.6479768753051758, + "learning_rate": 2.8493095044679126e-06, + "loss": 2.1675, + "step": 1349 + }, + { + "epoch": 1.4794520547945205, + "grad_norm": 1.5761340856552124, + "learning_rate": 2.8472786352558894e-06, + "loss": 2.1271, + "step": 1350 + }, + { + "epoch": 1.4794520547945205, + "eval_loss": 2.2767348289489746, + "eval_runtime": 37.7124, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 1350 + }, + { + "epoch": 1.4805479452054795, + "grad_norm": 1.577124834060669, + "learning_rate": 2.845247766043867e-06, + "loss": 2.0968, + "step": 1351 + }, + { + "epoch": 1.4816438356164383, + "grad_norm": 1.6901545524597168, + "learning_rate": 2.843216896831844e-06, + "loss": 2.092, + "step": 1352 + }, + { + "epoch": 1.4827397260273973, + "grad_norm": 1.582205891609192, + "learning_rate": 2.8411860276198217e-06, + "loss": 2.1067, + "step": 1353 + }, + { + "epoch": 1.4838356164383562, + "grad_norm": 1.6398444175720215, + "learning_rate": 2.8391551584077985e-06, + "loss": 2.1524, + "step": 1354 + }, + { + "epoch": 1.484931506849315, + "grad_norm": 1.6396856307983398, + "learning_rate": 2.8371242891957762e-06, + "loss": 2.1523, + "step": 1355 + }, + { + "epoch": 1.486027397260274, + "grad_norm": 1.7187350988388062, + "learning_rate": 2.835093419983753e-06, + "loss": 2.1186, + "step": 1356 + }, + { + "epoch": 1.4871232876712328, + "grad_norm": 1.6155378818511963, + "learning_rate": 2.833062550771731e-06, + "loss": 2.1119, + "step": 1357 + }, + { + "epoch": 1.4882191780821918, + "grad_norm": 1.6749935150146484, + "learning_rate": 2.8310316815597076e-06, + "loss": 2.0871, + "step": 1358 + }, + { + "epoch": 1.4893150684931507, + "grad_norm": 1.7644442319869995, + "learning_rate": 2.8290008123476853e-06, + "loss": 2.1296, + "step": 1359 + }, + { + "epoch": 1.4904109589041097, + "grad_norm": 1.7644187211990356, + "learning_rate": 2.826969943135662e-06, + "loss": 2.1168, + "step": 1360 + }, + { + "epoch": 1.4904109589041097, + "eval_loss": 2.275988817214966, + "eval_runtime": 37.7569, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1360 + }, + { + "epoch": 1.4915068493150685, + "grad_norm": 1.6250940561294556, + "learning_rate": 2.8249390739236395e-06, + "loss": 2.1126, + "step": 1361 + }, + { + "epoch": 1.4926027397260273, + "grad_norm": 1.545547366142273, + "learning_rate": 2.8229082047116167e-06, + "loss": 2.1014, + "step": 1362 + }, + { + "epoch": 1.4936986301369863, + "grad_norm": 1.6356806755065918, + "learning_rate": 2.820877335499594e-06, + "loss": 2.0994, + "step": 1363 + }, + { + "epoch": 1.4947945205479451, + "grad_norm": 1.6093254089355469, + "learning_rate": 2.8188464662875713e-06, + "loss": 2.0638, + "step": 1364 + }, + { + "epoch": 1.4958904109589042, + "grad_norm": 1.531909465789795, + "learning_rate": 2.8168155970755486e-06, + "loss": 2.0873, + "step": 1365 + }, + { + "epoch": 1.496986301369863, + "grad_norm": 1.7620573043823242, + "learning_rate": 2.814784727863526e-06, + "loss": 2.1124, + "step": 1366 + }, + { + "epoch": 1.498082191780822, + "grad_norm": 1.7096633911132812, + "learning_rate": 2.812753858651503e-06, + "loss": 2.1308, + "step": 1367 + }, + { + "epoch": 1.4991780821917808, + "grad_norm": 1.8180763721466064, + "learning_rate": 2.8107229894394804e-06, + "loss": 2.0961, + "step": 1368 + }, + { + "epoch": 1.5002739726027396, + "grad_norm": 1.7475993633270264, + "learning_rate": 2.8086921202274577e-06, + "loss": 2.0784, + "step": 1369 + }, + { + "epoch": 1.5013698630136987, + "grad_norm": 1.7892746925354004, + "learning_rate": 2.806661251015435e-06, + "loss": 2.1298, + "step": 1370 + }, + { + "epoch": 1.5013698630136987, + "eval_loss": 2.2772014141082764, + "eval_runtime": 37.709, + "eval_samples_per_second": 26.519, + "eval_steps_per_second": 1.114, + "step": 1370 + }, + { + "epoch": 1.5024657534246575, + "grad_norm": 1.700134038925171, + "learning_rate": 2.8046303818034122e-06, + "loss": 2.1571, + "step": 1371 + }, + { + "epoch": 1.5035616438356163, + "grad_norm": 1.7968006134033203, + "learning_rate": 2.802599512591389e-06, + "loss": 2.2039, + "step": 1372 + }, + { + "epoch": 1.5046575342465753, + "grad_norm": 1.5552955865859985, + "learning_rate": 2.8005686433793668e-06, + "loss": 2.1078, + "step": 1373 + }, + { + "epoch": 1.5057534246575344, + "grad_norm": 1.6243152618408203, + "learning_rate": 2.7985377741673436e-06, + "loss": 2.071, + "step": 1374 + }, + { + "epoch": 1.5068493150684932, + "grad_norm": 1.7453999519348145, + "learning_rate": 2.7965069049553213e-06, + "loss": 2.1228, + "step": 1375 + }, + { + "epoch": 1.507945205479452, + "grad_norm": 1.6567325592041016, + "learning_rate": 2.794476035743298e-06, + "loss": 2.158, + "step": 1376 + }, + { + "epoch": 1.509041095890411, + "grad_norm": 1.6160842180252075, + "learning_rate": 2.792445166531276e-06, + "loss": 2.1231, + "step": 1377 + }, + { + "epoch": 1.5101369863013698, + "grad_norm": 1.715888500213623, + "learning_rate": 2.7904142973192527e-06, + "loss": 2.1466, + "step": 1378 + }, + { + "epoch": 1.5112328767123286, + "grad_norm": 1.6263453960418701, + "learning_rate": 2.7883834281072304e-06, + "loss": 2.0926, + "step": 1379 + }, + { + "epoch": 1.5123287671232877, + "grad_norm": 1.7616580724716187, + "learning_rate": 2.7863525588952073e-06, + "loss": 2.0877, + "step": 1380 + }, + { + "epoch": 1.5123287671232877, + "eval_loss": 2.275289535522461, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 1380 + }, + { + "epoch": 1.5134246575342467, + "grad_norm": 1.6861932277679443, + "learning_rate": 2.784321689683185e-06, + "loss": 2.0867, + "step": 1381 + }, + { + "epoch": 1.5145205479452055, + "grad_norm": 1.570162296295166, + "learning_rate": 2.782290820471162e-06, + "loss": 2.0957, + "step": 1382 + }, + { + "epoch": 1.5156164383561643, + "grad_norm": 1.698014259338379, + "learning_rate": 2.7802599512591395e-06, + "loss": 2.1427, + "step": 1383 + }, + { + "epoch": 1.5167123287671234, + "grad_norm": 1.6773440837860107, + "learning_rate": 2.7782290820471164e-06, + "loss": 2.135, + "step": 1384 + }, + { + "epoch": 1.5178082191780822, + "grad_norm": 1.6762758493423462, + "learning_rate": 2.7761982128350932e-06, + "loss": 2.132, + "step": 1385 + }, + { + "epoch": 1.518904109589041, + "grad_norm": 1.5917284488677979, + "learning_rate": 2.774167343623071e-06, + "loss": 2.1616, + "step": 1386 + }, + { + "epoch": 1.52, + "grad_norm": 1.583791732788086, + "learning_rate": 2.772136474411048e-06, + "loss": 2.0901, + "step": 1387 + }, + { + "epoch": 1.521095890410959, + "grad_norm": 1.6780498027801514, + "learning_rate": 2.7701056051990255e-06, + "loss": 2.0577, + "step": 1388 + }, + { + "epoch": 1.5221917808219179, + "grad_norm": 1.6320325136184692, + "learning_rate": 2.7680747359870023e-06, + "loss": 2.1381, + "step": 1389 + }, + { + "epoch": 1.5232876712328767, + "grad_norm": 1.7060813903808594, + "learning_rate": 2.76604386677498e-06, + "loss": 2.115, + "step": 1390 + }, + { + "epoch": 1.5232876712328767, + "eval_loss": 2.27346134185791, + "eval_runtime": 37.778, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 1390 + }, + { + "epoch": 1.5243835616438357, + "grad_norm": 1.6142189502716064, + "learning_rate": 2.764012997562957e-06, + "loss": 2.1388, + "step": 1391 + }, + { + "epoch": 1.5254794520547945, + "grad_norm": 1.6605664491653442, + "learning_rate": 2.7619821283509346e-06, + "loss": 2.0784, + "step": 1392 + }, + { + "epoch": 1.5265753424657533, + "grad_norm": 1.7281928062438965, + "learning_rate": 2.7599512591389114e-06, + "loss": 2.1461, + "step": 1393 + }, + { + "epoch": 1.5276712328767124, + "grad_norm": 1.6661336421966553, + "learning_rate": 2.757920389926889e-06, + "loss": 2.0987, + "step": 1394 + }, + { + "epoch": 1.5287671232876714, + "grad_norm": 1.6454191207885742, + "learning_rate": 2.755889520714866e-06, + "loss": 2.0912, + "step": 1395 + }, + { + "epoch": 1.52986301369863, + "grad_norm": 1.6793323755264282, + "learning_rate": 2.7538586515028437e-06, + "loss": 2.0642, + "step": 1396 + }, + { + "epoch": 1.530958904109589, + "grad_norm": 1.5941072702407837, + "learning_rate": 2.7518277822908205e-06, + "loss": 2.1232, + "step": 1397 + }, + { + "epoch": 1.532054794520548, + "grad_norm": 1.5669342279434204, + "learning_rate": 2.749796913078798e-06, + "loss": 2.1406, + "step": 1398 + }, + { + "epoch": 1.5331506849315069, + "grad_norm": 1.5805522203445435, + "learning_rate": 2.747766043866775e-06, + "loss": 2.1096, + "step": 1399 + }, + { + "epoch": 1.5342465753424657, + "grad_norm": 1.6227574348449707, + "learning_rate": 2.7457351746547524e-06, + "loss": 2.15, + "step": 1400 + }, + { + "epoch": 1.5342465753424657, + "eval_loss": 2.27066707611084, + "eval_runtime": 37.708, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 1400 + }, + { + "epoch": 1.5353424657534247, + "grad_norm": 1.6500868797302246, + "learning_rate": 2.7437043054427297e-06, + "loss": 2.193, + "step": 1401 + }, + { + "epoch": 1.5364383561643835, + "grad_norm": 1.6337374448776245, + "learning_rate": 2.741673436230707e-06, + "loss": 2.1196, + "step": 1402 + }, + { + "epoch": 1.5375342465753423, + "grad_norm": 1.5784432888031006, + "learning_rate": 2.739642567018684e-06, + "loss": 2.0906, + "step": 1403 + }, + { + "epoch": 1.5386301369863014, + "grad_norm": 1.75803542137146, + "learning_rate": 2.7376116978066615e-06, + "loss": 2.124, + "step": 1404 + }, + { + "epoch": 1.5397260273972604, + "grad_norm": 1.7682627439498901, + "learning_rate": 2.7355808285946388e-06, + "loss": 2.121, + "step": 1405 + }, + { + "epoch": 1.5408219178082192, + "grad_norm": 1.7324312925338745, + "learning_rate": 2.733549959382616e-06, + "loss": 2.0975, + "step": 1406 + }, + { + "epoch": 1.541917808219178, + "grad_norm": 1.765071153640747, + "learning_rate": 2.7315190901705933e-06, + "loss": 2.1668, + "step": 1407 + }, + { + "epoch": 1.543013698630137, + "grad_norm": 1.7210100889205933, + "learning_rate": 2.7294882209585706e-06, + "loss": 2.1284, + "step": 1408 + }, + { + "epoch": 1.5441095890410959, + "grad_norm": 1.7064242362976074, + "learning_rate": 2.727457351746548e-06, + "loss": 2.1007, + "step": 1409 + }, + { + "epoch": 1.5452054794520547, + "grad_norm": 1.7347673177719116, + "learning_rate": 2.725426482534525e-06, + "loss": 2.0908, + "step": 1410 + }, + { + "epoch": 1.5452054794520547, + "eval_loss": 2.273494005203247, + "eval_runtime": 37.5003, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 1410 + }, + { + "epoch": 1.5463013698630137, + "grad_norm": 1.6207059621810913, + "learning_rate": 2.723395613322502e-06, + "loss": 2.1238, + "step": 1411 + }, + { + "epoch": 1.5473972602739727, + "grad_norm": 1.6258277893066406, + "learning_rate": 2.7213647441104797e-06, + "loss": 2.107, + "step": 1412 + }, + { + "epoch": 1.5484931506849315, + "grad_norm": 1.8286561965942383, + "learning_rate": 2.7193338748984565e-06, + "loss": 2.1198, + "step": 1413 + }, + { + "epoch": 1.5495890410958904, + "grad_norm": 1.6007434129714966, + "learning_rate": 2.7173030056864342e-06, + "loss": 2.2353, + "step": 1414 + }, + { + "epoch": 1.5506849315068494, + "grad_norm": 1.665379285812378, + "learning_rate": 2.715272136474411e-06, + "loss": 2.0716, + "step": 1415 + }, + { + "epoch": 1.5517808219178082, + "grad_norm": 1.6235698461532593, + "learning_rate": 2.713241267262389e-06, + "loss": 2.1239, + "step": 1416 + }, + { + "epoch": 1.552876712328767, + "grad_norm": 1.6042425632476807, + "learning_rate": 2.7112103980503656e-06, + "loss": 2.1033, + "step": 1417 + }, + { + "epoch": 1.553972602739726, + "grad_norm": 1.6660529375076294, + "learning_rate": 2.7091795288383433e-06, + "loss": 2.1326, + "step": 1418 + }, + { + "epoch": 1.555068493150685, + "grad_norm": 1.7682642936706543, + "learning_rate": 2.70714865962632e-06, + "loss": 2.0977, + "step": 1419 + }, + { + "epoch": 1.5561643835616439, + "grad_norm": 1.633492112159729, + "learning_rate": 2.705117790414298e-06, + "loss": 2.1078, + "step": 1420 + }, + { + "epoch": 1.5561643835616439, + "eval_loss": 2.2733564376831055, + "eval_runtime": 37.7364, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 1420 + }, + { + "epoch": 1.5572602739726027, + "grad_norm": 1.6661697626113892, + "learning_rate": 2.7030869212022747e-06, + "loss": 2.1395, + "step": 1421 + }, + { + "epoch": 1.5583561643835617, + "grad_norm": 1.678194284439087, + "learning_rate": 2.7010560519902516e-06, + "loss": 2.101, + "step": 1422 + }, + { + "epoch": 1.5594520547945205, + "grad_norm": 1.686339020729065, + "learning_rate": 2.6990251827782293e-06, + "loss": 2.1117, + "step": 1423 + }, + { + "epoch": 1.5605479452054793, + "grad_norm": 1.5932446718215942, + "learning_rate": 2.696994313566206e-06, + "loss": 2.1075, + "step": 1424 + }, + { + "epoch": 1.5616438356164384, + "grad_norm": 1.744378685951233, + "learning_rate": 2.694963444354184e-06, + "loss": 2.1658, + "step": 1425 + }, + { + "epoch": 1.5627397260273974, + "grad_norm": 1.5379083156585693, + "learning_rate": 2.6929325751421607e-06, + "loss": 2.1027, + "step": 1426 + }, + { + "epoch": 1.563835616438356, + "grad_norm": 1.582470417022705, + "learning_rate": 2.6909017059301384e-06, + "loss": 2.1343, + "step": 1427 + }, + { + "epoch": 1.564931506849315, + "grad_norm": 1.6151307821273804, + "learning_rate": 2.6888708367181153e-06, + "loss": 2.0837, + "step": 1428 + }, + { + "epoch": 1.566027397260274, + "grad_norm": 1.7137283086776733, + "learning_rate": 2.686839967506093e-06, + "loss": 2.135, + "step": 1429 + }, + { + "epoch": 1.5671232876712329, + "grad_norm": 1.6230741739273071, + "learning_rate": 2.68480909829407e-06, + "loss": 2.0593, + "step": 1430 + }, + { + "epoch": 1.5671232876712329, + "eval_loss": 2.2720754146575928, + "eval_runtime": 37.8152, + "eval_samples_per_second": 26.444, + "eval_steps_per_second": 1.111, + "step": 1430 + }, + { + "epoch": 1.5682191780821917, + "grad_norm": 1.638656497001648, + "learning_rate": 2.6827782290820475e-06, + "loss": 2.0652, + "step": 1431 + }, + { + "epoch": 1.5693150684931507, + "grad_norm": 1.565036654472351, + "learning_rate": 2.6807473598700244e-06, + "loss": 2.154, + "step": 1432 + }, + { + "epoch": 1.5704109589041098, + "grad_norm": 1.624617576599121, + "learning_rate": 2.678716490658002e-06, + "loss": 2.1264, + "step": 1433 + }, + { + "epoch": 1.5715068493150683, + "grad_norm": 1.5961869955062866, + "learning_rate": 2.676685621445979e-06, + "loss": 2.1189, + "step": 1434 + }, + { + "epoch": 1.5726027397260274, + "grad_norm": 1.5941061973571777, + "learning_rate": 2.674654752233956e-06, + "loss": 2.1229, + "step": 1435 + }, + { + "epoch": 1.5736986301369864, + "grad_norm": 1.5900241136550903, + "learning_rate": 2.6726238830219335e-06, + "loss": 2.1174, + "step": 1436 + }, + { + "epoch": 1.5747945205479452, + "grad_norm": 1.696982979774475, + "learning_rate": 2.6705930138099107e-06, + "loss": 2.1463, + "step": 1437 + }, + { + "epoch": 1.575890410958904, + "grad_norm": 1.5432531833648682, + "learning_rate": 2.668562144597888e-06, + "loss": 2.0515, + "step": 1438 + }, + { + "epoch": 1.576986301369863, + "grad_norm": 1.6274412870407104, + "learning_rate": 2.6665312753858653e-06, + "loss": 2.1281, + "step": 1439 + }, + { + "epoch": 1.5780821917808219, + "grad_norm": 1.5978962182998657, + "learning_rate": 2.6645004061738426e-06, + "loss": 2.123, + "step": 1440 + }, + { + "epoch": 1.5780821917808219, + "eval_loss": 2.2747466564178467, + "eval_runtime": 37.7905, + "eval_samples_per_second": 26.462, + "eval_steps_per_second": 1.111, + "step": 1440 + }, + { + "epoch": 1.5791780821917807, + "grad_norm": 1.684459924697876, + "learning_rate": 2.66246953696182e-06, + "loss": 2.09, + "step": 1441 + }, + { + "epoch": 1.5802739726027397, + "grad_norm": 1.6446969509124756, + "learning_rate": 2.660438667749797e-06, + "loss": 2.0723, + "step": 1442 + }, + { + "epoch": 1.5813698630136988, + "grad_norm": 1.6567888259887695, + "learning_rate": 2.6584077985377744e-06, + "loss": 2.1052, + "step": 1443 + }, + { + "epoch": 1.5824657534246576, + "grad_norm": 1.6030794382095337, + "learning_rate": 2.6563769293257517e-06, + "loss": 2.11, + "step": 1444 + }, + { + "epoch": 1.5835616438356164, + "grad_norm": 1.5979795455932617, + "learning_rate": 2.654346060113729e-06, + "loss": 2.0903, + "step": 1445 + }, + { + "epoch": 1.5846575342465754, + "grad_norm": 1.5626153945922852, + "learning_rate": 2.6523151909017062e-06, + "loss": 2.0998, + "step": 1446 + }, + { + "epoch": 1.5857534246575342, + "grad_norm": 1.6497416496276855, + "learning_rate": 2.6502843216896835e-06, + "loss": 2.0863, + "step": 1447 + }, + { + "epoch": 1.586849315068493, + "grad_norm": 1.6650139093399048, + "learning_rate": 2.6482534524776603e-06, + "loss": 2.1615, + "step": 1448 + }, + { + "epoch": 1.587945205479452, + "grad_norm": 1.6197447776794434, + "learning_rate": 2.646222583265638e-06, + "loss": 2.1374, + "step": 1449 + }, + { + "epoch": 1.589041095890411, + "grad_norm": 1.6509665250778198, + "learning_rate": 2.644191714053615e-06, + "loss": 2.1242, + "step": 1450 + }, + { + "epoch": 1.589041095890411, + "eval_loss": 2.272801160812378, + "eval_runtime": 37.769, + "eval_samples_per_second": 26.477, + "eval_steps_per_second": 1.112, + "step": 1450 + }, + { + "epoch": 1.59013698630137, + "grad_norm": 1.570002555847168, + "learning_rate": 2.6421608448415926e-06, + "loss": 2.0899, + "step": 1451 + }, + { + "epoch": 1.5912328767123287, + "grad_norm": 1.653418779373169, + "learning_rate": 2.6401299756295694e-06, + "loss": 2.0823, + "step": 1452 + }, + { + "epoch": 1.5923287671232877, + "grad_norm": 1.7645854949951172, + "learning_rate": 2.638099106417547e-06, + "loss": 2.1162, + "step": 1453 + }, + { + "epoch": 1.5934246575342466, + "grad_norm": 1.6598302125930786, + "learning_rate": 2.636068237205524e-06, + "loss": 2.0887, + "step": 1454 + }, + { + "epoch": 1.5945205479452054, + "grad_norm": 1.7615699768066406, + "learning_rate": 2.6340373679935017e-06, + "loss": 2.153, + "step": 1455 + }, + { + "epoch": 1.5956164383561644, + "grad_norm": 1.5611566305160522, + "learning_rate": 2.6320064987814785e-06, + "loss": 2.0635, + "step": 1456 + }, + { + "epoch": 1.5967123287671234, + "grad_norm": 1.7776182889938354, + "learning_rate": 2.6299756295694562e-06, + "loss": 2.1248, + "step": 1457 + }, + { + "epoch": 1.5978082191780822, + "grad_norm": 1.563895344734192, + "learning_rate": 2.627944760357433e-06, + "loss": 2.1591, + "step": 1458 + }, + { + "epoch": 1.598904109589041, + "grad_norm": 1.5638104677200317, + "learning_rate": 2.6259138911454104e-06, + "loss": 2.1594, + "step": 1459 + }, + { + "epoch": 1.6, + "grad_norm": 1.6918673515319824, + "learning_rate": 2.6238830219333877e-06, + "loss": 2.143, + "step": 1460 + }, + { + "epoch": 1.6, + "eval_loss": 2.27238130569458, + "eval_runtime": 37.7964, + "eval_samples_per_second": 26.458, + "eval_steps_per_second": 1.111, + "step": 1460 + }, + { + "epoch": 1.601095890410959, + "grad_norm": 1.6937850713729858, + "learning_rate": 2.621852152721365e-06, + "loss": 2.0692, + "step": 1461 + }, + { + "epoch": 1.6021917808219177, + "grad_norm": 1.6500011682510376, + "learning_rate": 2.619821283509342e-06, + "loss": 2.1432, + "step": 1462 + }, + { + "epoch": 1.6032876712328767, + "grad_norm": 1.583245038986206, + "learning_rate": 2.6177904142973195e-06, + "loss": 2.0811, + "step": 1463 + }, + { + "epoch": 1.6043835616438358, + "grad_norm": 1.7283663749694824, + "learning_rate": 2.6157595450852968e-06, + "loss": 2.1268, + "step": 1464 + }, + { + "epoch": 1.6054794520547944, + "grad_norm": 1.6247549057006836, + "learning_rate": 2.613728675873274e-06, + "loss": 2.1464, + "step": 1465 + }, + { + "epoch": 1.6065753424657534, + "grad_norm": 1.6460528373718262, + "learning_rate": 2.6116978066612513e-06, + "loss": 2.1428, + "step": 1466 + }, + { + "epoch": 1.6076712328767124, + "grad_norm": 1.6332457065582275, + "learning_rate": 2.6096669374492286e-06, + "loss": 2.1216, + "step": 1467 + }, + { + "epoch": 1.6087671232876712, + "grad_norm": 1.7202653884887695, + "learning_rate": 2.607636068237206e-06, + "loss": 2.0877, + "step": 1468 + }, + { + "epoch": 1.60986301369863, + "grad_norm": 1.6361600160598755, + "learning_rate": 2.605605199025183e-06, + "loss": 2.129, + "step": 1469 + }, + { + "epoch": 1.610958904109589, + "grad_norm": 1.6851226091384888, + "learning_rate": 2.6035743298131604e-06, + "loss": 2.1181, + "step": 1470 + }, + { + "epoch": 1.610958904109589, + "eval_loss": 2.276170492172241, + "eval_runtime": 37.7741, + "eval_samples_per_second": 26.473, + "eval_steps_per_second": 1.112, + "step": 1470 + }, + { + "epoch": 1.612054794520548, + "grad_norm": 1.7388262748718262, + "learning_rate": 2.6015434606011377e-06, + "loss": 2.137, + "step": 1471 + }, + { + "epoch": 1.6131506849315067, + "grad_norm": 1.8428372144699097, + "learning_rate": 2.5995125913891145e-06, + "loss": 2.1324, + "step": 1472 + }, + { + "epoch": 1.6142465753424657, + "grad_norm": 1.6778063774108887, + "learning_rate": 2.5974817221770922e-06, + "loss": 2.1207, + "step": 1473 + }, + { + "epoch": 1.6153424657534248, + "grad_norm": 1.5915114879608154, + "learning_rate": 2.595450852965069e-06, + "loss": 2.1297, + "step": 1474 + }, + { + "epoch": 1.6164383561643836, + "grad_norm": 1.6911993026733398, + "learning_rate": 2.593419983753047e-06, + "loss": 2.1574, + "step": 1475 + }, + { + "epoch": 1.6175342465753424, + "grad_norm": 1.568377137184143, + "learning_rate": 2.5913891145410236e-06, + "loss": 2.1046, + "step": 1476 + }, + { + "epoch": 1.6186301369863014, + "grad_norm": 1.6314934492111206, + "learning_rate": 2.5893582453290013e-06, + "loss": 2.145, + "step": 1477 + }, + { + "epoch": 1.6197260273972602, + "grad_norm": 1.6405373811721802, + "learning_rate": 2.587327376116978e-06, + "loss": 2.1017, + "step": 1478 + }, + { + "epoch": 1.620821917808219, + "grad_norm": 1.5760985612869263, + "learning_rate": 2.585296506904956e-06, + "loss": 2.1106, + "step": 1479 + }, + { + "epoch": 1.621917808219178, + "grad_norm": 1.667673945426941, + "learning_rate": 2.5832656376929327e-06, + "loss": 2.0898, + "step": 1480 + }, + { + "epoch": 1.621917808219178, + "eval_loss": 2.272311210632324, + "eval_runtime": 37.6893, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 1480 + }, + { + "epoch": 1.6230136986301371, + "grad_norm": 1.6694347858428955, + "learning_rate": 2.5812347684809104e-06, + "loss": 2.0515, + "step": 1481 + }, + { + "epoch": 1.624109589041096, + "grad_norm": 1.6225323677062988, + "learning_rate": 2.5792038992688873e-06, + "loss": 2.1824, + "step": 1482 + }, + { + "epoch": 1.6252054794520547, + "grad_norm": 1.7553166151046753, + "learning_rate": 2.577173030056865e-06, + "loss": 2.0484, + "step": 1483 + }, + { + "epoch": 1.6263013698630138, + "grad_norm": 1.6127912998199463, + "learning_rate": 2.575142160844842e-06, + "loss": 2.1306, + "step": 1484 + }, + { + "epoch": 1.6273972602739726, + "grad_norm": 1.728456974029541, + "learning_rate": 2.5731112916328187e-06, + "loss": 2.1371, + "step": 1485 + }, + { + "epoch": 1.6284931506849314, + "grad_norm": 1.655109167098999, + "learning_rate": 2.5710804224207964e-06, + "loss": 2.0918, + "step": 1486 + }, + { + "epoch": 1.6295890410958904, + "grad_norm": 1.7091511487960815, + "learning_rate": 2.5690495532087733e-06, + "loss": 2.0534, + "step": 1487 + }, + { + "epoch": 1.6306849315068495, + "grad_norm": 1.663798213005066, + "learning_rate": 2.567018683996751e-06, + "loss": 2.1262, + "step": 1488 + }, + { + "epoch": 1.6317808219178083, + "grad_norm": 1.7468243837356567, + "learning_rate": 2.564987814784728e-06, + "loss": 2.088, + "step": 1489 + }, + { + "epoch": 1.632876712328767, + "grad_norm": 1.662075400352478, + "learning_rate": 2.5629569455727055e-06, + "loss": 2.1063, + "step": 1490 + }, + { + "epoch": 1.632876712328767, + "eval_loss": 2.268484354019165, + "eval_runtime": 37.698, + "eval_samples_per_second": 26.527, + "eval_steps_per_second": 1.114, + "step": 1490 + }, + { + "epoch": 1.633972602739726, + "grad_norm": 1.6704449653625488, + "learning_rate": 2.5609260763606824e-06, + "loss": 2.0996, + "step": 1491 + }, + { + "epoch": 1.635068493150685, + "grad_norm": 1.622410774230957, + "learning_rate": 2.55889520714866e-06, + "loss": 2.1792, + "step": 1492 + }, + { + "epoch": 1.6361643835616437, + "grad_norm": 1.6591641902923584, + "learning_rate": 2.556864337936637e-06, + "loss": 2.0999, + "step": 1493 + }, + { + "epoch": 1.6372602739726028, + "grad_norm": 1.7249618768692017, + "learning_rate": 2.5548334687246146e-06, + "loss": 2.1152, + "step": 1494 + }, + { + "epoch": 1.6383561643835618, + "grad_norm": 1.8214666843414307, + "learning_rate": 2.5528025995125915e-06, + "loss": 2.1512, + "step": 1495 + }, + { + "epoch": 1.6394520547945204, + "grad_norm": 1.6735785007476807, + "learning_rate": 2.550771730300569e-06, + "loss": 2.1375, + "step": 1496 + }, + { + "epoch": 1.6405479452054794, + "grad_norm": 1.8190255165100098, + "learning_rate": 2.548740861088546e-06, + "loss": 2.1425, + "step": 1497 + }, + { + "epoch": 1.6416438356164385, + "grad_norm": 1.6991790533065796, + "learning_rate": 2.5467099918765233e-06, + "loss": 2.0658, + "step": 1498 + }, + { + "epoch": 1.6427397260273973, + "grad_norm": 1.6228128671646118, + "learning_rate": 2.5446791226645006e-06, + "loss": 2.0505, + "step": 1499 + }, + { + "epoch": 1.643835616438356, + "grad_norm": 1.9889343976974487, + "learning_rate": 2.542648253452478e-06, + "loss": 2.1374, + "step": 1500 + }, + { + "epoch": 1.643835616438356, + "eval_loss": 2.2690248489379883, + "eval_runtime": 37.6989, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 1500 + }, + { + "epoch": 1.644931506849315, + "grad_norm": 1.6495356559753418, + "learning_rate": 2.540617384240455e-06, + "loss": 2.1171, + "step": 1501 + }, + { + "epoch": 1.6460273972602741, + "grad_norm": 1.7420153617858887, + "learning_rate": 2.5385865150284324e-06, + "loss": 2.092, + "step": 1502 + }, + { + "epoch": 1.6471232876712327, + "grad_norm": 1.7456740140914917, + "learning_rate": 2.5365556458164097e-06, + "loss": 2.1029, + "step": 1503 + }, + { + "epoch": 1.6482191780821918, + "grad_norm": 1.607245922088623, + "learning_rate": 2.534524776604387e-06, + "loss": 2.1173, + "step": 1504 + }, + { + "epoch": 1.6493150684931508, + "grad_norm": 1.648555040359497, + "learning_rate": 2.5324939073923642e-06, + "loss": 2.0275, + "step": 1505 + }, + { + "epoch": 1.6504109589041096, + "grad_norm": 1.753891944885254, + "learning_rate": 2.5304630381803415e-06, + "loss": 2.103, + "step": 1506 + }, + { + "epoch": 1.6515068493150684, + "grad_norm": 1.6390420198440552, + "learning_rate": 2.5284321689683188e-06, + "loss": 2.111, + "step": 1507 + }, + { + "epoch": 1.6526027397260274, + "grad_norm": 1.6427407264709473, + "learning_rate": 2.526401299756296e-06, + "loss": 2.0834, + "step": 1508 + }, + { + "epoch": 1.6536986301369863, + "grad_norm": 1.7193669080734253, + "learning_rate": 2.524370430544273e-06, + "loss": 2.1125, + "step": 1509 + }, + { + "epoch": 1.654794520547945, + "grad_norm": 1.6364102363586426, + "learning_rate": 2.5223395613322506e-06, + "loss": 2.0839, + "step": 1510 + }, + { + "epoch": 1.654794520547945, + "eval_loss": 2.271984577178955, + "eval_runtime": 37.4343, + "eval_samples_per_second": 26.713, + "eval_steps_per_second": 1.122, + "step": 1510 + }, + { + "epoch": 1.655890410958904, + "grad_norm": 1.6273912191390991, + "learning_rate": 2.5203086921202274e-06, + "loss": 2.1296, + "step": 1511 + }, + { + "epoch": 1.6569863013698631, + "grad_norm": 1.631685733795166, + "learning_rate": 2.518277822908205e-06, + "loss": 2.0956, + "step": 1512 + }, + { + "epoch": 1.658082191780822, + "grad_norm": 1.668209195137024, + "learning_rate": 2.516246953696182e-06, + "loss": 2.1216, + "step": 1513 + }, + { + "epoch": 1.6591780821917808, + "grad_norm": 1.7267968654632568, + "learning_rate": 2.5142160844841597e-06, + "loss": 2.1421, + "step": 1514 + }, + { + "epoch": 1.6602739726027398, + "grad_norm": 1.730349063873291, + "learning_rate": 2.5121852152721365e-06, + "loss": 2.1286, + "step": 1515 + }, + { + "epoch": 1.6613698630136986, + "grad_norm": 1.6914033889770508, + "learning_rate": 2.5101543460601142e-06, + "loss": 2.123, + "step": 1516 + }, + { + "epoch": 1.6624657534246574, + "grad_norm": 1.6512525081634521, + "learning_rate": 2.508123476848091e-06, + "loss": 2.0714, + "step": 1517 + }, + { + "epoch": 1.6635616438356164, + "grad_norm": 1.7400277853012085, + "learning_rate": 2.506092607636069e-06, + "loss": 2.1508, + "step": 1518 + }, + { + "epoch": 1.6646575342465755, + "grad_norm": 1.6652593612670898, + "learning_rate": 2.5040617384240457e-06, + "loss": 2.0897, + "step": 1519 + }, + { + "epoch": 1.6657534246575343, + "grad_norm": 1.8187719583511353, + "learning_rate": 2.5020308692120234e-06, + "loss": 2.0946, + "step": 1520 + }, + { + "epoch": 1.6657534246575343, + "eval_loss": 2.2724390029907227, + "eval_runtime": 37.6493, + "eval_samples_per_second": 26.561, + "eval_steps_per_second": 1.116, + "step": 1520 + }, + { + "epoch": 1.666849315068493, + "grad_norm": 1.6140000820159912, + "learning_rate": 2.5e-06, + "loss": 2.1534, + "step": 1521 + }, + { + "epoch": 1.6679452054794521, + "grad_norm": 1.7010711431503296, + "learning_rate": 2.4979691307879775e-06, + "loss": 2.1028, + "step": 1522 + }, + { + "epoch": 1.669041095890411, + "grad_norm": 1.661583423614502, + "learning_rate": 2.4959382615759548e-06, + "loss": 2.0763, + "step": 1523 + }, + { + "epoch": 1.6701369863013698, + "grad_norm": 1.6757241487503052, + "learning_rate": 2.493907392363932e-06, + "loss": 2.1602, + "step": 1524 + }, + { + "epoch": 1.6712328767123288, + "grad_norm": 1.5544378757476807, + "learning_rate": 2.4918765231519093e-06, + "loss": 2.1251, + "step": 1525 + }, + { + "epoch": 1.6723287671232878, + "grad_norm": 1.7000782489776611, + "learning_rate": 2.4898456539398866e-06, + "loss": 2.1044, + "step": 1526 + }, + { + "epoch": 1.6734246575342466, + "grad_norm": 1.584068775177002, + "learning_rate": 2.487814784727864e-06, + "loss": 2.166, + "step": 1527 + }, + { + "epoch": 1.6745205479452054, + "grad_norm": 1.6422476768493652, + "learning_rate": 2.4857839155158407e-06, + "loss": 2.1079, + "step": 1528 + }, + { + "epoch": 1.6756164383561645, + "grad_norm": 1.621401071548462, + "learning_rate": 2.483753046303818e-06, + "loss": 2.107, + "step": 1529 + }, + { + "epoch": 1.6767123287671233, + "grad_norm": 1.6382123231887817, + "learning_rate": 2.4817221770917953e-06, + "loss": 2.1161, + "step": 1530 + }, + { + "epoch": 1.6767123287671233, + "eval_loss": 2.2733490467071533, + "eval_runtime": 37.713, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 1530 + }, + { + "epoch": 1.677808219178082, + "grad_norm": 1.7815386056900024, + "learning_rate": 2.4796913078797725e-06, + "loss": 2.138, + "step": 1531 + }, + { + "epoch": 1.6789041095890411, + "grad_norm": 1.7218440771102905, + "learning_rate": 2.47766043866775e-06, + "loss": 2.1342, + "step": 1532 + }, + { + "epoch": 1.6800000000000002, + "grad_norm": 1.5806915760040283, + "learning_rate": 2.475629569455727e-06, + "loss": 2.1229, + "step": 1533 + }, + { + "epoch": 1.6810958904109587, + "grad_norm": 1.5785084962844849, + "learning_rate": 2.4735987002437044e-06, + "loss": 2.0883, + "step": 1534 + }, + { + "epoch": 1.6821917808219178, + "grad_norm": 1.587751030921936, + "learning_rate": 2.4715678310316816e-06, + "loss": 2.0956, + "step": 1535 + }, + { + "epoch": 1.6832876712328768, + "grad_norm": 1.604783296585083, + "learning_rate": 2.469536961819659e-06, + "loss": 2.1014, + "step": 1536 + }, + { + "epoch": 1.6843835616438356, + "grad_norm": 1.6625169515609741, + "learning_rate": 2.467506092607636e-06, + "loss": 2.1099, + "step": 1537 + }, + { + "epoch": 1.6854794520547944, + "grad_norm": 1.623159646987915, + "learning_rate": 2.4654752233956135e-06, + "loss": 2.1012, + "step": 1538 + }, + { + "epoch": 1.6865753424657535, + "grad_norm": 1.7528589963912964, + "learning_rate": 2.4634443541835907e-06, + "loss": 2.1635, + "step": 1539 + }, + { + "epoch": 1.6876712328767123, + "grad_norm": 1.6822079420089722, + "learning_rate": 2.461413484971568e-06, + "loss": 2.1057, + "step": 1540 + }, + { + "epoch": 1.6876712328767123, + "eval_loss": 2.269556999206543, + "eval_runtime": 37.6668, + "eval_samples_per_second": 26.549, + "eval_steps_per_second": 1.115, + "step": 1540 + }, + { + "epoch": 1.688767123287671, + "grad_norm": 1.5787808895111084, + "learning_rate": 2.4593826157595453e-06, + "loss": 2.1831, + "step": 1541 + }, + { + "epoch": 1.6898630136986301, + "grad_norm": 1.5939741134643555, + "learning_rate": 2.4573517465475226e-06, + "loss": 2.0706, + "step": 1542 + }, + { + "epoch": 1.6909589041095892, + "grad_norm": 1.796216607093811, + "learning_rate": 2.4553208773355e-06, + "loss": 2.0959, + "step": 1543 + }, + { + "epoch": 1.692054794520548, + "grad_norm": 1.6734933853149414, + "learning_rate": 2.453290008123477e-06, + "loss": 2.1098, + "step": 1544 + }, + { + "epoch": 1.6931506849315068, + "grad_norm": 1.5036938190460205, + "learning_rate": 2.4512591389114544e-06, + "loss": 2.12, + "step": 1545 + }, + { + "epoch": 1.6942465753424658, + "grad_norm": 1.631889820098877, + "learning_rate": 2.4492282696994317e-06, + "loss": 2.1207, + "step": 1546 + }, + { + "epoch": 1.6953424657534246, + "grad_norm": 1.6895856857299805, + "learning_rate": 2.447197400487409e-06, + "loss": 2.115, + "step": 1547 + }, + { + "epoch": 1.6964383561643834, + "grad_norm": 1.581718921661377, + "learning_rate": 2.4451665312753862e-06, + "loss": 2.1337, + "step": 1548 + }, + { + "epoch": 1.6975342465753425, + "grad_norm": 1.6116989850997925, + "learning_rate": 2.4431356620633635e-06, + "loss": 2.1269, + "step": 1549 + }, + { + "epoch": 1.6986301369863015, + "grad_norm": 1.5890625715255737, + "learning_rate": 2.4411047928513408e-06, + "loss": 2.1115, + "step": 1550 + }, + { + "epoch": 1.6986301369863015, + "eval_loss": 2.269603729248047, + "eval_runtime": 37.7242, + "eval_samples_per_second": 26.508, + "eval_steps_per_second": 1.113, + "step": 1550 + }, + { + "epoch": 1.6997260273972603, + "grad_norm": 1.6351535320281982, + "learning_rate": 2.439073923639318e-06, + "loss": 2.1243, + "step": 1551 + }, + { + "epoch": 1.7008219178082191, + "grad_norm": 1.628990650177002, + "learning_rate": 2.4370430544272953e-06, + "loss": 2.0874, + "step": 1552 + }, + { + "epoch": 1.7019178082191782, + "grad_norm": 1.6842089891433716, + "learning_rate": 2.435012185215272e-06, + "loss": 2.0774, + "step": 1553 + }, + { + "epoch": 1.703013698630137, + "grad_norm": 1.6843653917312622, + "learning_rate": 2.4329813160032495e-06, + "loss": 2.09, + "step": 1554 + }, + { + "epoch": 1.7041095890410958, + "grad_norm": 1.6409231424331665, + "learning_rate": 2.4309504467912267e-06, + "loss": 2.1029, + "step": 1555 + }, + { + "epoch": 1.7052054794520548, + "grad_norm": 1.6237034797668457, + "learning_rate": 2.428919577579204e-06, + "loss": 2.1014, + "step": 1556 + }, + { + "epoch": 1.7063013698630138, + "grad_norm": 1.7377818822860718, + "learning_rate": 2.4268887083671813e-06, + "loss": 2.1786, + "step": 1557 + }, + { + "epoch": 1.7073972602739727, + "grad_norm": 1.6430894136428833, + "learning_rate": 2.4248578391551586e-06, + "loss": 2.0985, + "step": 1558 + }, + { + "epoch": 1.7084931506849315, + "grad_norm": 1.5623244047164917, + "learning_rate": 2.422826969943136e-06, + "loss": 2.0913, + "step": 1559 + }, + { + "epoch": 1.7095890410958905, + "grad_norm": 1.6417165994644165, + "learning_rate": 2.420796100731113e-06, + "loss": 2.0673, + "step": 1560 + }, + { + "epoch": 1.7095890410958905, + "eval_loss": 2.268665075302124, + "eval_runtime": 37.7309, + "eval_samples_per_second": 26.504, + "eval_steps_per_second": 1.113, + "step": 1560 + }, + { + "epoch": 1.7106849315068493, + "grad_norm": 1.605925440788269, + "learning_rate": 2.4187652315190904e-06, + "loss": 2.1087, + "step": 1561 + }, + { + "epoch": 1.7117808219178081, + "grad_norm": 1.5908552408218384, + "learning_rate": 2.4167343623070677e-06, + "loss": 2.1254, + "step": 1562 + }, + { + "epoch": 1.7128767123287671, + "grad_norm": 1.6716837882995605, + "learning_rate": 2.414703493095045e-06, + "loss": 2.109, + "step": 1563 + }, + { + "epoch": 1.7139726027397262, + "grad_norm": 1.735393762588501, + "learning_rate": 2.4126726238830222e-06, + "loss": 2.1181, + "step": 1564 + }, + { + "epoch": 1.7150684931506848, + "grad_norm": 1.6103562116622925, + "learning_rate": 2.410641754670999e-06, + "loss": 2.1, + "step": 1565 + }, + { + "epoch": 1.7161643835616438, + "grad_norm": 1.6393123865127563, + "learning_rate": 2.4086108854589763e-06, + "loss": 2.1363, + "step": 1566 + }, + { + "epoch": 1.7172602739726028, + "grad_norm": 1.6986010074615479, + "learning_rate": 2.4065800162469536e-06, + "loss": 2.0909, + "step": 1567 + }, + { + "epoch": 1.7183561643835616, + "grad_norm": 1.5851256847381592, + "learning_rate": 2.404549147034931e-06, + "loss": 2.1046, + "step": 1568 + }, + { + "epoch": 1.7194520547945205, + "grad_norm": 1.7347966432571411, + "learning_rate": 2.402518277822908e-06, + "loss": 2.1517, + "step": 1569 + }, + { + "epoch": 1.7205479452054795, + "grad_norm": 1.6968820095062256, + "learning_rate": 2.4004874086108854e-06, + "loss": 2.0702, + "step": 1570 + }, + { + "epoch": 1.7205479452054795, + "eval_loss": 2.26865816116333, + "eval_runtime": 37.6876, + "eval_samples_per_second": 26.534, + "eval_steps_per_second": 1.114, + "step": 1570 + }, + { + "epoch": 1.7216438356164385, + "grad_norm": 1.5828689336776733, + "learning_rate": 2.3984565393988627e-06, + "loss": 2.1416, + "step": 1571 + }, + { + "epoch": 1.7227397260273971, + "grad_norm": 1.672504186630249, + "learning_rate": 2.39642567018684e-06, + "loss": 2.09, + "step": 1572 + }, + { + "epoch": 1.7238356164383561, + "grad_norm": 1.6303755044937134, + "learning_rate": 2.3943948009748173e-06, + "loss": 2.1042, + "step": 1573 + }, + { + "epoch": 1.7249315068493152, + "grad_norm": 1.6052359342575073, + "learning_rate": 2.3923639317627945e-06, + "loss": 2.1246, + "step": 1574 + }, + { + "epoch": 1.726027397260274, + "grad_norm": 1.5766305923461914, + "learning_rate": 2.390333062550772e-06, + "loss": 2.0823, + "step": 1575 + }, + { + "epoch": 1.7271232876712328, + "grad_norm": 1.6897045373916626, + "learning_rate": 2.388302193338749e-06, + "loss": 2.0112, + "step": 1576 + }, + { + "epoch": 1.7282191780821918, + "grad_norm": 1.732053518295288, + "learning_rate": 2.3862713241267264e-06, + "loss": 2.1817, + "step": 1577 + }, + { + "epoch": 1.7293150684931506, + "grad_norm": 1.7444037199020386, + "learning_rate": 2.3842404549147037e-06, + "loss": 2.0971, + "step": 1578 + }, + { + "epoch": 1.7304109589041095, + "grad_norm": 1.6277426481246948, + "learning_rate": 2.382209585702681e-06, + "loss": 2.0825, + "step": 1579 + }, + { + "epoch": 1.7315068493150685, + "grad_norm": 1.765645146369934, + "learning_rate": 2.380178716490658e-06, + "loss": 2.146, + "step": 1580 + }, + { + "epoch": 1.7315068493150685, + "eval_loss": 2.2666616439819336, + "eval_runtime": 37.7024, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 1580 + }, + { + "epoch": 1.7326027397260275, + "grad_norm": 1.6550240516662598, + "learning_rate": 2.3781478472786355e-06, + "loss": 2.1107, + "step": 1581 + }, + { + "epoch": 1.7336986301369863, + "grad_norm": 1.5257171392440796, + "learning_rate": 2.3761169780666128e-06, + "loss": 2.1346, + "step": 1582 + }, + { + "epoch": 1.7347945205479451, + "grad_norm": 1.6080381870269775, + "learning_rate": 2.37408610885459e-06, + "loss": 2.1402, + "step": 1583 + }, + { + "epoch": 1.7358904109589042, + "grad_norm": 1.6197988986968994, + "learning_rate": 2.3720552396425673e-06, + "loss": 2.0894, + "step": 1584 + }, + { + "epoch": 1.736986301369863, + "grad_norm": 1.5313374996185303, + "learning_rate": 2.3700243704305446e-06, + "loss": 2.1612, + "step": 1585 + }, + { + "epoch": 1.7380821917808218, + "grad_norm": 1.5613369941711426, + "learning_rate": 2.367993501218522e-06, + "loss": 2.1149, + "step": 1586 + }, + { + "epoch": 1.7391780821917808, + "grad_norm": 1.616064429283142, + "learning_rate": 2.365962632006499e-06, + "loss": 2.1186, + "step": 1587 + }, + { + "epoch": 1.7402739726027399, + "grad_norm": 1.6820755004882812, + "learning_rate": 2.3639317627944764e-06, + "loss": 2.1373, + "step": 1588 + }, + { + "epoch": 1.7413698630136987, + "grad_norm": 1.6518917083740234, + "learning_rate": 2.3619008935824537e-06, + "loss": 2.1316, + "step": 1589 + }, + { + "epoch": 1.7424657534246575, + "grad_norm": 1.6021987199783325, + "learning_rate": 2.3598700243704305e-06, + "loss": 2.1281, + "step": 1590 + }, + { + "epoch": 1.7424657534246575, + "eval_loss": 2.2650322914123535, + "eval_runtime": 37.6736, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 1590 + }, + { + "epoch": 1.7435616438356165, + "grad_norm": 1.571864366531372, + "learning_rate": 2.357839155158408e-06, + "loss": 2.1237, + "step": 1591 + }, + { + "epoch": 1.7446575342465753, + "grad_norm": 1.5786241292953491, + "learning_rate": 2.355808285946385e-06, + "loss": 2.116, + "step": 1592 + }, + { + "epoch": 1.7457534246575341, + "grad_norm": 1.8110829591751099, + "learning_rate": 2.3537774167343624e-06, + "loss": 2.1398, + "step": 1593 + }, + { + "epoch": 1.7468493150684932, + "grad_norm": 1.648918628692627, + "learning_rate": 2.3517465475223396e-06, + "loss": 2.1183, + "step": 1594 + }, + { + "epoch": 1.7479452054794522, + "grad_norm": 1.6972789764404297, + "learning_rate": 2.349715678310317e-06, + "loss": 2.133, + "step": 1595 + }, + { + "epoch": 1.749041095890411, + "grad_norm": 1.6375590562820435, + "learning_rate": 2.347684809098294e-06, + "loss": 2.0891, + "step": 1596 + }, + { + "epoch": 1.7501369863013698, + "grad_norm": 1.7189046144485474, + "learning_rate": 2.3456539398862715e-06, + "loss": 2.1456, + "step": 1597 + }, + { + "epoch": 1.7512328767123289, + "grad_norm": 1.729980230331421, + "learning_rate": 2.3436230706742487e-06, + "loss": 2.0876, + "step": 1598 + }, + { + "epoch": 1.7523287671232877, + "grad_norm": 1.8501031398773193, + "learning_rate": 2.341592201462226e-06, + "loss": 2.1071, + "step": 1599 + }, + { + "epoch": 1.7534246575342465, + "grad_norm": 1.6985840797424316, + "learning_rate": 2.3395613322502033e-06, + "loss": 2.1006, + "step": 1600 + }, + { + "epoch": 1.7534246575342465, + "eval_loss": 2.2667760848999023, + "eval_runtime": 37.7374, + "eval_samples_per_second": 26.499, + "eval_steps_per_second": 1.113, + "step": 1600 + }, + { + "epoch": 1.7545205479452055, + "grad_norm": 1.6635565757751465, + "learning_rate": 2.3375304630381806e-06, + "loss": 2.0745, + "step": 1601 + }, + { + "epoch": 1.7556164383561645, + "grad_norm": 1.69534170627594, + "learning_rate": 2.335499593826158e-06, + "loss": 2.0823, + "step": 1602 + }, + { + "epoch": 1.7567123287671231, + "grad_norm": 1.8325625658035278, + "learning_rate": 2.333468724614135e-06, + "loss": 2.1176, + "step": 1603 + }, + { + "epoch": 1.7578082191780822, + "grad_norm": 1.6108698844909668, + "learning_rate": 2.3314378554021124e-06, + "loss": 2.1426, + "step": 1604 + }, + { + "epoch": 1.7589041095890412, + "grad_norm": 1.7432985305786133, + "learning_rate": 2.3294069861900897e-06, + "loss": 2.0901, + "step": 1605 + }, + { + "epoch": 1.76, + "grad_norm": 1.5642728805541992, + "learning_rate": 2.327376116978067e-06, + "loss": 2.0718, + "step": 1606 + }, + { + "epoch": 1.7610958904109588, + "grad_norm": 1.7221434116363525, + "learning_rate": 2.3253452477660442e-06, + "loss": 2.1437, + "step": 1607 + }, + { + "epoch": 1.7621917808219179, + "grad_norm": 1.595481514930725, + "learning_rate": 2.3233143785540215e-06, + "loss": 2.0853, + "step": 1608 + }, + { + "epoch": 1.7632876712328767, + "grad_norm": 1.659735083580017, + "learning_rate": 2.3212835093419988e-06, + "loss": 2.1324, + "step": 1609 + }, + { + "epoch": 1.7643835616438355, + "grad_norm": 1.591446876525879, + "learning_rate": 2.319252640129976e-06, + "loss": 2.0842, + "step": 1610 + }, + { + "epoch": 1.7643835616438355, + "eval_loss": 2.265777111053467, + "eval_runtime": 37.4623, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 1610 + }, + { + "epoch": 1.7654794520547945, + "grad_norm": 1.6175882816314697, + "learning_rate": 2.3172217709179533e-06, + "loss": 2.1646, + "step": 1611 + }, + { + "epoch": 1.7665753424657535, + "grad_norm": 1.730900526046753, + "learning_rate": 2.3151909017059306e-06, + "loss": 2.1396, + "step": 1612 + }, + { + "epoch": 1.7676712328767124, + "grad_norm": 1.6827929019927979, + "learning_rate": 2.313160032493908e-06, + "loss": 2.0645, + "step": 1613 + }, + { + "epoch": 1.7687671232876712, + "grad_norm": 1.621931552886963, + "learning_rate": 2.311129163281885e-06, + "loss": 2.1179, + "step": 1614 + }, + { + "epoch": 1.7698630136986302, + "grad_norm": 1.7136013507843018, + "learning_rate": 2.309098294069862e-06, + "loss": 2.0894, + "step": 1615 + }, + { + "epoch": 1.770958904109589, + "grad_norm": 1.5491700172424316, + "learning_rate": 2.3070674248578393e-06, + "loss": 2.0717, + "step": 1616 + }, + { + "epoch": 1.7720547945205478, + "grad_norm": 1.702519416809082, + "learning_rate": 2.3050365556458166e-06, + "loss": 2.1221, + "step": 1617 + }, + { + "epoch": 1.7731506849315068, + "grad_norm": 1.6896106004714966, + "learning_rate": 2.303005686433794e-06, + "loss": 2.1273, + "step": 1618 + }, + { + "epoch": 1.7742465753424659, + "grad_norm": 1.6859010457992554, + "learning_rate": 2.300974817221771e-06, + "loss": 2.0903, + "step": 1619 + }, + { + "epoch": 1.7753424657534247, + "grad_norm": 1.5701655149459839, + "learning_rate": 2.2989439480097484e-06, + "loss": 2.0701, + "step": 1620 + }, + { + "epoch": 1.7753424657534247, + "eval_loss": 2.266063690185547, + "eval_runtime": 37.7002, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 1620 + }, + { + "epoch": 1.7764383561643835, + "grad_norm": 1.7901954650878906, + "learning_rate": 2.2969130787977257e-06, + "loss": 2.1066, + "step": 1621 + }, + { + "epoch": 1.7775342465753425, + "grad_norm": 1.6261320114135742, + "learning_rate": 2.294882209585703e-06, + "loss": 2.1322, + "step": 1622 + }, + { + "epoch": 1.7786301369863013, + "grad_norm": 1.5909242630004883, + "learning_rate": 2.2928513403736802e-06, + "loss": 2.0891, + "step": 1623 + }, + { + "epoch": 1.7797260273972602, + "grad_norm": 1.597174048423767, + "learning_rate": 2.2908204711616575e-06, + "loss": 2.1065, + "step": 1624 + }, + { + "epoch": 1.7808219178082192, + "grad_norm": 1.6266214847564697, + "learning_rate": 2.2887896019496348e-06, + "loss": 2.1491, + "step": 1625 + }, + { + "epoch": 1.7819178082191782, + "grad_norm": 1.6197980642318726, + "learning_rate": 2.286758732737612e-06, + "loss": 2.1004, + "step": 1626 + }, + { + "epoch": 1.783013698630137, + "grad_norm": 1.6115243434906006, + "learning_rate": 2.284727863525589e-06, + "loss": 2.1053, + "step": 1627 + }, + { + "epoch": 1.7841095890410958, + "grad_norm": 1.7729809284210205, + "learning_rate": 2.282696994313566e-06, + "loss": 2.1433, + "step": 1628 + }, + { + "epoch": 1.7852054794520549, + "grad_norm": 1.6845381259918213, + "learning_rate": 2.2806661251015434e-06, + "loss": 2.1294, + "step": 1629 + }, + { + "epoch": 1.7863013698630137, + "grad_norm": 1.6369025707244873, + "learning_rate": 2.2786352558895207e-06, + "loss": 2.101, + "step": 1630 + }, + { + "epoch": 1.7863013698630137, + "eval_loss": 2.2655839920043945, + "eval_runtime": 37.7129, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 1630 + }, + { + "epoch": 1.7873972602739725, + "grad_norm": 1.6781350374221802, + "learning_rate": 2.276604386677498e-06, + "loss": 2.0706, + "step": 1631 + }, + { + "epoch": 1.7884931506849315, + "grad_norm": 1.8017534017562866, + "learning_rate": 2.2745735174654753e-06, + "loss": 2.0901, + "step": 1632 + }, + { + "epoch": 1.7895890410958906, + "grad_norm": 1.6327341794967651, + "learning_rate": 2.2725426482534526e-06, + "loss": 2.1282, + "step": 1633 + }, + { + "epoch": 1.7906849315068492, + "grad_norm": 1.6167854070663452, + "learning_rate": 2.27051177904143e-06, + "loss": 2.1161, + "step": 1634 + }, + { + "epoch": 1.7917808219178082, + "grad_norm": 1.834930658340454, + "learning_rate": 2.268480909829407e-06, + "loss": 2.103, + "step": 1635 + }, + { + "epoch": 1.7928767123287672, + "grad_norm": 1.6765397787094116, + "learning_rate": 2.2664500406173844e-06, + "loss": 2.1187, + "step": 1636 + }, + { + "epoch": 1.793972602739726, + "grad_norm": 1.5943162441253662, + "learning_rate": 2.2644191714053617e-06, + "loss": 2.1316, + "step": 1637 + }, + { + "epoch": 1.7950684931506848, + "grad_norm": 1.5250434875488281, + "learning_rate": 2.262388302193339e-06, + "loss": 2.1093, + "step": 1638 + }, + { + "epoch": 1.7961643835616439, + "grad_norm": 1.6569294929504395, + "learning_rate": 2.260357432981316e-06, + "loss": 2.1105, + "step": 1639 + }, + { + "epoch": 1.7972602739726027, + "grad_norm": 1.671498417854309, + "learning_rate": 2.2583265637692935e-06, + "loss": 2.1226, + "step": 1640 + }, + { + "epoch": 1.7972602739726027, + "eval_loss": 2.2658486366271973, + "eval_runtime": 37.7223, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 1640 + }, + { + "epoch": 1.7983561643835615, + "grad_norm": 1.7031270265579224, + "learning_rate": 2.2562956945572708e-06, + "loss": 2.1118, + "step": 1641 + }, + { + "epoch": 1.7994520547945205, + "grad_norm": 1.7372241020202637, + "learning_rate": 2.254264825345248e-06, + "loss": 2.0964, + "step": 1642 + }, + { + "epoch": 1.8005479452054796, + "grad_norm": 1.7946152687072754, + "learning_rate": 2.2522339561332253e-06, + "loss": 2.109, + "step": 1643 + }, + { + "epoch": 1.8016438356164384, + "grad_norm": 1.658363938331604, + "learning_rate": 2.2502030869212026e-06, + "loss": 2.107, + "step": 1644 + }, + { + "epoch": 1.8027397260273972, + "grad_norm": 1.742719054222107, + "learning_rate": 2.24817221770918e-06, + "loss": 2.15, + "step": 1645 + }, + { + "epoch": 1.8038356164383562, + "grad_norm": 1.7087559700012207, + "learning_rate": 2.246141348497157e-06, + "loss": 2.1304, + "step": 1646 + }, + { + "epoch": 1.804931506849315, + "grad_norm": 1.6908596754074097, + "learning_rate": 2.2441104792851344e-06, + "loss": 2.0938, + "step": 1647 + }, + { + "epoch": 1.8060273972602738, + "grad_norm": 1.6445401906967163, + "learning_rate": 2.2420796100731117e-06, + "loss": 2.126, + "step": 1648 + }, + { + "epoch": 1.8071232876712329, + "grad_norm": 1.9059737920761108, + "learning_rate": 2.240048740861089e-06, + "loss": 2.1372, + "step": 1649 + }, + { + "epoch": 1.808219178082192, + "grad_norm": 1.5548077821731567, + "learning_rate": 2.2380178716490662e-06, + "loss": 2.0276, + "step": 1650 + }, + { + "epoch": 1.808219178082192, + "eval_loss": 2.2644405364990234, + "eval_runtime": 37.7453, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 1650 + }, + { + "epoch": 1.8093150684931507, + "grad_norm": 1.6281542778015137, + "learning_rate": 2.2359870024370435e-06, + "loss": 2.1285, + "step": 1651 + }, + { + "epoch": 1.8104109589041095, + "grad_norm": 1.705527901649475, + "learning_rate": 2.2339561332250204e-06, + "loss": 2.1349, + "step": 1652 + }, + { + "epoch": 1.8115068493150686, + "grad_norm": 1.7579933404922485, + "learning_rate": 2.2319252640129976e-06, + "loss": 2.1183, + "step": 1653 + }, + { + "epoch": 1.8126027397260274, + "grad_norm": 1.791908621788025, + "learning_rate": 2.229894394800975e-06, + "loss": 2.0732, + "step": 1654 + }, + { + "epoch": 1.8136986301369862, + "grad_norm": 1.6568973064422607, + "learning_rate": 2.227863525588952e-06, + "loss": 2.0453, + "step": 1655 + }, + { + "epoch": 1.8147945205479452, + "grad_norm": 1.7207781076431274, + "learning_rate": 2.2258326563769295e-06, + "loss": 2.0717, + "step": 1656 + }, + { + "epoch": 1.8158904109589042, + "grad_norm": 1.7678815126419067, + "learning_rate": 2.2238017871649067e-06, + "loss": 2.0606, + "step": 1657 + }, + { + "epoch": 1.816986301369863, + "grad_norm": 1.7331746816635132, + "learning_rate": 2.221770917952884e-06, + "loss": 2.1165, + "step": 1658 + }, + { + "epoch": 1.8180821917808219, + "grad_norm": 1.7091835737228394, + "learning_rate": 2.2197400487408613e-06, + "loss": 2.1383, + "step": 1659 + }, + { + "epoch": 1.819178082191781, + "grad_norm": 1.733068823814392, + "learning_rate": 2.2177091795288386e-06, + "loss": 2.0699, + "step": 1660 + }, + { + "epoch": 1.819178082191781, + "eval_loss": 2.2659995555877686, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 1660 + }, + { + "epoch": 1.8202739726027397, + "grad_norm": 1.7482062578201294, + "learning_rate": 2.215678310316816e-06, + "loss": 2.1038, + "step": 1661 + }, + { + "epoch": 1.8213698630136985, + "grad_norm": 1.769737958908081, + "learning_rate": 2.213647441104793e-06, + "loss": 2.1132, + "step": 1662 + }, + { + "epoch": 1.8224657534246576, + "grad_norm": 1.699000358581543, + "learning_rate": 2.2116165718927704e-06, + "loss": 2.0996, + "step": 1663 + }, + { + "epoch": 1.8235616438356166, + "grad_norm": 1.5369913578033447, + "learning_rate": 2.2095857026807473e-06, + "loss": 2.1191, + "step": 1664 + }, + { + "epoch": 1.8246575342465754, + "grad_norm": 1.798531174659729, + "learning_rate": 2.2075548334687245e-06, + "loss": 2.074, + "step": 1665 + }, + { + "epoch": 1.8257534246575342, + "grad_norm": 1.5534207820892334, + "learning_rate": 2.205523964256702e-06, + "loss": 2.0802, + "step": 1666 + }, + { + "epoch": 1.8268493150684932, + "grad_norm": 1.6919447183609009, + "learning_rate": 2.203493095044679e-06, + "loss": 2.114, + "step": 1667 + }, + { + "epoch": 1.827945205479452, + "grad_norm": 1.7718945741653442, + "learning_rate": 2.2014622258326564e-06, + "loss": 2.1328, + "step": 1668 + }, + { + "epoch": 1.8290410958904109, + "grad_norm": 1.5809446573257446, + "learning_rate": 2.1994313566206336e-06, + "loss": 2.1472, + "step": 1669 + }, + { + "epoch": 1.83013698630137, + "grad_norm": 1.671584963798523, + "learning_rate": 2.197400487408611e-06, + "loss": 2.0894, + "step": 1670 + }, + { + "epoch": 1.83013698630137, + "eval_loss": 2.2637691497802734, + "eval_runtime": 37.7573, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1670 + }, + { + "epoch": 1.831232876712329, + "grad_norm": 1.7733304500579834, + "learning_rate": 2.195369618196588e-06, + "loss": 2.0732, + "step": 1671 + }, + { + "epoch": 1.8323287671232875, + "grad_norm": 1.680348515510559, + "learning_rate": 2.1933387489845655e-06, + "loss": 2.1032, + "step": 1672 + }, + { + "epoch": 1.8334246575342465, + "grad_norm": 1.734384298324585, + "learning_rate": 2.1913078797725427e-06, + "loss": 2.123, + "step": 1673 + }, + { + "epoch": 1.8345205479452056, + "grad_norm": 1.6625633239746094, + "learning_rate": 2.18927701056052e-06, + "loss": 2.1599, + "step": 1674 + }, + { + "epoch": 1.8356164383561644, + "grad_norm": 2.6964328289031982, + "learning_rate": 2.1872461413484973e-06, + "loss": 2.0988, + "step": 1675 + }, + { + "epoch": 1.8367123287671232, + "grad_norm": 1.7807667255401611, + "learning_rate": 2.1852152721364746e-06, + "loss": 2.0959, + "step": 1676 + }, + { + "epoch": 1.8378082191780822, + "grad_norm": 1.6420763731002808, + "learning_rate": 2.183184402924452e-06, + "loss": 2.072, + "step": 1677 + }, + { + "epoch": 1.838904109589041, + "grad_norm": 1.6583257913589478, + "learning_rate": 2.181153533712429e-06, + "loss": 2.1293, + "step": 1678 + }, + { + "epoch": 1.8399999999999999, + "grad_norm": 1.7795226573944092, + "learning_rate": 2.1791226645004064e-06, + "loss": 2.1025, + "step": 1679 + }, + { + "epoch": 1.841095890410959, + "grad_norm": 1.5687706470489502, + "learning_rate": 2.1770917952883837e-06, + "loss": 2.094, + "step": 1680 + }, + { + "epoch": 1.841095890410959, + "eval_loss": 2.265833854675293, + "eval_runtime": 37.757, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 1680 + }, + { + "epoch": 1.842191780821918, + "grad_norm": 1.5199975967407227, + "learning_rate": 2.175060926076361e-06, + "loss": 2.1515, + "step": 1681 + }, + { + "epoch": 1.8432876712328767, + "grad_norm": 1.6898490190505981, + "learning_rate": 2.1730300568643382e-06, + "loss": 2.1615, + "step": 1682 + }, + { + "epoch": 1.8443835616438355, + "grad_norm": 1.7239112854003906, + "learning_rate": 2.1709991876523155e-06, + "loss": 2.0923, + "step": 1683 + }, + { + "epoch": 1.8454794520547946, + "grad_norm": 1.8588145971298218, + "learning_rate": 2.1689683184402928e-06, + "loss": 2.0615, + "step": 1684 + }, + { + "epoch": 1.8465753424657534, + "grad_norm": 1.7058383226394653, + "learning_rate": 2.16693744922827e-06, + "loss": 2.1324, + "step": 1685 + }, + { + "epoch": 1.8476712328767122, + "grad_norm": 1.5865023136138916, + "learning_rate": 2.1649065800162473e-06, + "loss": 2.0982, + "step": 1686 + }, + { + "epoch": 1.8487671232876712, + "grad_norm": 1.8552294969558716, + "learning_rate": 2.1628757108042246e-06, + "loss": 2.0894, + "step": 1687 + }, + { + "epoch": 1.8498630136986303, + "grad_norm": 1.6848362684249878, + "learning_rate": 2.160844841592202e-06, + "loss": 2.0789, + "step": 1688 + }, + { + "epoch": 1.850958904109589, + "grad_norm": 1.7172846794128418, + "learning_rate": 2.1588139723801787e-06, + "loss": 2.0784, + "step": 1689 + }, + { + "epoch": 1.8520547945205479, + "grad_norm": 1.598373532295227, + "learning_rate": 2.156783103168156e-06, + "loss": 2.089, + "step": 1690 + }, + { + "epoch": 1.8520547945205479, + "eval_loss": 2.265944719314575, + "eval_runtime": 37.7878, + "eval_samples_per_second": 26.464, + "eval_steps_per_second": 1.111, + "step": 1690 + }, + { + "epoch": 1.853150684931507, + "grad_norm": 1.7011735439300537, + "learning_rate": 2.1547522339561333e-06, + "loss": 2.081, + "step": 1691 + }, + { + "epoch": 1.8542465753424657, + "grad_norm": 1.5379940271377563, + "learning_rate": 2.1527213647441106e-06, + "loss": 2.1199, + "step": 1692 + }, + { + "epoch": 1.8553424657534245, + "grad_norm": 1.6489131450653076, + "learning_rate": 2.150690495532088e-06, + "loss": 2.1497, + "step": 1693 + }, + { + "epoch": 1.8564383561643836, + "grad_norm": 1.6760814189910889, + "learning_rate": 2.148659626320065e-06, + "loss": 2.1207, + "step": 1694 + }, + { + "epoch": 1.8575342465753426, + "grad_norm": 1.668662190437317, + "learning_rate": 2.1466287571080424e-06, + "loss": 2.124, + "step": 1695 + }, + { + "epoch": 1.8586301369863014, + "grad_norm": 1.6035821437835693, + "learning_rate": 2.1445978878960197e-06, + "loss": 2.0949, + "step": 1696 + }, + { + "epoch": 1.8597260273972602, + "grad_norm": 1.8574438095092773, + "learning_rate": 2.142567018683997e-06, + "loss": 2.098, + "step": 1697 + }, + { + "epoch": 1.8608219178082193, + "grad_norm": 1.642325520515442, + "learning_rate": 2.140536149471974e-06, + "loss": 2.1454, + "step": 1698 + }, + { + "epoch": 1.861917808219178, + "grad_norm": 1.869927167892456, + "learning_rate": 2.1385052802599515e-06, + "loss": 2.0833, + "step": 1699 + }, + { + "epoch": 1.8630136986301369, + "grad_norm": 1.815918207168579, + "learning_rate": 2.1364744110479288e-06, + "loss": 2.0935, + "step": 1700 + }, + { + "epoch": 1.8630136986301369, + "eval_loss": 2.2657382488250732, + "eval_runtime": 37.7421, + "eval_samples_per_second": 26.496, + "eval_steps_per_second": 1.113, + "step": 1700 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 5.299425995915264e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/latest new file mode 100644 index 0000000000000000000000000000000000000000..12cae1adf3af8546b4141c6f62261c8e99839a54 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/latest @@ -0,0 +1 @@ +global_step600 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..ac018ae7a11305a20ba31cc797a2e0ec72e61056 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/trainer_state.json @@ -0,0 +1,4713 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.6575342465753424, + "eval_steps": 10, + "global_step": 600, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9794018268585205, + "learning_rate": 1.551094890510949e-06, + "loss": 2.3047, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.396454334259033, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.3671, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 2.0826210975646973, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.4105, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.3590970039367676, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.3405, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 2.3945844173431396, + "learning_rate": 1.624087591240876e-06, + "loss": 2.2922, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 2.185126781463623, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.3399, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.2687461376190186, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.3547, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.920499324798584, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.28, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.127774715423584, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.3409, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 1.8677122592926025, + "learning_rate": 1.715328467153285e-06, + "loss": 2.3406, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.4506173133850098, + "eval_runtime": 37.5008, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9803308248519897, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.3642, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 1.7733794450759888, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.3666, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 1.7639747858047485, + "learning_rate": 1.77007299270073e-06, + "loss": 2.3926, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 1.9172815084457397, + "learning_rate": 1.788321167883212e-06, + "loss": 2.3431, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 1.580621361732483, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.2967, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.0889124870300293, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.331, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.6659903526306152, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.3181, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.028651237487793, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.3237, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.681212306022644, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.32, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 1.9986696243286133, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.3425, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.4408469200134277, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 1.832582950592041, + "learning_rate": 1.916058394160584e-06, + "loss": 2.2548, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 1.8609144687652588, + "learning_rate": 1.934306569343066e-06, + "loss": 2.3024, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 1.8699817657470703, + "learning_rate": 1.952554744525548e-06, + "loss": 2.3044, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 1.81354558467865, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.3098, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 1.7789452075958252, + "learning_rate": 1.989051094890511e-06, + "loss": 2.3623, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 1.6322330236434937, + "learning_rate": 2.007299270072993e-06, + "loss": 2.303, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 1.8742916584014893, + "learning_rate": 2.025547445255475e-06, + "loss": 2.3183, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.7790557146072388, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.3168, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 1.7624462842941284, + "learning_rate": 2.062043795620438e-06, + "loss": 2.3629, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.781114101409912, + "learning_rate": 2.08029197080292e-06, + "loss": 2.2813, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.426335334777832, + "eval_runtime": 37.7146, + "eval_samples_per_second": 26.515, + "eval_steps_per_second": 1.114, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.769988775253296, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.3396, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.0640275478363037, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.3688, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.7884094715118408, + "learning_rate": 2.135036496350365e-06, + "loss": 2.2775, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.831847071647644, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.2966, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9044525623321533, + "learning_rate": 2.171532846715329e-06, + "loss": 2.3359, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.9899096488952637, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.3099, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8570183515548706, + "learning_rate": 2.208029197080292e-06, + "loss": 2.3016, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 1.8255733251571655, + "learning_rate": 2.226277372262774e-06, + "loss": 2.335, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.8759729862213135, + "learning_rate": 2.244525547445256e-06, + "loss": 2.3357, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.9031389951705933, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.3014, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.4213266372680664, + "eval_runtime": 37.6757, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.1044118404388428, + "learning_rate": 2.281021897810219e-06, + "loss": 2.2946, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.005735158920288, + "learning_rate": 2.299270072992701e-06, + "loss": 2.3132, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.8821312189102173, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.2816, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7866829633712769, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.309, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.8452801704406738, + "learning_rate": 2.354014598540146e-06, + "loss": 2.3222, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.5781638622283936, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.2688, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 1.9746729135513306, + "learning_rate": 2.39051094890511e-06, + "loss": 2.2864, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 2.0312447547912598, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.2914, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.675674557685852, + "learning_rate": 2.427007299270073e-06, + "loss": 2.3441, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 1.8987798690795898, + "learning_rate": 2.445255474452555e-06, + "loss": 2.2589, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.4149818420410156, + "eval_runtime": 37.6627, + "eval_samples_per_second": 26.551, + "eval_steps_per_second": 1.115, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 2.041752338409424, + "learning_rate": 2.463503649635037e-06, + "loss": 2.2457, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.7845569849014282, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.3376, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 2.489330530166626, + "learning_rate": 2.5e-06, + "loss": 2.258, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 2.3793296813964844, + "learning_rate": 2.518248175182482e-06, + "loss": 2.2561, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 2.0474460124969482, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.2992, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.7181849479675293, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.3003, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 2.3854947090148926, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.2866, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.7228457927703857, + "learning_rate": 2.591240875912409e-06, + "loss": 2.3027, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.3928253650665283, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.3233, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.136934757232666, + "learning_rate": 2.627737226277373e-06, + "loss": 2.2949, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.410449266433716, + "eval_runtime": 37.6837, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.8906159400939941, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.2693, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 1.8668863773345947, + "learning_rate": 2.664233576642336e-06, + "loss": 2.2944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.319363594055176, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.2791, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0869688987731934, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.32, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 1.8910208940505981, + "learning_rate": 2.718978102189781e-06, + "loss": 2.3059, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.157984495162964, + "learning_rate": 2.737226277372263e-06, + "loss": 2.2756, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.387314558029175, + "learning_rate": 2.755474452554745e-06, + "loss": 2.3597, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.183333158493042, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.322, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.158015489578247, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.2663, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 1.9594391584396362, + "learning_rate": 2.81021897810219e-06, + "loss": 2.3, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.411637306213379, + "eval_runtime": 37.6428, + "eval_samples_per_second": 26.565, + "eval_steps_per_second": 1.116, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.28957462310791, + "learning_rate": 2.828467153284672e-06, + "loss": 2.2258, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.8085683584213257, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.2739, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 1.9662787914276123, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.2398, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 2.337609052658081, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.3216, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 1.6008479595184326, + "learning_rate": 2.901459854014599e-06, + "loss": 2.2997, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 2.177690029144287, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.2893, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8746743202209473, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.3303, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 1.7862998247146606, + "learning_rate": 2.956204379562044e-06, + "loss": 2.3007, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.7578285932540894, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.2591, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.1532702445983887, + "learning_rate": 2.992700729927008e-06, + "loss": 2.2544, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.3992533683776855, + "eval_runtime": 37.6571, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.0168659687042236, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.3108, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 1.8336135149002075, + "learning_rate": 3.029197080291971e-06, + "loss": 2.3044, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.6675384044647217, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.2978, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 1.8866558074951172, + "learning_rate": 3.065693430656935e-06, + "loss": 2.3169, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0588760375976562, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.3041, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.9300280809402466, + "learning_rate": 3.102189781021898e-06, + "loss": 2.2494, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 1.837496042251587, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.2339, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.155080795288086, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.324, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 1.9562510251998901, + "learning_rate": 3.156934306569343e-06, + "loss": 2.2937, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 2.522388219833374, + "learning_rate": 3.175182481751825e-06, + "loss": 2.2683, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.3936216831207275, + "eval_runtime": 37.6499, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 1.7409396171569824, + "learning_rate": 3.193430656934307e-06, + "loss": 2.2833, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.176687002182007, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.2748, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 1.8576542139053345, + "learning_rate": 3.2299270072992707e-06, + "loss": 2.2144, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.9564870595932007, + "learning_rate": 3.248175182481752e-06, + "loss": 2.2898, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.896238088607788, + "learning_rate": 3.266423357664234e-06, + "loss": 2.2793, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.735960841178894, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.2869, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.018303394317627, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.2775, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 1.7448145151138306, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.2736, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8245689868927002, + "learning_rate": 3.339416058394161e-06, + "loss": 2.2799, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 1.821118950843811, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.2823, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.3893656730651855, + "eval_runtime": 37.676, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 1.9112446308135986, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.287, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 1.6978867053985596, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.2865, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9376300573349, + "learning_rate": 3.412408759124088e-06, + "loss": 2.2326, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.8067563772201538, + "learning_rate": 3.43065693430657e-06, + "loss": 2.2739, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.8476370573043823, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.2328, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.927923321723938, + "learning_rate": 3.467153284671533e-06, + "loss": 2.2679, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7956453561782837, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.2607, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 1.857413649559021, + "learning_rate": 3.503649635036497e-06, + "loss": 2.2854, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 1.8384093046188354, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.3021, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.7346371412277222, + "learning_rate": 3.54014598540146e-06, + "loss": 2.3402, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.390045642852783, + "eval_runtime": 37.4358, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.7546677589416504, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.3214, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 1.7437881231307983, + "learning_rate": 3.576642335766424e-06, + "loss": 2.2004, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9509227275848389, + "learning_rate": 3.5948905109489054e-06, + "loss": 2.2044, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 1.6918139457702637, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.2369, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9664885997772217, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.254, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 1.7942848205566406, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.2296, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 1.8610197305679321, + "learning_rate": 3.667883211678833e-06, + "loss": 2.2273, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 1.8405956029891968, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.3034, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 1.7151877880096436, + "learning_rate": 3.704379562043796e-06, + "loss": 2.2689, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 1.8795795440673828, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.2363, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.384481430053711, + "eval_runtime": 37.6512, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 1.7494690418243408, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.257, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 1.7700639963150024, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.2457, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 1.846092700958252, + "learning_rate": 3.777372262773723e-06, + "loss": 2.2425, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 1.6699751615524292, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.2734, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 1.8978508710861206, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.2435, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.6387466192245483, + "learning_rate": 3.832116788321168e-06, + "loss": 2.2241, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 1.6999541521072388, + "learning_rate": 3.85036496350365e-06, + "loss": 2.2468, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 1.6322710514068604, + "learning_rate": 3.868613138686132e-06, + "loss": 2.2642, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.5633572340011597, + "learning_rate": 3.886861313868613e-06, + "loss": 2.2862, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 1.7642648220062256, + "learning_rate": 3.905109489051096e-06, + "loss": 2.2414, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.3775289058685303, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.76724374294281, + "learning_rate": 3.923357664233577e-06, + "loss": 2.2976, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.6049190759658813, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.3034, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.6591981649398804, + "learning_rate": 3.95985401459854e-06, + "loss": 2.1957, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.650676965713501, + "learning_rate": 3.978102189781022e-06, + "loss": 2.275, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.6665960550308228, + "learning_rate": 3.996350364963504e-06, + "loss": 2.1853, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.5785901546478271, + "learning_rate": 4.014598540145986e-06, + "loss": 2.2399, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.976162075996399, + "learning_rate": 4.032846715328467e-06, + "loss": 2.2499, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.642327904701233, + "learning_rate": 4.05109489051095e-06, + "loss": 2.235, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 2.0126261711120605, + "learning_rate": 4.069343065693431e-06, + "loss": 2.2466, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.6737161874771118, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.2984, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.375901222229004, + "eval_runtime": 37.7072, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 2.094257354736328, + "learning_rate": 4.105839416058395e-06, + "loss": 2.251, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.655393123626709, + "learning_rate": 4.124087591240876e-06, + "loss": 2.2366, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 2.5909082889556885, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.2986, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.8278532028198242, + "learning_rate": 4.16058394160584e-06, + "loss": 2.1977, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9404895305633545, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.2578, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.6525582075119019, + "learning_rate": 4.197080291970803e-06, + "loss": 2.2263, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 2.021822929382324, + "learning_rate": 4.215328467153285e-06, + "loss": 2.2326, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 1.9136033058166504, + "learning_rate": 4.233576642335767e-06, + "loss": 2.2457, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.389714002609253, + "learning_rate": 4.251824817518249e-06, + "loss": 2.2256, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.7886695861816406, + "learning_rate": 4.27007299270073e-06, + "loss": 2.2222, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.3729844093322754, + "eval_runtime": 37.7638, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.9499074220657349, + "learning_rate": 4.288321167883212e-06, + "loss": 2.2026, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 1.834797978401184, + "learning_rate": 4.306569343065693e-06, + "loss": 2.2564, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 2.0415048599243164, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.2462, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 1.6527886390686035, + "learning_rate": 4.343065693430658e-06, + "loss": 2.2493, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.0504889488220215, + "learning_rate": 4.361313868613139e-06, + "loss": 2.2734, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.004772424697876, + "learning_rate": 4.379562043795621e-06, + "loss": 2.2469, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.1973347663879395, + "learning_rate": 4.397810218978102e-06, + "loss": 2.2479, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.8222625255584717, + "learning_rate": 4.416058394160584e-06, + "loss": 2.2206, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 2.1759912967681885, + "learning_rate": 4.434306569343066e-06, + "loss": 2.261, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.8344624042510986, + "learning_rate": 4.452554744525548e-06, + "loss": 2.2605, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.3706893920898438, + "eval_runtime": 37.734, + "eval_samples_per_second": 26.501, + "eval_steps_per_second": 1.113, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.7783704996109009, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.284, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 2.101193428039551, + "learning_rate": 4.489051094890512e-06, + "loss": 2.2524, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 1.7923569679260254, + "learning_rate": 4.507299270072993e-06, + "loss": 2.2434, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 2.170883893966675, + "learning_rate": 4.525547445255475e-06, + "loss": 2.2173, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.6749192476272583, + "learning_rate": 4.543795620437956e-06, + "loss": 2.283, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.0484230518341064, + "learning_rate": 4.562043795620438e-06, + "loss": 2.2931, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.9212749004364014, + "learning_rate": 4.580291970802921e-06, + "loss": 2.2661, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.6729028224945068, + "learning_rate": 4.598540145985402e-06, + "loss": 2.2269, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 2.1190619468688965, + "learning_rate": 4.6167883211678835e-06, + "loss": 2.2278, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.7444697618484497, + "learning_rate": 4.635036496350365e-06, + "loss": 2.2634, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.3706092834472656, + "eval_runtime": 37.7391, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 1.8921526670455933, + "learning_rate": 4.653284671532847e-06, + "loss": 2.2697, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7347339391708374, + "learning_rate": 4.671532846715329e-06, + "loss": 2.2406, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 2.1617953777313232, + "learning_rate": 4.689781021897811e-06, + "loss": 2.2891, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.873384714126587, + "learning_rate": 4.708029197080292e-06, + "loss": 2.2554, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.801356554031372, + "learning_rate": 4.726277372262774e-06, + "loss": 2.2073, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.9250633716583252, + "learning_rate": 4.744525547445255e-06, + "loss": 2.2501, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 1.8360860347747803, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.2561, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 1.8711498975753784, + "learning_rate": 4.78102189781022e-06, + "loss": 2.2247, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.6604810953140259, + "learning_rate": 4.799270072992701e-06, + "loss": 2.2448, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 1.689201831817627, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.2494, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.3682427406311035, + "eval_runtime": 37.7233, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.720961093902588, + "learning_rate": 4.835766423357665e-06, + "loss": 2.2013, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.7863153219223022, + "learning_rate": 4.854014598540146e-06, + "loss": 2.2183, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 1.8583060503005981, + "learning_rate": 4.872262773722628e-06, + "loss": 2.2013, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 1.722428560256958, + "learning_rate": 4.89051094890511e-06, + "loss": 2.2399, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.8408993482589722, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.2489, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.887604832649231, + "learning_rate": 4.927007299270074e-06, + "loss": 2.2145, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 1.7112189531326294, + "learning_rate": 4.945255474452555e-06, + "loss": 2.2915, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 1.8314064741134644, + "learning_rate": 4.963503649635037e-06, + "loss": 2.2786, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.7001186609268188, + "learning_rate": 4.981751824817518e-06, + "loss": 2.249, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.5820505619049072, + "learning_rate": 5e-06, + "loss": 2.254, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.3624653816223145, + "eval_runtime": 37.7283, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 1.8516303300857544, + "learning_rate": 4.997969130787978e-06, + "loss": 2.1894, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 1.8591808080673218, + "learning_rate": 4.995938261575955e-06, + "loss": 2.2537, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.6822415590286255, + "learning_rate": 4.993907392363932e-06, + "loss": 2.1978, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 2.0779759883880615, + "learning_rate": 4.9918765231519095e-06, + "loss": 2.2286, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.6937508583068848, + "learning_rate": 4.989845653939887e-06, + "loss": 2.2339, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.9342437982559204, + "learning_rate": 4.987814784727864e-06, + "loss": 2.2145, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.773258924484253, + "learning_rate": 4.985783915515841e-06, + "loss": 2.2368, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.8196488618850708, + "learning_rate": 4.983753046303819e-06, + "loss": 2.2224, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.6786775588989258, + "learning_rate": 4.981722177091796e-06, + "loss": 2.2439, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.7761427164077759, + "learning_rate": 4.979691307879773e-06, + "loss": 2.236, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.3614282608032227, + "eval_runtime": 37.7314, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.8389159440994263, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.2745, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.7774744033813477, + "learning_rate": 4.975629569455728e-06, + "loss": 2.2122, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.7862017154693604, + "learning_rate": 4.973598700243705e-06, + "loss": 2.2097, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.8851722478866577, + "learning_rate": 4.971567831031681e-06, + "loss": 2.118, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.068608045578003, + "learning_rate": 4.9695369618196595e-06, + "loss": 2.1884, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.8885847330093384, + "learning_rate": 4.967506092607636e-06, + "loss": 2.2097, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.9441415071487427, + "learning_rate": 4.965475223395614e-06, + "loss": 2.2068, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 2.313577890396118, + "learning_rate": 4.9634443541835905e-06, + "loss": 2.2355, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 1.8542472124099731, + "learning_rate": 4.961413484971569e-06, + "loss": 2.2299, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9382838010787964, + "learning_rate": 4.959382615759545e-06, + "loss": 2.2072, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.361767053604126, + "eval_runtime": 37.4603, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.081242322921753, + "learning_rate": 4.957351746547523e-06, + "loss": 2.3167, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 1.9109065532684326, + "learning_rate": 4.9553208773355e-06, + "loss": 2.2336, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.124897003173828, + "learning_rate": 4.953290008123478e-06, + "loss": 2.184, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.8820433616638184, + "learning_rate": 4.951259138911454e-06, + "loss": 2.2836, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.016204595565796, + "learning_rate": 4.9492282696994315e-06, + "loss": 2.2221, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.0679776668548584, + "learning_rate": 4.947197400487409e-06, + "loss": 2.2113, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 1.8040891885757446, + "learning_rate": 4.945166531275386e-06, + "loss": 2.2292, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.0315871238708496, + "learning_rate": 4.943135662063363e-06, + "loss": 2.2657, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.71120023727417, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.2517, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0443193912506104, + "learning_rate": 4.939073923639318e-06, + "loss": 2.2105, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.3582265377044678, + "eval_runtime": 37.6749, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.9363837242126465, + "learning_rate": 4.937043054427295e-06, + "loss": 2.1944, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 2.0878167152404785, + "learning_rate": 4.935012185215272e-06, + "loss": 2.2477, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.9059010744094849, + "learning_rate": 4.93298131600325e-06, + "loss": 2.2278, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 2.220851421356201, + "learning_rate": 4.930950446791227e-06, + "loss": 2.2391, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.7555570602416992, + "learning_rate": 4.928919577579204e-06, + "loss": 2.2206, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 2.1546502113342285, + "learning_rate": 4.9268887083671815e-06, + "loss": 2.2216, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 1.9688911437988281, + "learning_rate": 4.924857839155159e-06, + "loss": 2.2423, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 1.983587622642517, + "learning_rate": 4.922826969943136e-06, + "loss": 2.1738, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 2.1990034580230713, + "learning_rate": 4.920796100731113e-06, + "loss": 2.2341, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 1.9690128564834595, + "learning_rate": 4.918765231519091e-06, + "loss": 2.1914, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.36171817779541, + "eval_runtime": 37.727, + "eval_samples_per_second": 26.506, + "eval_steps_per_second": 1.113, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 2.5950212478637695, + "learning_rate": 4.916734362307068e-06, + "loss": 2.1921, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 2.023130416870117, + "learning_rate": 4.914703493095045e-06, + "loss": 2.1932, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2486252784729004, + "learning_rate": 4.912672623883022e-06, + "loss": 2.2174, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 1.723677635192871, + "learning_rate": 4.910641754671e-06, + "loss": 2.2173, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.94509756565094, + "learning_rate": 4.908610885458977e-06, + "loss": 2.2539, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 1.6934322118759155, + "learning_rate": 4.906580016246954e-06, + "loss": 2.2345, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 1.6882652044296265, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.231, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.6730620861053467, + "learning_rate": 4.902518277822909e-06, + "loss": 2.2394, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.7261624336242676, + "learning_rate": 4.900487408610886e-06, + "loss": 2.2524, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.8310920000076294, + "learning_rate": 4.898456539398863e-06, + "loss": 2.3004, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.3536295890808105, + "eval_runtime": 37.7461, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.7025809288024902, + "learning_rate": 4.89642567018684e-06, + "loss": 2.1923, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.7285082340240479, + "learning_rate": 4.894394800974818e-06, + "loss": 2.241, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 1.7589807510375977, + "learning_rate": 4.892363931762794e-06, + "loss": 2.2019, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.6664931774139404, + "learning_rate": 4.8903330625507725e-06, + "loss": 2.1822, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 1.8308027982711792, + "learning_rate": 4.888302193338749e-06, + "loss": 2.2651, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.6426894664764404, + "learning_rate": 4.886271324126727e-06, + "loss": 2.2514, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.9765205383300781, + "learning_rate": 4.8842404549147034e-06, + "loss": 2.226, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.7872203588485718, + "learning_rate": 4.8822095857026816e-06, + "loss": 2.18, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 2.0645813941955566, + "learning_rate": 4.880178716490658e-06, + "loss": 2.2023, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 1.9823166131973267, + "learning_rate": 4.878147847278636e-06, + "loss": 2.1801, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.353776454925537, + "eval_runtime": 37.771, + "eval_samples_per_second": 26.475, + "eval_steps_per_second": 1.112, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.831152319908142, + "learning_rate": 4.8761169780666125e-06, + "loss": 2.1858, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 1.9321790933609009, + "learning_rate": 4.874086108854591e-06, + "loss": 2.1638, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.9097596406936646, + "learning_rate": 4.872055239642567e-06, + "loss": 2.2206, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8821815252304077, + "learning_rate": 4.870024370430544e-06, + "loss": 2.2338, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.7959749698638916, + "learning_rate": 4.867993501218522e-06, + "loss": 2.1882, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 1.625320315361023, + "learning_rate": 4.865962632006499e-06, + "loss": 2.2137, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 1.6914536952972412, + "learning_rate": 4.863931762794476e-06, + "loss": 2.2374, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0801584720611572, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.2517, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 1.6874804496765137, + "learning_rate": 4.859870024370431e-06, + "loss": 2.2621, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 1.9347832202911377, + "learning_rate": 4.857839155158408e-06, + "loss": 2.2326, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.3468236923217773, + "eval_runtime": 37.7184, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.6798334121704102, + "learning_rate": 4.855808285946385e-06, + "loss": 2.2222, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 1.8993884325027466, + "learning_rate": 4.8537774167343626e-06, + "loss": 2.2178, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 1.697892189025879, + "learning_rate": 4.85174654752234e-06, + "loss": 2.2217, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8484156131744385, + "learning_rate": 4.849715678310317e-06, + "loss": 2.2069, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 1.8125075101852417, + "learning_rate": 4.847684809098294e-06, + "loss": 2.2285, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 1.8215030431747437, + "learning_rate": 4.845653939886272e-06, + "loss": 2.2029, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 1.777129054069519, + "learning_rate": 4.843623070674249e-06, + "loss": 2.2512, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.7327351570129395, + "learning_rate": 4.841592201462226e-06, + "loss": 2.1767, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.7363202571868896, + "learning_rate": 4.8395613322502035e-06, + "loss": 2.2174, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9183942079544067, + "learning_rate": 4.837530463038181e-06, + "loss": 2.2383, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.345961332321167, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.7290767431259155, + "learning_rate": 4.835499593826158e-06, + "loss": 2.22, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 1.7892694473266602, + "learning_rate": 4.833468724614135e-06, + "loss": 2.1403, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.6671900749206543, + "learning_rate": 4.831437855402113e-06, + "loss": 2.1989, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 1.848751425743103, + "learning_rate": 4.82940698619009e-06, + "loss": 2.245, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.6957710981369019, + "learning_rate": 4.827376116978067e-06, + "loss": 2.2239, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.650583267211914, + "learning_rate": 4.8253452477660444e-06, + "loss": 2.2101, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.6510974168777466, + "learning_rate": 4.823314378554022e-06, + "loss": 2.204, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 1.9469026327133179, + "learning_rate": 4.821283509341998e-06, + "loss": 2.221, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 1.6031194925308228, + "learning_rate": 4.819252640129976e-06, + "loss": 2.1926, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.120325803756714, + "learning_rate": 4.817221770917953e-06, + "loss": 2.2166, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.3419017791748047, + "eval_runtime": 37.6854, + "eval_samples_per_second": 26.535, + "eval_steps_per_second": 1.114, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.5533722639083862, + "learning_rate": 4.815190901705931e-06, + "loss": 2.1981, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9650007486343384, + "learning_rate": 4.813160032493907e-06, + "loss": 2.2319, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 1.6874849796295166, + "learning_rate": 4.811129163281885e-06, + "loss": 2.2263, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.787996768951416, + "learning_rate": 4.809098294069862e-06, + "loss": 2.2546, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 1.7853440046310425, + "learning_rate": 4.80706742485784e-06, + "loss": 2.1994, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.7882641553878784, + "learning_rate": 4.805036555645816e-06, + "loss": 2.2449, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 1.8206003904342651, + "learning_rate": 4.8030056864337945e-06, + "loss": 2.2144, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.7012749910354614, + "learning_rate": 4.800974817221771e-06, + "loss": 2.2122, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 1.8711670637130737, + "learning_rate": 4.798943948009749e-06, + "loss": 2.2132, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 1.6880611181259155, + "learning_rate": 4.7969130787977254e-06, + "loss": 2.2336, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.340393543243408, + "eval_runtime": 37.6885, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 2.0543065071105957, + "learning_rate": 4.794882209585703e-06, + "loss": 2.1899, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 1.5774480104446411, + "learning_rate": 4.79285134037368e-06, + "loss": 2.1982, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 1.7154620885849, + "learning_rate": 4.790820471161657e-06, + "loss": 2.2096, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.8200151920318604, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.2391, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 1.6368298530578613, + "learning_rate": 4.786758732737612e-06, + "loss": 2.1855, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 1.6851340532302856, + "learning_rate": 4.784727863525589e-06, + "loss": 2.2114, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 1.8293818235397339, + "learning_rate": 4.782696994313566e-06, + "loss": 2.1636, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 1.6213513612747192, + "learning_rate": 4.780666125101544e-06, + "loss": 2.1884, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.7791825532913208, + "learning_rate": 4.778635255889521e-06, + "loss": 2.1977, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 1.6627994775772095, + "learning_rate": 4.776604386677498e-06, + "loss": 2.2444, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.3396637439727783, + "eval_runtime": 37.7069, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 1.7597277164459229, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.2369, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.7754322290420532, + "learning_rate": 4.772542648253453e-06, + "loss": 2.1859, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 1.7897601127624512, + "learning_rate": 4.77051177904143e-06, + "loss": 2.2245, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 1.7294282913208008, + "learning_rate": 4.768480909829407e-06, + "loss": 2.1998, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 1.8406974077224731, + "learning_rate": 4.766450040617385e-06, + "loss": 2.2353, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 1.849493384361267, + "learning_rate": 4.764419171405362e-06, + "loss": 2.2165, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.5568877458572388, + "learning_rate": 4.762388302193339e-06, + "loss": 2.1989, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 1.8816689252853394, + "learning_rate": 4.760357432981316e-06, + "loss": 2.2062, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 1.7648332118988037, + "learning_rate": 4.758326563769294e-06, + "loss": 2.2204, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.753944754600525, + "learning_rate": 4.756295694557271e-06, + "loss": 2.1748, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.339653253555298, + "eval_runtime": 37.4438, + "eval_samples_per_second": 26.707, + "eval_steps_per_second": 1.122, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.7343329191207886, + "learning_rate": 4.754264825345248e-06, + "loss": 2.1843, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 1.8125407695770264, + "learning_rate": 4.7522339561332255e-06, + "loss": 2.1492, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.5255184173583984, + "learning_rate": 4.750203086921203e-06, + "loss": 2.2178, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 1.6692545413970947, + "learning_rate": 4.74817221770918e-06, + "loss": 2.2516, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 1.7804819345474243, + "learning_rate": 4.7461413484971565e-06, + "loss": 2.1592, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8048933744430542, + "learning_rate": 4.744110479285135e-06, + "loss": 2.1869, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 1.6439207792282104, + "learning_rate": 4.742079610073111e-06, + "loss": 2.1995, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.7375173568725586, + "learning_rate": 4.740048740861089e-06, + "loss": 2.2067, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 1.9471659660339355, + "learning_rate": 4.738017871649066e-06, + "loss": 2.2513, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 1.6524564027786255, + "learning_rate": 4.735987002437044e-06, + "loss": 2.2643, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.337153434753418, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 1.8574563264846802, + "learning_rate": 4.73395613322502e-06, + "loss": 2.2373, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 1.836464524269104, + "learning_rate": 4.731925264012998e-06, + "loss": 2.1769, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 1.7012423276901245, + "learning_rate": 4.729894394800975e-06, + "loss": 2.1494, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 1.953180193901062, + "learning_rate": 4.727863525588953e-06, + "loss": 2.186, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 1.8468470573425293, + "learning_rate": 4.725832656376929e-06, + "loss": 2.2279, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 2.005077362060547, + "learning_rate": 4.723801787164907e-06, + "loss": 2.1798, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 1.8246952295303345, + "learning_rate": 4.721770917952885e-06, + "loss": 2.1916, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 1.8558839559555054, + "learning_rate": 4.719740048740861e-06, + "loss": 2.2237, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 1.7670999765396118, + "learning_rate": 4.717709179528839e-06, + "loss": 2.2179, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.8544219732284546, + "learning_rate": 4.715678310316816e-06, + "loss": 2.1811, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.333791732788086, + "eval_runtime": 37.7254, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 1.5792925357818604, + "learning_rate": 4.713647441104794e-06, + "loss": 2.1749, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 1.5678222179412842, + "learning_rate": 4.71161657189277e-06, + "loss": 2.2294, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 1.9467812776565552, + "learning_rate": 4.709585702680748e-06, + "loss": 2.2684, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 1.6647964715957642, + "learning_rate": 4.707554833468725e-06, + "loss": 2.1722, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 1.7037622928619385, + "learning_rate": 4.705523964256703e-06, + "loss": 2.1915, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.5857148170471191, + "learning_rate": 4.703493095044679e-06, + "loss": 2.1466, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 1.5792173147201538, + "learning_rate": 4.701462225832657e-06, + "loss": 2.2472, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 1.5998480319976807, + "learning_rate": 4.699431356620634e-06, + "loss": 2.2129, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.6728194952011108, + "learning_rate": 4.697400487408612e-06, + "loss": 2.2057, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 1.5402852296829224, + "learning_rate": 4.695369618196588e-06, + "loss": 2.1903, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.3325934410095215, + "eval_runtime": 37.7463, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 1.5890164375305176, + "learning_rate": 4.693338748984566e-06, + "loss": 2.1584, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 1.805923581123352, + "learning_rate": 4.691307879772543e-06, + "loss": 2.2263, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.5932700634002686, + "learning_rate": 4.68927701056052e-06, + "loss": 2.2109, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 1.5611685514450073, + "learning_rate": 4.6872461413484975e-06, + "loss": 2.192, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 1.6976194381713867, + "learning_rate": 4.685215272136475e-06, + "loss": 2.1998, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.7149630784988403, + "learning_rate": 4.683184402924452e-06, + "loss": 2.2599, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 1.7597423791885376, + "learning_rate": 4.681153533712429e-06, + "loss": 2.1765, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 1.539652943611145, + "learning_rate": 4.679122664500407e-06, + "loss": 2.1635, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.690402865409851, + "learning_rate": 4.677091795288384e-06, + "loss": 2.2417, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.7267673015594482, + "learning_rate": 4.675060926076361e-06, + "loss": 2.166, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.331728935241699, + "eval_runtime": 37.7133, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.6383073329925537, + "learning_rate": 4.673030056864338e-06, + "loss": 2.1624, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.5364140272140503, + "learning_rate": 4.670999187652316e-06, + "loss": 2.1696, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.591477870941162, + "learning_rate": 4.668968318440293e-06, + "loss": 2.2306, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 1.8121449947357178, + "learning_rate": 4.66693744922827e-06, + "loss": 2.1449, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.515539288520813, + "learning_rate": 4.6649065800162475e-06, + "loss": 2.2108, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 1.6521453857421875, + "learning_rate": 4.662875710804225e-06, + "loss": 2.1398, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 1.7074251174926758, + "learning_rate": 4.660844841592202e-06, + "loss": 2.2342, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.7688119411468506, + "learning_rate": 4.658813972380179e-06, + "loss": 2.1287, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.7170679569244385, + "learning_rate": 4.656783103168157e-06, + "loss": 2.2143, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 1.7401350736618042, + "learning_rate": 4.654752233956134e-06, + "loss": 2.1855, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.335416555404663, + "eval_runtime": 37.7115, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.6344956159591675, + "learning_rate": 4.652721364744111e-06, + "loss": 2.2205, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.629184603691101, + "learning_rate": 4.6506904955320885e-06, + "loss": 2.2271, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.769491195678711, + "learning_rate": 4.648659626320066e-06, + "loss": 2.1945, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.6513251066207886, + "learning_rate": 4.646628757108043e-06, + "loss": 2.1705, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9357609748840332, + "learning_rate": 4.6445978878960194e-06, + "loss": 2.1968, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 1.75972318649292, + "learning_rate": 4.6425670186839976e-06, + "loss": 2.2229, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 1.686574935913086, + "learning_rate": 4.640536149471974e-06, + "loss": 2.2361, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 1.655661702156067, + "learning_rate": 4.638505280259952e-06, + "loss": 2.2064, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.7583094835281372, + "learning_rate": 4.6364744110479285e-06, + "loss": 2.2352, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.5906383991241455, + "learning_rate": 4.634443541835907e-06, + "loss": 2.187, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.332923412322998, + "eval_runtime": 37.7496, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 1.708019733428955, + "learning_rate": 4.632412672623883e-06, + "loss": 2.1032, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 1.6350252628326416, + "learning_rate": 4.630381803411861e-06, + "loss": 2.1976, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.9037389755249023, + "learning_rate": 4.628350934199838e-06, + "loss": 2.224, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 1.7126816511154175, + "learning_rate": 4.626320064987816e-06, + "loss": 2.1925, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 1.5828406810760498, + "learning_rate": 4.624289195775792e-06, + "loss": 2.1928, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.6797102689743042, + "learning_rate": 4.62225832656377e-06, + "loss": 2.1902, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 1.6871609687805176, + "learning_rate": 4.620227457351747e-06, + "loss": 2.2024, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.8785804510116577, + "learning_rate": 4.618196588139724e-06, + "loss": 2.1778, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.6175488233566284, + "learning_rate": 4.616165718927701e-06, + "loss": 2.2438, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.7724019289016724, + "learning_rate": 4.6141348497156786e-06, + "loss": 2.2272, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.328923225402832, + "eval_runtime": 37.7443, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.7230926752090454, + "learning_rate": 4.612103980503656e-06, + "loss": 2.2021, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.7899917364120483, + "learning_rate": 4.610073111291633e-06, + "loss": 2.1854, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.5864417552947998, + "learning_rate": 4.60804224207961e-06, + "loss": 2.2042, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.875375509262085, + "learning_rate": 4.606011372867588e-06, + "loss": 2.1844, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 1.830597162246704, + "learning_rate": 4.603980503655565e-06, + "loss": 2.1474, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 1.8624461889266968, + "learning_rate": 4.601949634443542e-06, + "loss": 2.158, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.7161756753921509, + "learning_rate": 4.5999187652315195e-06, + "loss": 2.1838, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.6212890148162842, + "learning_rate": 4.597887896019497e-06, + "loss": 2.1911, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.8586416244506836, + "learning_rate": 4.595857026807474e-06, + "loss": 2.1794, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.653976559638977, + "learning_rate": 4.593826157595451e-06, + "loss": 2.2428, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.3279106616973877, + "eval_runtime": 37.739, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 2.054537057876587, + "learning_rate": 4.591795288383429e-06, + "loss": 2.2003, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.730942964553833, + "learning_rate": 4.589764419171406e-06, + "loss": 2.1816, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.7264728546142578, + "learning_rate": 4.587733549959383e-06, + "loss": 2.1615, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 1.8276281356811523, + "learning_rate": 4.5857026807473604e-06, + "loss": 2.1442, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 1.9947556257247925, + "learning_rate": 4.583671811535338e-06, + "loss": 2.2192, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 1.770601511001587, + "learning_rate": 4.581640942323315e-06, + "loss": 2.233, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 1.8946491479873657, + "learning_rate": 4.579610073111292e-06, + "loss": 2.2249, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 1.7839634418487549, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.2469, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.6298807859420776, + "learning_rate": 4.575548334687247e-06, + "loss": 2.1723, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 1.9310041666030884, + "learning_rate": 4.573517465475224e-06, + "loss": 2.2156, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.3249237537384033, + "eval_runtime": 37.7505, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 1.7901647090911865, + "learning_rate": 4.571486596263201e-06, + "loss": 2.2169, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.7739574909210205, + "learning_rate": 4.569455727051178e-06, + "loss": 2.2005, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 1.7310121059417725, + "learning_rate": 4.567424857839156e-06, + "loss": 2.2369, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.004469156265259, + "learning_rate": 4.565393988627132e-06, + "loss": 2.2051, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 1.7906666994094849, + "learning_rate": 4.5633631194151105e-06, + "loss": 2.1908, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 1.7224774360656738, + "learning_rate": 4.561332250203087e-06, + "loss": 2.1529, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 1.9950156211853027, + "learning_rate": 4.559301380991065e-06, + "loss": 2.1776, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 1.8257590532302856, + "learning_rate": 4.5572705117790414e-06, + "loss": 2.2267, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.038490056991577, + "learning_rate": 4.5552396425670196e-06, + "loss": 2.1703, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 1.7314448356628418, + "learning_rate": 4.553208773354996e-06, + "loss": 2.2409, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.327078104019165, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 2.0013606548309326, + "learning_rate": 4.551177904142974e-06, + "loss": 2.1732, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.7352690696716309, + "learning_rate": 4.5491470349309505e-06, + "loss": 2.2027, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9024865627288818, + "learning_rate": 4.547116165718929e-06, + "loss": 2.1806, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8870717287063599, + "learning_rate": 4.545085296506905e-06, + "loss": 2.1664, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.781226634979248, + "learning_rate": 4.543054427294882e-06, + "loss": 2.166, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.8994221687316895, + "learning_rate": 4.54102355808286e-06, + "loss": 2.1951, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.6999517679214478, + "learning_rate": 4.538992688870837e-06, + "loss": 2.1786, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.899590253829956, + "learning_rate": 4.536961819658814e-06, + "loss": 2.2051, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.8054970502853394, + "learning_rate": 4.5349309504467915e-06, + "loss": 2.1675, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.6732861995697021, + "learning_rate": 4.532900081234769e-06, + "loss": 2.1997, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.3242335319519043, + "eval_runtime": 37.6748, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 2.0063157081604004, + "learning_rate": 4.530869212022746e-06, + "loss": 2.1909, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.7750036716461182, + "learning_rate": 4.528838342810723e-06, + "loss": 2.1935, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9993029832839966, + "learning_rate": 4.526807473598701e-06, + "loss": 2.1753, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.6742994785308838, + "learning_rate": 4.524776604386678e-06, + "loss": 2.2205, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 1.9300428628921509, + "learning_rate": 4.522745735174655e-06, + "loss": 2.2081, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.8770017623901367, + "learning_rate": 4.520714865962632e-06, + "loss": 2.2074, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.847765326499939, + "learning_rate": 4.51868399675061e-06, + "loss": 2.1775, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.905117392539978, + "learning_rate": 4.516653127538587e-06, + "loss": 2.1511, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.7623670101165771, + "learning_rate": 4.514622258326564e-06, + "loss": 2.188, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.9232176542282104, + "learning_rate": 4.5125913891145415e-06, + "loss": 2.2109, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.3252475261688232, + "eval_runtime": 37.6923, + "eval_samples_per_second": 26.531, + "eval_steps_per_second": 1.114, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.6523535251617432, + "learning_rate": 4.510560519902519e-06, + "loss": 2.1499, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9934184551239014, + "learning_rate": 4.508529650690496e-06, + "loss": 2.1383, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.704522728919983, + "learning_rate": 4.506498781478473e-06, + "loss": 2.1892, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.9909989833831787, + "learning_rate": 4.504467912266451e-06, + "loss": 2.2422, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.7490800619125366, + "learning_rate": 4.502437043054428e-06, + "loss": 2.2008, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8620145320892334, + "learning_rate": 4.500406173842405e-06, + "loss": 2.1892, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.7069931030273438, + "learning_rate": 4.4983753046303824e-06, + "loss": 2.1816, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 2.152491807937622, + "learning_rate": 4.49634443541836e-06, + "loss": 2.2331, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.8526151180267334, + "learning_rate": 4.494313566206336e-06, + "loss": 2.1811, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.6865259408950806, + "learning_rate": 4.492282696994314e-06, + "loss": 2.1599, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.3265392780303955, + "eval_runtime": 37.7255, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.7710354328155518, + "learning_rate": 4.490251827782291e-06, + "loss": 2.1636, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.8444210290908813, + "learning_rate": 4.488220958570269e-06, + "loss": 2.1345, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.9095014333724976, + "learning_rate": 4.486190089358245e-06, + "loss": 2.2236, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.7406108379364014, + "learning_rate": 4.484159220146223e-06, + "loss": 2.2078, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.712534785270691, + "learning_rate": 4.4821283509342e-06, + "loss": 2.1809, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8921459913253784, + "learning_rate": 4.480097481722178e-06, + "loss": 2.158, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.805146336555481, + "learning_rate": 4.478066612510154e-06, + "loss": 2.1765, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 2.1620516777038574, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.2832, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.661618947982788, + "learning_rate": 4.474004874086109e-06, + "loss": 2.1397, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.732483983039856, + "learning_rate": 4.471974004874087e-06, + "loss": 2.225, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.3276846408843994, + "eval_runtime": 37.7156, + "eval_samples_per_second": 26.514, + "eval_steps_per_second": 1.114, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.8693495988845825, + "learning_rate": 4.4699431356620635e-06, + "loss": 2.1959, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8201690912246704, + "learning_rate": 4.467912266450041e-06, + "loss": 2.1882, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.9229968786239624, + "learning_rate": 4.465881397238018e-06, + "loss": 2.2006, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.6330770254135132, + "learning_rate": 4.463850528025995e-06, + "loss": 2.209, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9654744863510132, + "learning_rate": 4.4618196588139726e-06, + "loss": 2.1963, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.8808507919311523, + "learning_rate": 4.45978878960195e-06, + "loss": 2.1538, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 1.8481040000915527, + "learning_rate": 4.457757920389927e-06, + "loss": 2.1769, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 2.114439010620117, + "learning_rate": 4.455727051177904e-06, + "loss": 2.2241, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.5492331981658936, + "learning_rate": 4.453696181965882e-06, + "loss": 2.206, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 2.11787486076355, + "learning_rate": 4.451665312753859e-06, + "loss": 2.2328, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.3215034008026123, + "eval_runtime": 37.7471, + "eval_samples_per_second": 26.492, + "eval_steps_per_second": 1.113, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.6561832427978516, + "learning_rate": 4.449634443541836e-06, + "loss": 2.1607, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.7510966062545776, + "learning_rate": 4.4476035743298135e-06, + "loss": 2.1699, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.9424657821655273, + "learning_rate": 4.445572705117791e-06, + "loss": 2.2298, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 1.5836161375045776, + "learning_rate": 4.443541835905768e-06, + "loss": 2.1711, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.895216464996338, + "learning_rate": 4.441510966693745e-06, + "loss": 2.2252, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.3365776538848877, + "learning_rate": 4.439480097481723e-06, + "loss": 2.1309, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.6988815069198608, + "learning_rate": 4.4374492282697e-06, + "loss": 2.2023, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 2.228900671005249, + "learning_rate": 4.435418359057677e-06, + "loss": 2.2021, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.5536632537841797, + "learning_rate": 4.433387489845654e-06, + "loss": 2.1706, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.8463878631591797, + "learning_rate": 4.431356620633632e-06, + "loss": 2.1715, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.320664167404175, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.924741506576538, + "learning_rate": 4.429325751421609e-06, + "loss": 2.1894, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 1.5921828746795654, + "learning_rate": 4.427294882209586e-06, + "loss": 2.1358, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 2.4617693424224854, + "learning_rate": 4.4252640129975635e-06, + "loss": 2.1918, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.6239196062088013, + "learning_rate": 4.423233143785541e-06, + "loss": 2.1996, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.952112078666687, + "learning_rate": 4.421202274573518e-06, + "loss": 2.112, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.663409948348999, + "learning_rate": 4.4191714053614945e-06, + "loss": 2.149, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.831711769104004, + "learning_rate": 4.417140536149473e-06, + "loss": 2.209, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 1.9281738996505737, + "learning_rate": 4.415109666937449e-06, + "loss": 2.1637, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.5661600828170776, + "learning_rate": 4.413078797725427e-06, + "loss": 2.1832, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.6930522918701172, + "learning_rate": 4.411047928513404e-06, + "loss": 2.196, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.318861246109009, + "eval_runtime": 37.6994, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.74558424949646, + "learning_rate": 4.409017059301382e-06, + "loss": 2.2097, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.664880633354187, + "learning_rate": 4.406986190089358e-06, + "loss": 2.2076, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.5981497764587402, + "learning_rate": 4.404955320877336e-06, + "loss": 2.1706, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.6035112142562866, + "learning_rate": 4.402924451665313e-06, + "loss": 2.1538, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 1.6819255352020264, + "learning_rate": 4.400893582453291e-06, + "loss": 2.1816, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.7082457542419434, + "learning_rate": 4.398862713241267e-06, + "loss": 2.1804, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.6893151998519897, + "learning_rate": 4.396831844029245e-06, + "loss": 2.1783, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 1.8639923334121704, + "learning_rate": 4.394800974817222e-06, + "loss": 2.1989, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.6407119035720825, + "learning_rate": 4.392770105605199e-06, + "loss": 2.2013, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.6435848474502563, + "learning_rate": 4.390739236393176e-06, + "loss": 2.2049, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.318084478378296, + "eval_runtime": 37.6999, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.7974674701690674, + "learning_rate": 4.388708367181154e-06, + "loss": 2.188, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 1.8340431451797485, + "learning_rate": 4.386677497969131e-06, + "loss": 2.2305, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.55428946018219, + "learning_rate": 4.384646628757108e-06, + "loss": 2.1837, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.7258585691452026, + "learning_rate": 4.3826157595450855e-06, + "loss": 2.1426, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.7347694635391235, + "learning_rate": 4.380584890333063e-06, + "loss": 2.1161, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.6082992553710938, + "learning_rate": 4.37855402112104e-06, + "loss": 2.2284, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.9446425437927246, + "learning_rate": 4.376523151909017e-06, + "loss": 2.1571, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 1.8597826957702637, + "learning_rate": 4.3744922826969946e-06, + "loss": 2.1982, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.8589327335357666, + "learning_rate": 4.372461413484972e-06, + "loss": 2.2097, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.6825448274612427, + "learning_rate": 4.370430544272949e-06, + "loss": 2.2077, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.315889596939087, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 600 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 1.870385645617152e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/generation_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/latest b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/latest new file mode 100644 index 0000000000000000000000000000000000000000..4b10acccf3e8395339ff8799cea202bbc54d7f7d --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/latest @@ -0,0 +1 @@ +global_step900 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/model.safetensors.index.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/special_tokens_map.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/tokenizer_config.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/tokenizer_config.json @@ -0,0 +1,2070 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/trainer_state.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..a4495e7cb9ea7e6bda86a9714a5e7db1ec79b7cf --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/trainer_state.json @@ -0,0 +1,7053 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.9863013698630136, + "eval_steps": 10, + "global_step": 900, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001095890410958904, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0998, + "step": 1 + }, + { + "epoch": 0.002191780821917808, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0858, + "step": 2 + }, + { + "epoch": 0.0032876712328767125, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0672, + "step": 3 + }, + { + "epoch": 0.004383561643835616, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0802, + "step": 4 + }, + { + "epoch": 0.005479452054794521, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0636, + "step": 5 + }, + { + "epoch": 0.006575342465753425, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1672, + "step": 6 + }, + { + "epoch": 0.007671232876712329, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0961, + "step": 7 + }, + { + "epoch": 0.008767123287671232, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1258, + "step": 8 + }, + { + "epoch": 0.009863013698630137, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0953, + "step": 9 + }, + { + "epoch": 0.010958904109589041, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1404, + "step": 10 + }, + { + "epoch": 0.010958904109589041, + "eval_loss": 3.1298255920410156, + "eval_runtime": 38.6417, + "eval_samples_per_second": 25.879, + "eval_steps_per_second": 1.087, + "step": 10 + }, + { + "epoch": 0.012054794520547946, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.1187, + "step": 11 + }, + { + "epoch": 0.01315068493150685, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0593, + "step": 12 + }, + { + "epoch": 0.014246575342465753, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.111, + "step": 13 + }, + { + "epoch": 0.015342465753424657, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 3.0974, + "step": 14 + }, + { + "epoch": 0.01643835616438356, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.135, + "step": 15 + }, + { + "epoch": 0.017534246575342465, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0799, + "step": 16 + }, + { + "epoch": 0.01863013698630137, + "grad_norm": 5.991453647613525, + "learning_rate": 1.8248175182481753e-08, + "loss": 3.0851, + "step": 17 + }, + { + "epoch": 0.019726027397260273, + "grad_norm": 10.659249305725098, + "learning_rate": 3.6496350364963505e-08, + "loss": 3.1008, + "step": 18 + }, + { + "epoch": 0.020821917808219178, + "grad_norm": 8.556068420410156, + "learning_rate": 5.4744525547445264e-08, + "loss": 3.0962, + "step": 19 + }, + { + "epoch": 0.021917808219178082, + "grad_norm": 5.162924766540527, + "learning_rate": 7.299270072992701e-08, + "loss": 3.0638, + "step": 20 + }, + { + "epoch": 0.021917808219178082, + "eval_loss": 3.0002310276031494, + "eval_runtime": 38.1663, + "eval_samples_per_second": 26.201, + "eval_steps_per_second": 1.1, + "step": 20 + }, + { + "epoch": 0.023013698630136987, + "grad_norm": 4.7945966720581055, + "learning_rate": 9.124087591240876e-08, + "loss": 2.9786, + "step": 21 + }, + { + "epoch": 0.02410958904109589, + "grad_norm": 5.523077487945557, + "learning_rate": 1.0948905109489053e-07, + "loss": 2.8652, + "step": 22 + }, + { + "epoch": 0.025205479452054796, + "grad_norm": 2.859027862548828, + "learning_rate": 1.2773722627737227e-07, + "loss": 2.8468, + "step": 23 + }, + { + "epoch": 0.0263013698630137, + "grad_norm": 4.805753707885742, + "learning_rate": 1.4598540145985402e-07, + "loss": 2.8626, + "step": 24 + }, + { + "epoch": 0.0273972602739726, + "grad_norm": 3.050151824951172, + "learning_rate": 1.642335766423358e-07, + "loss": 2.8487, + "step": 25 + }, + { + "epoch": 0.028493150684931506, + "grad_norm": 2.5798933506011963, + "learning_rate": 1.8248175182481753e-07, + "loss": 2.8154, + "step": 26 + }, + { + "epoch": 0.02958904109589041, + "grad_norm": 2.379521369934082, + "learning_rate": 2.0072992700729928e-07, + "loss": 2.8028, + "step": 27 + }, + { + "epoch": 0.030684931506849315, + "grad_norm": 2.509614944458008, + "learning_rate": 2.1897810218978106e-07, + "loss": 2.7474, + "step": 28 + }, + { + "epoch": 0.03178082191780822, + "grad_norm": 2.5612199306488037, + "learning_rate": 2.3722627737226278e-07, + "loss": 2.7206, + "step": 29 + }, + { + "epoch": 0.03287671232876712, + "grad_norm": 2.068345069885254, + "learning_rate": 2.5547445255474454e-07, + "loss": 2.6676, + "step": 30 + }, + { + "epoch": 0.03287671232876712, + "eval_loss": 2.762876033782959, + "eval_runtime": 37.7329, + "eval_samples_per_second": 26.502, + "eval_steps_per_second": 1.113, + "step": 30 + }, + { + "epoch": 0.03397260273972603, + "grad_norm": 2.1446118354797363, + "learning_rate": 2.737226277372263e-07, + "loss": 2.692, + "step": 31 + }, + { + "epoch": 0.03506849315068493, + "grad_norm": 2.0872929096221924, + "learning_rate": 2.9197080291970804e-07, + "loss": 2.6617, + "step": 32 + }, + { + "epoch": 0.03616438356164384, + "grad_norm": 2.2962639331817627, + "learning_rate": 3.102189781021898e-07, + "loss": 2.6252, + "step": 33 + }, + { + "epoch": 0.03726027397260274, + "grad_norm": 2.0494158267974854, + "learning_rate": 3.284671532846716e-07, + "loss": 2.6068, + "step": 34 + }, + { + "epoch": 0.038356164383561646, + "grad_norm": 2.2687947750091553, + "learning_rate": 3.467153284671533e-07, + "loss": 2.6194, + "step": 35 + }, + { + "epoch": 0.03945205479452055, + "grad_norm": 3.141436815261841, + "learning_rate": 3.6496350364963505e-07, + "loss": 2.6304, + "step": 36 + }, + { + "epoch": 0.040547945205479455, + "grad_norm": 1.7621077299118042, + "learning_rate": 3.8321167883211686e-07, + "loss": 2.5855, + "step": 37 + }, + { + "epoch": 0.041643835616438356, + "grad_norm": 2.396498441696167, + "learning_rate": 4.0145985401459856e-07, + "loss": 2.5497, + "step": 38 + }, + { + "epoch": 0.042739726027397264, + "grad_norm": 1.8160557746887207, + "learning_rate": 4.197080291970803e-07, + "loss": 2.6161, + "step": 39 + }, + { + "epoch": 0.043835616438356165, + "grad_norm": 2.3418030738830566, + "learning_rate": 4.379562043795621e-07, + "loss": 2.5377, + "step": 40 + }, + { + "epoch": 0.043835616438356165, + "eval_loss": 2.6371867656707764, + "eval_runtime": 37.7063, + "eval_samples_per_second": 26.521, + "eval_steps_per_second": 1.114, + "step": 40 + }, + { + "epoch": 0.044931506849315066, + "grad_norm": 1.572534203529358, + "learning_rate": 4.562043795620438e-07, + "loss": 2.5333, + "step": 41 + }, + { + "epoch": 0.046027397260273974, + "grad_norm": 2.2532575130462646, + "learning_rate": 4.7445255474452557e-07, + "loss": 2.5797, + "step": 42 + }, + { + "epoch": 0.047123287671232875, + "grad_norm": 1.5870578289031982, + "learning_rate": 4.927007299270073e-07, + "loss": 2.5446, + "step": 43 + }, + { + "epoch": 0.04821917808219178, + "grad_norm": 1.848671317100525, + "learning_rate": 5.109489051094891e-07, + "loss": 2.561, + "step": 44 + }, + { + "epoch": 0.049315068493150684, + "grad_norm": 1.8653115034103394, + "learning_rate": 5.291970802919708e-07, + "loss": 2.5415, + "step": 45 + }, + { + "epoch": 0.05041095890410959, + "grad_norm": 1.8786529302597046, + "learning_rate": 5.474452554744526e-07, + "loss": 2.5034, + "step": 46 + }, + { + "epoch": 0.05150684931506849, + "grad_norm": 1.9696972370147705, + "learning_rate": 5.656934306569343e-07, + "loss": 2.4951, + "step": 47 + }, + { + "epoch": 0.0526027397260274, + "grad_norm": 1.6735143661499023, + "learning_rate": 5.839416058394161e-07, + "loss": 2.5062, + "step": 48 + }, + { + "epoch": 0.0536986301369863, + "grad_norm": 1.860193133354187, + "learning_rate": 6.021897810218979e-07, + "loss": 2.4351, + "step": 49 + }, + { + "epoch": 0.0547945205479452, + "grad_norm": 1.6798452138900757, + "learning_rate": 6.204379562043796e-07, + "loss": 2.5634, + "step": 50 + }, + { + "epoch": 0.0547945205479452, + "eval_loss": 2.569493055343628, + "eval_runtime": 37.7225, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 50 + }, + { + "epoch": 0.05589041095890411, + "grad_norm": 1.7167212963104248, + "learning_rate": 6.386861313868614e-07, + "loss": 2.4502, + "step": 51 + }, + { + "epoch": 0.05698630136986301, + "grad_norm": 1.6960467100143433, + "learning_rate": 6.569343065693432e-07, + "loss": 2.4877, + "step": 52 + }, + { + "epoch": 0.05808219178082192, + "grad_norm": 1.4868950843811035, + "learning_rate": 6.751824817518248e-07, + "loss": 2.426, + "step": 53 + }, + { + "epoch": 0.05917808219178082, + "grad_norm": 1.633758544921875, + "learning_rate": 6.934306569343066e-07, + "loss": 2.4438, + "step": 54 + }, + { + "epoch": 0.06027397260273973, + "grad_norm": 1.5697351694107056, + "learning_rate": 7.116788321167883e-07, + "loss": 2.4505, + "step": 55 + }, + { + "epoch": 0.06136986301369863, + "grad_norm": 1.7102861404418945, + "learning_rate": 7.299270072992701e-07, + "loss": 2.4659, + "step": 56 + }, + { + "epoch": 0.06246575342465754, + "grad_norm": 1.6420042514801025, + "learning_rate": 7.48175182481752e-07, + "loss": 2.4728, + "step": 57 + }, + { + "epoch": 0.06356164383561644, + "grad_norm": 1.7134343385696411, + "learning_rate": 7.664233576642337e-07, + "loss": 2.4531, + "step": 58 + }, + { + "epoch": 0.06465753424657535, + "grad_norm": 1.7748825550079346, + "learning_rate": 7.846715328467154e-07, + "loss": 2.4263, + "step": 59 + }, + { + "epoch": 0.06575342465753424, + "grad_norm": 1.7815611362457275, + "learning_rate": 8.029197080291971e-07, + "loss": 2.4049, + "step": 60 + }, + { + "epoch": 0.06575342465753424, + "eval_loss": 2.531026840209961, + "eval_runtime": 37.7203, + "eval_samples_per_second": 26.511, + "eval_steps_per_second": 1.113, + "step": 60 + }, + { + "epoch": 0.06684931506849315, + "grad_norm": 1.9545243978500366, + "learning_rate": 8.211678832116789e-07, + "loss": 2.4158, + "step": 61 + }, + { + "epoch": 0.06794520547945206, + "grad_norm": 1.5614768266677856, + "learning_rate": 8.394160583941606e-07, + "loss": 2.4248, + "step": 62 + }, + { + "epoch": 0.06904109589041096, + "grad_norm": 1.7685444355010986, + "learning_rate": 8.576642335766425e-07, + "loss": 2.4311, + "step": 63 + }, + { + "epoch": 0.07013698630136986, + "grad_norm": 1.742304801940918, + "learning_rate": 8.759124087591242e-07, + "loss": 2.4514, + "step": 64 + }, + { + "epoch": 0.07123287671232877, + "grad_norm": 1.4872788190841675, + "learning_rate": 8.94160583941606e-07, + "loss": 2.4034, + "step": 65 + }, + { + "epoch": 0.07232876712328767, + "grad_norm": 1.735102653503418, + "learning_rate": 9.124087591240876e-07, + "loss": 2.426, + "step": 66 + }, + { + "epoch": 0.07342465753424658, + "grad_norm": 1.9607082605361938, + "learning_rate": 9.306569343065694e-07, + "loss": 2.4188, + "step": 67 + }, + { + "epoch": 0.07452054794520548, + "grad_norm": 1.8506171703338623, + "learning_rate": 9.489051094890511e-07, + "loss": 2.4516, + "step": 68 + }, + { + "epoch": 0.07561643835616438, + "grad_norm": 2.18963360786438, + "learning_rate": 9.67153284671533e-07, + "loss": 2.4068, + "step": 69 + }, + { + "epoch": 0.07671232876712329, + "grad_norm": 3.0092129707336426, + "learning_rate": 9.854014598540146e-07, + "loss": 2.3709, + "step": 70 + }, + { + "epoch": 0.07671232876712329, + "eval_loss": 2.5034940242767334, + "eval_runtime": 37.8469, + "eval_samples_per_second": 26.422, + "eval_steps_per_second": 1.11, + "step": 70 + }, + { + "epoch": 0.07780821917808219, + "grad_norm": 1.9387335777282715, + "learning_rate": 1.0036496350364965e-06, + "loss": 2.4394, + "step": 71 + }, + { + "epoch": 0.0789041095890411, + "grad_norm": 3.146247386932373, + "learning_rate": 1.0218978102189781e-06, + "loss": 2.355, + "step": 72 + }, + { + "epoch": 0.08, + "grad_norm": 2.447643518447876, + "learning_rate": 1.04014598540146e-06, + "loss": 2.3574, + "step": 73 + }, + { + "epoch": 0.08109589041095891, + "grad_norm": 2.1435651779174805, + "learning_rate": 1.0583941605839416e-06, + "loss": 2.374, + "step": 74 + }, + { + "epoch": 0.0821917808219178, + "grad_norm": 2.143691301345825, + "learning_rate": 1.0766423357664233e-06, + "loss": 2.4076, + "step": 75 + }, + { + "epoch": 0.08328767123287671, + "grad_norm": 1.8463176488876343, + "learning_rate": 1.0948905109489052e-06, + "loss": 2.4193, + "step": 76 + }, + { + "epoch": 0.08438356164383562, + "grad_norm": 1.906254529953003, + "learning_rate": 1.113138686131387e-06, + "loss": 2.3622, + "step": 77 + }, + { + "epoch": 0.08547945205479453, + "grad_norm": 1.7407734394073486, + "learning_rate": 1.1313868613138687e-06, + "loss": 2.3389, + "step": 78 + }, + { + "epoch": 0.08657534246575342, + "grad_norm": 1.9932411909103394, + "learning_rate": 1.1496350364963505e-06, + "loss": 2.3601, + "step": 79 + }, + { + "epoch": 0.08767123287671233, + "grad_norm": 1.8244212865829468, + "learning_rate": 1.1678832116788322e-06, + "loss": 2.354, + "step": 80 + }, + { + "epoch": 0.08767123287671233, + "eval_loss": 2.4792213439941406, + "eval_runtime": 37.7188, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 80 + }, + { + "epoch": 0.08876712328767124, + "grad_norm": 1.6612335443496704, + "learning_rate": 1.1861313868613138e-06, + "loss": 2.3719, + "step": 81 + }, + { + "epoch": 0.08986301369863013, + "grad_norm": 1.7231117486953735, + "learning_rate": 1.2043795620437959e-06, + "loss": 2.4127, + "step": 82 + }, + { + "epoch": 0.09095890410958904, + "grad_norm": 1.6009923219680786, + "learning_rate": 1.2226277372262775e-06, + "loss": 2.37, + "step": 83 + }, + { + "epoch": 0.09205479452054795, + "grad_norm": 2.51488995552063, + "learning_rate": 1.2408759124087592e-06, + "loss": 2.3874, + "step": 84 + }, + { + "epoch": 0.09315068493150686, + "grad_norm": 2.2683658599853516, + "learning_rate": 1.259124087591241e-06, + "loss": 2.3787, + "step": 85 + }, + { + "epoch": 0.09424657534246575, + "grad_norm": 1.7536020278930664, + "learning_rate": 1.2773722627737229e-06, + "loss": 2.3383, + "step": 86 + }, + { + "epoch": 0.09534246575342466, + "grad_norm": 1.9104911088943481, + "learning_rate": 1.2956204379562045e-06, + "loss": 2.3597, + "step": 87 + }, + { + "epoch": 0.09643835616438357, + "grad_norm": 1.9581283330917358, + "learning_rate": 1.3138686131386864e-06, + "loss": 2.3134, + "step": 88 + }, + { + "epoch": 0.09753424657534246, + "grad_norm": 2.0857372283935547, + "learning_rate": 1.332116788321168e-06, + "loss": 2.3527, + "step": 89 + }, + { + "epoch": 0.09863013698630137, + "grad_norm": 1.8731858730316162, + "learning_rate": 1.3503649635036497e-06, + "loss": 2.2812, + "step": 90 + }, + { + "epoch": 0.09863013698630137, + "eval_loss": 2.46639347076416, + "eval_runtime": 37.7185, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 90 + }, + { + "epoch": 0.09972602739726028, + "grad_norm": 2.262418746948242, + "learning_rate": 1.3686131386861315e-06, + "loss": 2.3857, + "step": 91 + }, + { + "epoch": 0.10082191780821918, + "grad_norm": 1.8655835390090942, + "learning_rate": 1.3868613138686132e-06, + "loss": 2.3513, + "step": 92 + }, + { + "epoch": 0.10191780821917808, + "grad_norm": 2.1694376468658447, + "learning_rate": 1.405109489051095e-06, + "loss": 2.3804, + "step": 93 + }, + { + "epoch": 0.10301369863013699, + "grad_norm": 1.9875420331954956, + "learning_rate": 1.4233576642335767e-06, + "loss": 2.3432, + "step": 94 + }, + { + "epoch": 0.10410958904109589, + "grad_norm": 2.093919515609741, + "learning_rate": 1.4416058394160583e-06, + "loss": 2.3258, + "step": 95 + }, + { + "epoch": 0.1052054794520548, + "grad_norm": 2.2373430728912354, + "learning_rate": 1.4598540145985402e-06, + "loss": 2.3926, + "step": 96 + }, + { + "epoch": 0.1063013698630137, + "grad_norm": 1.830695629119873, + "learning_rate": 1.478102189781022e-06, + "loss": 2.3596, + "step": 97 + }, + { + "epoch": 0.1073972602739726, + "grad_norm": 1.9554585218429565, + "learning_rate": 1.496350364963504e-06, + "loss": 2.3817, + "step": 98 + }, + { + "epoch": 0.10849315068493151, + "grad_norm": 1.7795532941818237, + "learning_rate": 1.5145985401459856e-06, + "loss": 2.3516, + "step": 99 + }, + { + "epoch": 0.1095890410958904, + "grad_norm": 2.1863558292388916, + "learning_rate": 1.5328467153284674e-06, + "loss": 2.3108, + "step": 100 + }, + { + "epoch": 0.1095890410958904, + "eval_loss": 2.4552853107452393, + "eval_runtime": 37.7096, + "eval_samples_per_second": 26.518, + "eval_steps_per_second": 1.114, + "step": 100 + }, + { + "epoch": 0.11068493150684931, + "grad_norm": 1.9794018268585205, + "learning_rate": 1.551094890510949e-06, + "loss": 2.3047, + "step": 101 + }, + { + "epoch": 0.11178082191780822, + "grad_norm": 2.396454334259033, + "learning_rate": 1.5693430656934307e-06, + "loss": 2.3671, + "step": 102 + }, + { + "epoch": 0.11287671232876713, + "grad_norm": 2.0826210975646973, + "learning_rate": 1.5875912408759126e-06, + "loss": 2.4105, + "step": 103 + }, + { + "epoch": 0.11397260273972602, + "grad_norm": 2.3590970039367676, + "learning_rate": 1.6058394160583942e-06, + "loss": 2.3405, + "step": 104 + }, + { + "epoch": 0.11506849315068493, + "grad_norm": 2.3945844173431396, + "learning_rate": 1.624087591240876e-06, + "loss": 2.2922, + "step": 105 + }, + { + "epoch": 0.11616438356164384, + "grad_norm": 2.185126781463623, + "learning_rate": 1.6423357664233577e-06, + "loss": 2.3399, + "step": 106 + }, + { + "epoch": 0.11726027397260275, + "grad_norm": 2.2687461376190186, + "learning_rate": 1.6605839416058394e-06, + "loss": 2.3547, + "step": 107 + }, + { + "epoch": 0.11835616438356164, + "grad_norm": 1.920499324798584, + "learning_rate": 1.6788321167883212e-06, + "loss": 2.28, + "step": 108 + }, + { + "epoch": 0.11945205479452055, + "grad_norm": 2.127774715423584, + "learning_rate": 1.6970802919708029e-06, + "loss": 2.3409, + "step": 109 + }, + { + "epoch": 0.12054794520547946, + "grad_norm": 1.8677122592926025, + "learning_rate": 1.715328467153285e-06, + "loss": 2.3406, + "step": 110 + }, + { + "epoch": 0.12054794520547946, + "eval_loss": 2.4506173133850098, + "eval_runtime": 37.5008, + "eval_samples_per_second": 26.666, + "eval_steps_per_second": 1.12, + "step": 110 + }, + { + "epoch": 0.12164383561643835, + "grad_norm": 1.9803308248519897, + "learning_rate": 1.7335766423357666e-06, + "loss": 2.3642, + "step": 111 + }, + { + "epoch": 0.12273972602739726, + "grad_norm": 1.7733794450759888, + "learning_rate": 1.7518248175182485e-06, + "loss": 2.3666, + "step": 112 + }, + { + "epoch": 0.12383561643835617, + "grad_norm": 1.7639747858047485, + "learning_rate": 1.77007299270073e-06, + "loss": 2.3926, + "step": 113 + }, + { + "epoch": 0.12493150684931507, + "grad_norm": 1.9172815084457397, + "learning_rate": 1.788321167883212e-06, + "loss": 2.3431, + "step": 114 + }, + { + "epoch": 0.12602739726027398, + "grad_norm": 1.580621361732483, + "learning_rate": 1.8065693430656936e-06, + "loss": 2.2967, + "step": 115 + }, + { + "epoch": 0.12712328767123288, + "grad_norm": 2.0889124870300293, + "learning_rate": 1.8248175182481753e-06, + "loss": 2.331, + "step": 116 + }, + { + "epoch": 0.12821917808219177, + "grad_norm": 1.6659903526306152, + "learning_rate": 1.8430656934306571e-06, + "loss": 2.3181, + "step": 117 + }, + { + "epoch": 0.1293150684931507, + "grad_norm": 2.028651237487793, + "learning_rate": 1.8613138686131388e-06, + "loss": 2.3237, + "step": 118 + }, + { + "epoch": 0.1304109589041096, + "grad_norm": 1.681212306022644, + "learning_rate": 1.8795620437956206e-06, + "loss": 2.32, + "step": 119 + }, + { + "epoch": 0.13150684931506848, + "grad_norm": 1.9986696243286133, + "learning_rate": 1.8978102189781023e-06, + "loss": 2.3425, + "step": 120 + }, + { + "epoch": 0.13150684931506848, + "eval_loss": 2.4408469200134277, + "eval_runtime": 37.7136, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 120 + }, + { + "epoch": 0.1326027397260274, + "grad_norm": 1.832582950592041, + "learning_rate": 1.916058394160584e-06, + "loss": 2.2548, + "step": 121 + }, + { + "epoch": 0.1336986301369863, + "grad_norm": 1.8609144687652588, + "learning_rate": 1.934306569343066e-06, + "loss": 2.3024, + "step": 122 + }, + { + "epoch": 0.13479452054794522, + "grad_norm": 1.8699817657470703, + "learning_rate": 1.952554744525548e-06, + "loss": 2.3044, + "step": 123 + }, + { + "epoch": 0.1358904109589041, + "grad_norm": 1.81354558467865, + "learning_rate": 1.9708029197080293e-06, + "loss": 2.3098, + "step": 124 + }, + { + "epoch": 0.136986301369863, + "grad_norm": 1.7789452075958252, + "learning_rate": 1.989051094890511e-06, + "loss": 2.3623, + "step": 125 + }, + { + "epoch": 0.13808219178082193, + "grad_norm": 1.6322330236434937, + "learning_rate": 2.007299270072993e-06, + "loss": 2.303, + "step": 126 + }, + { + "epoch": 0.13917808219178082, + "grad_norm": 1.8742916584014893, + "learning_rate": 2.025547445255475e-06, + "loss": 2.3183, + "step": 127 + }, + { + "epoch": 0.14027397260273972, + "grad_norm": 1.7790557146072388, + "learning_rate": 2.0437956204379563e-06, + "loss": 2.3168, + "step": 128 + }, + { + "epoch": 0.14136986301369864, + "grad_norm": 1.7624462842941284, + "learning_rate": 2.062043795620438e-06, + "loss": 2.3629, + "step": 129 + }, + { + "epoch": 0.14246575342465753, + "grad_norm": 1.781114101409912, + "learning_rate": 2.08029197080292e-06, + "loss": 2.2813, + "step": 130 + }, + { + "epoch": 0.14246575342465753, + "eval_loss": 2.426335334777832, + "eval_runtime": 37.7146, + "eval_samples_per_second": 26.515, + "eval_steps_per_second": 1.114, + "step": 130 + }, + { + "epoch": 0.14356164383561643, + "grad_norm": 1.769988775253296, + "learning_rate": 2.0985401459854014e-06, + "loss": 2.3396, + "step": 131 + }, + { + "epoch": 0.14465753424657535, + "grad_norm": 2.0640275478363037, + "learning_rate": 2.1167883211678833e-06, + "loss": 2.3688, + "step": 132 + }, + { + "epoch": 0.14575342465753424, + "grad_norm": 1.7884094715118408, + "learning_rate": 2.135036496350365e-06, + "loss": 2.2775, + "step": 133 + }, + { + "epoch": 0.14684931506849316, + "grad_norm": 1.831847071647644, + "learning_rate": 2.1532846715328466e-06, + "loss": 2.2966, + "step": 134 + }, + { + "epoch": 0.14794520547945206, + "grad_norm": 1.9044525623321533, + "learning_rate": 2.171532846715329e-06, + "loss": 2.3359, + "step": 135 + }, + { + "epoch": 0.14904109589041095, + "grad_norm": 1.9899096488952637, + "learning_rate": 2.1897810218978103e-06, + "loss": 2.3099, + "step": 136 + }, + { + "epoch": 0.15013698630136987, + "grad_norm": 1.8570183515548706, + "learning_rate": 2.208029197080292e-06, + "loss": 2.3016, + "step": 137 + }, + { + "epoch": 0.15123287671232877, + "grad_norm": 1.8255733251571655, + "learning_rate": 2.226277372262774e-06, + "loss": 2.335, + "step": 138 + }, + { + "epoch": 0.15232876712328766, + "grad_norm": 1.8759729862213135, + "learning_rate": 2.244525547445256e-06, + "loss": 2.3357, + "step": 139 + }, + { + "epoch": 0.15342465753424658, + "grad_norm": 1.9031389951705933, + "learning_rate": 2.2627737226277373e-06, + "loss": 2.3014, + "step": 140 + }, + { + "epoch": 0.15342465753424658, + "eval_loss": 2.4213266372680664, + "eval_runtime": 37.6757, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 140 + }, + { + "epoch": 0.15452054794520548, + "grad_norm": 2.1044118404388428, + "learning_rate": 2.281021897810219e-06, + "loss": 2.2946, + "step": 141 + }, + { + "epoch": 0.15561643835616437, + "grad_norm": 2.005735158920288, + "learning_rate": 2.299270072992701e-06, + "loss": 2.3132, + "step": 142 + }, + { + "epoch": 0.1567123287671233, + "grad_norm": 1.8821312189102173, + "learning_rate": 2.3175182481751825e-06, + "loss": 2.2816, + "step": 143 + }, + { + "epoch": 0.1578082191780822, + "grad_norm": 1.7866829633712769, + "learning_rate": 2.3357664233576643e-06, + "loss": 2.309, + "step": 144 + }, + { + "epoch": 0.1589041095890411, + "grad_norm": 1.8452801704406738, + "learning_rate": 2.354014598540146e-06, + "loss": 2.3222, + "step": 145 + }, + { + "epoch": 0.16, + "grad_norm": 1.5781638622283936, + "learning_rate": 2.3722627737226276e-06, + "loss": 2.2688, + "step": 146 + }, + { + "epoch": 0.1610958904109589, + "grad_norm": 1.9746729135513306, + "learning_rate": 2.39051094890511e-06, + "loss": 2.2864, + "step": 147 + }, + { + "epoch": 0.16219178082191782, + "grad_norm": 2.0312447547912598, + "learning_rate": 2.4087591240875918e-06, + "loss": 2.2914, + "step": 148 + }, + { + "epoch": 0.1632876712328767, + "grad_norm": 1.675674557685852, + "learning_rate": 2.427007299270073e-06, + "loss": 2.3441, + "step": 149 + }, + { + "epoch": 0.1643835616438356, + "grad_norm": 1.8987798690795898, + "learning_rate": 2.445255474452555e-06, + "loss": 2.2589, + "step": 150 + }, + { + "epoch": 0.1643835616438356, + "eval_loss": 2.4149818420410156, + "eval_runtime": 37.6627, + "eval_samples_per_second": 26.551, + "eval_steps_per_second": 1.115, + "step": 150 + }, + { + "epoch": 0.16547945205479453, + "grad_norm": 2.041752338409424, + "learning_rate": 2.463503649635037e-06, + "loss": 2.2457, + "step": 151 + }, + { + "epoch": 0.16657534246575342, + "grad_norm": 1.7845569849014282, + "learning_rate": 2.4817518248175183e-06, + "loss": 2.3376, + "step": 152 + }, + { + "epoch": 0.16767123287671232, + "grad_norm": 2.489330530166626, + "learning_rate": 2.5e-06, + "loss": 2.258, + "step": 153 + }, + { + "epoch": 0.16876712328767124, + "grad_norm": 2.3793296813964844, + "learning_rate": 2.518248175182482e-06, + "loss": 2.2561, + "step": 154 + }, + { + "epoch": 0.16986301369863013, + "grad_norm": 2.0474460124969482, + "learning_rate": 2.5364963503649635e-06, + "loss": 2.2992, + "step": 155 + }, + { + "epoch": 0.17095890410958905, + "grad_norm": 2.7181849479675293, + "learning_rate": 2.5547445255474458e-06, + "loss": 2.3003, + "step": 156 + }, + { + "epoch": 0.17205479452054795, + "grad_norm": 2.3854947090148926, + "learning_rate": 2.5729927007299272e-06, + "loss": 2.2866, + "step": 157 + }, + { + "epoch": 0.17315068493150684, + "grad_norm": 1.7228457927703857, + "learning_rate": 2.591240875912409e-06, + "loss": 2.3027, + "step": 158 + }, + { + "epoch": 0.17424657534246576, + "grad_norm": 2.3928253650665283, + "learning_rate": 2.6094890510948905e-06, + "loss": 2.3233, + "step": 159 + }, + { + "epoch": 0.17534246575342466, + "grad_norm": 2.136934757232666, + "learning_rate": 2.627737226277373e-06, + "loss": 2.2949, + "step": 160 + }, + { + "epoch": 0.17534246575342466, + "eval_loss": 2.410449266433716, + "eval_runtime": 37.6837, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 160 + }, + { + "epoch": 0.17643835616438355, + "grad_norm": 1.8906159400939941, + "learning_rate": 2.6459854014598542e-06, + "loss": 2.2693, + "step": 161 + }, + { + "epoch": 0.17753424657534247, + "grad_norm": 1.8668863773345947, + "learning_rate": 2.664233576642336e-06, + "loss": 2.2944, + "step": 162 + }, + { + "epoch": 0.17863013698630137, + "grad_norm": 2.319363594055176, + "learning_rate": 2.6824817518248175e-06, + "loss": 2.2791, + "step": 163 + }, + { + "epoch": 0.17972602739726026, + "grad_norm": 2.0869688987731934, + "learning_rate": 2.7007299270072994e-06, + "loss": 2.32, + "step": 164 + }, + { + "epoch": 0.18082191780821918, + "grad_norm": 1.8910208940505981, + "learning_rate": 2.718978102189781e-06, + "loss": 2.3059, + "step": 165 + }, + { + "epoch": 0.18191780821917808, + "grad_norm": 2.157984495162964, + "learning_rate": 2.737226277372263e-06, + "loss": 2.2756, + "step": 166 + }, + { + "epoch": 0.18301369863013697, + "grad_norm": 2.387314558029175, + "learning_rate": 2.755474452554745e-06, + "loss": 2.3597, + "step": 167 + }, + { + "epoch": 0.1841095890410959, + "grad_norm": 2.183333158493042, + "learning_rate": 2.7737226277372264e-06, + "loss": 2.322, + "step": 168 + }, + { + "epoch": 0.1852054794520548, + "grad_norm": 2.158015489578247, + "learning_rate": 2.7919708029197082e-06, + "loss": 2.2663, + "step": 169 + }, + { + "epoch": 0.1863013698630137, + "grad_norm": 1.9594391584396362, + "learning_rate": 2.81021897810219e-06, + "loss": 2.3, + "step": 170 + }, + { + "epoch": 0.1863013698630137, + "eval_loss": 2.411637306213379, + "eval_runtime": 37.6428, + "eval_samples_per_second": 26.565, + "eval_steps_per_second": 1.116, + "step": 170 + }, + { + "epoch": 0.1873972602739726, + "grad_norm": 2.28957462310791, + "learning_rate": 2.828467153284672e-06, + "loss": 2.2258, + "step": 171 + }, + { + "epoch": 0.1884931506849315, + "grad_norm": 1.8085683584213257, + "learning_rate": 2.8467153284671534e-06, + "loss": 2.2739, + "step": 172 + }, + { + "epoch": 0.18958904109589042, + "grad_norm": 1.9662787914276123, + "learning_rate": 2.8649635036496353e-06, + "loss": 2.2398, + "step": 173 + }, + { + "epoch": 0.19068493150684931, + "grad_norm": 2.337609052658081, + "learning_rate": 2.8832116788321167e-06, + "loss": 2.3216, + "step": 174 + }, + { + "epoch": 0.1917808219178082, + "grad_norm": 1.6008479595184326, + "learning_rate": 2.901459854014599e-06, + "loss": 2.2997, + "step": 175 + }, + { + "epoch": 0.19287671232876713, + "grad_norm": 2.177690029144287, + "learning_rate": 2.9197080291970804e-06, + "loss": 2.2893, + "step": 176 + }, + { + "epoch": 0.19397260273972602, + "grad_norm": 1.8746743202209473, + "learning_rate": 2.9379562043795623e-06, + "loss": 2.3303, + "step": 177 + }, + { + "epoch": 0.19506849315068492, + "grad_norm": 1.7862998247146606, + "learning_rate": 2.956204379562044e-06, + "loss": 2.3007, + "step": 178 + }, + { + "epoch": 0.19616438356164384, + "grad_norm": 1.7578285932540894, + "learning_rate": 2.9744525547445256e-06, + "loss": 2.2591, + "step": 179 + }, + { + "epoch": 0.19726027397260273, + "grad_norm": 2.1532702445983887, + "learning_rate": 2.992700729927008e-06, + "loss": 2.2544, + "step": 180 + }, + { + "epoch": 0.19726027397260273, + "eval_loss": 2.3992533683776855, + "eval_runtime": 37.6571, + "eval_samples_per_second": 26.555, + "eval_steps_per_second": 1.115, + "step": 180 + }, + { + "epoch": 0.19835616438356166, + "grad_norm": 2.0168659687042236, + "learning_rate": 3.0109489051094893e-06, + "loss": 2.3108, + "step": 181 + }, + { + "epoch": 0.19945205479452055, + "grad_norm": 1.8336135149002075, + "learning_rate": 3.029197080291971e-06, + "loss": 2.3044, + "step": 182 + }, + { + "epoch": 0.20054794520547944, + "grad_norm": 1.6675384044647217, + "learning_rate": 3.0474452554744526e-06, + "loss": 2.2978, + "step": 183 + }, + { + "epoch": 0.20164383561643837, + "grad_norm": 1.8866558074951172, + "learning_rate": 3.065693430656935e-06, + "loss": 2.3169, + "step": 184 + }, + { + "epoch": 0.20273972602739726, + "grad_norm": 2.0588760375976562, + "learning_rate": 3.0839416058394163e-06, + "loss": 2.3041, + "step": 185 + }, + { + "epoch": 0.20383561643835615, + "grad_norm": 1.9300280809402466, + "learning_rate": 3.102189781021898e-06, + "loss": 2.2494, + "step": 186 + }, + { + "epoch": 0.20493150684931508, + "grad_norm": 1.837496042251587, + "learning_rate": 3.1204379562043796e-06, + "loss": 2.2339, + "step": 187 + }, + { + "epoch": 0.20602739726027397, + "grad_norm": 2.155080795288086, + "learning_rate": 3.1386861313868614e-06, + "loss": 2.324, + "step": 188 + }, + { + "epoch": 0.20712328767123286, + "grad_norm": 1.9562510251998901, + "learning_rate": 3.156934306569343e-06, + "loss": 2.2937, + "step": 189 + }, + { + "epoch": 0.20821917808219179, + "grad_norm": 2.522388219833374, + "learning_rate": 3.175182481751825e-06, + "loss": 2.2683, + "step": 190 + }, + { + "epoch": 0.20821917808219179, + "eval_loss": 2.3936216831207275, + "eval_runtime": 37.6499, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 190 + }, + { + "epoch": 0.20931506849315068, + "grad_norm": 1.7409396171569824, + "learning_rate": 3.193430656934307e-06, + "loss": 2.2833, + "step": 191 + }, + { + "epoch": 0.2104109589041096, + "grad_norm": 2.176687002182007, + "learning_rate": 3.2116788321167884e-06, + "loss": 2.2748, + "step": 192 + }, + { + "epoch": 0.2115068493150685, + "grad_norm": 1.8576542139053345, + "learning_rate": 3.2299270072992707e-06, + "loss": 2.2144, + "step": 193 + }, + { + "epoch": 0.2126027397260274, + "grad_norm": 1.9564870595932007, + "learning_rate": 3.248175182481752e-06, + "loss": 2.2898, + "step": 194 + }, + { + "epoch": 0.2136986301369863, + "grad_norm": 1.896238088607788, + "learning_rate": 3.266423357664234e-06, + "loss": 2.2793, + "step": 195 + }, + { + "epoch": 0.2147945205479452, + "grad_norm": 1.735960841178894, + "learning_rate": 3.2846715328467155e-06, + "loss": 2.2869, + "step": 196 + }, + { + "epoch": 0.2158904109589041, + "grad_norm": 2.018303394317627, + "learning_rate": 3.3029197080291973e-06, + "loss": 2.2775, + "step": 197 + }, + { + "epoch": 0.21698630136986302, + "grad_norm": 1.7448145151138306, + "learning_rate": 3.3211678832116788e-06, + "loss": 2.2736, + "step": 198 + }, + { + "epoch": 0.21808219178082192, + "grad_norm": 1.8245689868927002, + "learning_rate": 3.339416058394161e-06, + "loss": 2.2799, + "step": 199 + }, + { + "epoch": 0.2191780821917808, + "grad_norm": 1.821118950843811, + "learning_rate": 3.3576642335766425e-06, + "loss": 2.2823, + "step": 200 + }, + { + "epoch": 0.2191780821917808, + "eval_loss": 2.3893656730651855, + "eval_runtime": 37.676, + "eval_samples_per_second": 26.542, + "eval_steps_per_second": 1.115, + "step": 200 + }, + { + "epoch": 0.22027397260273973, + "grad_norm": 1.9112446308135986, + "learning_rate": 3.3759124087591243e-06, + "loss": 2.287, + "step": 201 + }, + { + "epoch": 0.22136986301369863, + "grad_norm": 1.6978867053985596, + "learning_rate": 3.3941605839416058e-06, + "loss": 2.2865, + "step": 202 + }, + { + "epoch": 0.22246575342465755, + "grad_norm": 1.9376300573349, + "learning_rate": 3.412408759124088e-06, + "loss": 2.2326, + "step": 203 + }, + { + "epoch": 0.22356164383561644, + "grad_norm": 1.8067563772201538, + "learning_rate": 3.43065693430657e-06, + "loss": 2.2739, + "step": 204 + }, + { + "epoch": 0.22465753424657534, + "grad_norm": 1.8476370573043823, + "learning_rate": 3.4489051094890513e-06, + "loss": 2.2328, + "step": 205 + }, + { + "epoch": 0.22575342465753426, + "grad_norm": 1.927923321723938, + "learning_rate": 3.467153284671533e-06, + "loss": 2.2679, + "step": 206 + }, + { + "epoch": 0.22684931506849315, + "grad_norm": 1.7956453561782837, + "learning_rate": 3.4854014598540146e-06, + "loss": 2.2607, + "step": 207 + }, + { + "epoch": 0.22794520547945205, + "grad_norm": 1.857413649559021, + "learning_rate": 3.503649635036497e-06, + "loss": 2.2854, + "step": 208 + }, + { + "epoch": 0.22904109589041097, + "grad_norm": 1.8384093046188354, + "learning_rate": 3.5218978102189783e-06, + "loss": 2.3021, + "step": 209 + }, + { + "epoch": 0.23013698630136986, + "grad_norm": 1.7346371412277222, + "learning_rate": 3.54014598540146e-06, + "loss": 2.3402, + "step": 210 + }, + { + "epoch": 0.23013698630136986, + "eval_loss": 2.390045642852783, + "eval_runtime": 37.4358, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 210 + }, + { + "epoch": 0.23123287671232876, + "grad_norm": 1.7546677589416504, + "learning_rate": 3.5583941605839416e-06, + "loss": 2.3214, + "step": 211 + }, + { + "epoch": 0.23232876712328768, + "grad_norm": 1.7437881231307983, + "learning_rate": 3.576642335766424e-06, + "loss": 2.2004, + "step": 212 + }, + { + "epoch": 0.23342465753424657, + "grad_norm": 1.9509227275848389, + "learning_rate": 3.5948905109489054e-06, + "loss": 2.2044, + "step": 213 + }, + { + "epoch": 0.2345205479452055, + "grad_norm": 1.6918139457702637, + "learning_rate": 3.6131386861313872e-06, + "loss": 2.2369, + "step": 214 + }, + { + "epoch": 0.2356164383561644, + "grad_norm": 1.9664885997772217, + "learning_rate": 3.6313868613138687e-06, + "loss": 2.254, + "step": 215 + }, + { + "epoch": 0.23671232876712328, + "grad_norm": 1.7942848205566406, + "learning_rate": 3.6496350364963505e-06, + "loss": 2.2296, + "step": 216 + }, + { + "epoch": 0.2378082191780822, + "grad_norm": 1.8610197305679321, + "learning_rate": 3.667883211678833e-06, + "loss": 2.2273, + "step": 217 + }, + { + "epoch": 0.2389041095890411, + "grad_norm": 1.8405956029891968, + "learning_rate": 3.6861313868613142e-06, + "loss": 2.3034, + "step": 218 + }, + { + "epoch": 0.24, + "grad_norm": 1.7151877880096436, + "learning_rate": 3.704379562043796e-06, + "loss": 2.2689, + "step": 219 + }, + { + "epoch": 0.2410958904109589, + "grad_norm": 1.8795795440673828, + "learning_rate": 3.7226277372262775e-06, + "loss": 2.2363, + "step": 220 + }, + { + "epoch": 0.2410958904109589, + "eval_loss": 2.384481430053711, + "eval_runtime": 37.6512, + "eval_samples_per_second": 26.56, + "eval_steps_per_second": 1.116, + "step": 220 + }, + { + "epoch": 0.2421917808219178, + "grad_norm": 1.7494690418243408, + "learning_rate": 3.7408759124087594e-06, + "loss": 2.257, + "step": 221 + }, + { + "epoch": 0.2432876712328767, + "grad_norm": 1.7700639963150024, + "learning_rate": 3.7591240875912412e-06, + "loss": 2.2457, + "step": 222 + }, + { + "epoch": 0.24438356164383562, + "grad_norm": 1.846092700958252, + "learning_rate": 3.777372262773723e-06, + "loss": 2.2425, + "step": 223 + }, + { + "epoch": 0.24547945205479452, + "grad_norm": 1.6699751615524292, + "learning_rate": 3.7956204379562045e-06, + "loss": 2.2734, + "step": 224 + }, + { + "epoch": 0.2465753424657534, + "grad_norm": 1.8978508710861206, + "learning_rate": 3.8138686131386864e-06, + "loss": 2.2435, + "step": 225 + }, + { + "epoch": 0.24767123287671233, + "grad_norm": 1.6387466192245483, + "learning_rate": 3.832116788321168e-06, + "loss": 2.2241, + "step": 226 + }, + { + "epoch": 0.24876712328767123, + "grad_norm": 1.6999541521072388, + "learning_rate": 3.85036496350365e-06, + "loss": 2.2468, + "step": 227 + }, + { + "epoch": 0.24986301369863015, + "grad_norm": 1.6322710514068604, + "learning_rate": 3.868613138686132e-06, + "loss": 2.2642, + "step": 228 + }, + { + "epoch": 0.25095890410958904, + "grad_norm": 1.5633572340011597, + "learning_rate": 3.886861313868613e-06, + "loss": 2.2862, + "step": 229 + }, + { + "epoch": 0.25205479452054796, + "grad_norm": 1.7642648220062256, + "learning_rate": 3.905109489051096e-06, + "loss": 2.2414, + "step": 230 + }, + { + "epoch": 0.25205479452054796, + "eval_loss": 2.3775289058685303, + "eval_runtime": 37.7434, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 230 + }, + { + "epoch": 0.25315068493150683, + "grad_norm": 1.76724374294281, + "learning_rate": 3.923357664233577e-06, + "loss": 2.2976, + "step": 231 + }, + { + "epoch": 0.25424657534246575, + "grad_norm": 1.6049190759658813, + "learning_rate": 3.9416058394160585e-06, + "loss": 2.3034, + "step": 232 + }, + { + "epoch": 0.2553424657534247, + "grad_norm": 1.6591981649398804, + "learning_rate": 3.95985401459854e-06, + "loss": 2.1957, + "step": 233 + }, + { + "epoch": 0.25643835616438354, + "grad_norm": 1.650676965713501, + "learning_rate": 3.978102189781022e-06, + "loss": 2.275, + "step": 234 + }, + { + "epoch": 0.25753424657534246, + "grad_norm": 1.6665960550308228, + "learning_rate": 3.996350364963504e-06, + "loss": 2.1853, + "step": 235 + }, + { + "epoch": 0.2586301369863014, + "grad_norm": 1.5785901546478271, + "learning_rate": 4.014598540145986e-06, + "loss": 2.2399, + "step": 236 + }, + { + "epoch": 0.25972602739726025, + "grad_norm": 1.976162075996399, + "learning_rate": 4.032846715328467e-06, + "loss": 2.2499, + "step": 237 + }, + { + "epoch": 0.2608219178082192, + "grad_norm": 1.642327904701233, + "learning_rate": 4.05109489051095e-06, + "loss": 2.235, + "step": 238 + }, + { + "epoch": 0.2619178082191781, + "grad_norm": 2.0126261711120605, + "learning_rate": 4.069343065693431e-06, + "loss": 2.2466, + "step": 239 + }, + { + "epoch": 0.26301369863013696, + "grad_norm": 1.6737161874771118, + "learning_rate": 4.0875912408759126e-06, + "loss": 2.2984, + "step": 240 + }, + { + "epoch": 0.26301369863013696, + "eval_loss": 2.375901222229004, + "eval_runtime": 37.7072, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 240 + }, + { + "epoch": 0.2641095890410959, + "grad_norm": 2.094257354736328, + "learning_rate": 4.105839416058395e-06, + "loss": 2.251, + "step": 241 + }, + { + "epoch": 0.2652054794520548, + "grad_norm": 1.655393123626709, + "learning_rate": 4.124087591240876e-06, + "loss": 2.2366, + "step": 242 + }, + { + "epoch": 0.2663013698630137, + "grad_norm": 2.5909082889556885, + "learning_rate": 4.1423357664233586e-06, + "loss": 2.2986, + "step": 243 + }, + { + "epoch": 0.2673972602739726, + "grad_norm": 1.8278532028198242, + "learning_rate": 4.16058394160584e-06, + "loss": 2.1977, + "step": 244 + }, + { + "epoch": 0.2684931506849315, + "grad_norm": 1.9404895305633545, + "learning_rate": 4.1788321167883214e-06, + "loss": 2.2578, + "step": 245 + }, + { + "epoch": 0.26958904109589044, + "grad_norm": 1.6525582075119019, + "learning_rate": 4.197080291970803e-06, + "loss": 2.2263, + "step": 246 + }, + { + "epoch": 0.2706849315068493, + "grad_norm": 2.021822929382324, + "learning_rate": 4.215328467153285e-06, + "loss": 2.2326, + "step": 247 + }, + { + "epoch": 0.2717808219178082, + "grad_norm": 1.9136033058166504, + "learning_rate": 4.233576642335767e-06, + "loss": 2.2457, + "step": 248 + }, + { + "epoch": 0.27287671232876715, + "grad_norm": 2.389714002609253, + "learning_rate": 4.251824817518249e-06, + "loss": 2.2256, + "step": 249 + }, + { + "epoch": 0.273972602739726, + "grad_norm": 1.7886695861816406, + "learning_rate": 4.27007299270073e-06, + "loss": 2.2222, + "step": 250 + }, + { + "epoch": 0.273972602739726, + "eval_loss": 2.3729844093322754, + "eval_runtime": 37.7638, + "eval_samples_per_second": 26.48, + "eval_steps_per_second": 1.112, + "step": 250 + }, + { + "epoch": 0.27506849315068493, + "grad_norm": 1.9499074220657349, + "learning_rate": 4.288321167883212e-06, + "loss": 2.2026, + "step": 251 + }, + { + "epoch": 0.27616438356164386, + "grad_norm": 1.834797978401184, + "learning_rate": 4.306569343065693e-06, + "loss": 2.2564, + "step": 252 + }, + { + "epoch": 0.2772602739726027, + "grad_norm": 2.0415048599243164, + "learning_rate": 4.3248175182481755e-06, + "loss": 2.2462, + "step": 253 + }, + { + "epoch": 0.27835616438356164, + "grad_norm": 1.6527886390686035, + "learning_rate": 4.343065693430658e-06, + "loss": 2.2493, + "step": 254 + }, + { + "epoch": 0.27945205479452057, + "grad_norm": 2.0504889488220215, + "learning_rate": 4.361313868613139e-06, + "loss": 2.2734, + "step": 255 + }, + { + "epoch": 0.28054794520547943, + "grad_norm": 2.004772424697876, + "learning_rate": 4.379562043795621e-06, + "loss": 2.2469, + "step": 256 + }, + { + "epoch": 0.28164383561643835, + "grad_norm": 2.1973347663879395, + "learning_rate": 4.397810218978102e-06, + "loss": 2.2479, + "step": 257 + }, + { + "epoch": 0.2827397260273973, + "grad_norm": 1.8222625255584717, + "learning_rate": 4.416058394160584e-06, + "loss": 2.2206, + "step": 258 + }, + { + "epoch": 0.28383561643835614, + "grad_norm": 2.1759912967681885, + "learning_rate": 4.434306569343066e-06, + "loss": 2.261, + "step": 259 + }, + { + "epoch": 0.28493150684931506, + "grad_norm": 1.8344624042510986, + "learning_rate": 4.452554744525548e-06, + "loss": 2.2605, + "step": 260 + }, + { + "epoch": 0.28493150684931506, + "eval_loss": 2.3706893920898438, + "eval_runtime": 37.734, + "eval_samples_per_second": 26.501, + "eval_steps_per_second": 1.113, + "step": 260 + }, + { + "epoch": 0.286027397260274, + "grad_norm": 1.7783704996109009, + "learning_rate": 4.4708029197080295e-06, + "loss": 2.284, + "step": 261 + }, + { + "epoch": 0.28712328767123285, + "grad_norm": 2.101193428039551, + "learning_rate": 4.489051094890512e-06, + "loss": 2.2524, + "step": 262 + }, + { + "epoch": 0.2882191780821918, + "grad_norm": 1.7923569679260254, + "learning_rate": 4.507299270072993e-06, + "loss": 2.2434, + "step": 263 + }, + { + "epoch": 0.2893150684931507, + "grad_norm": 2.170883893966675, + "learning_rate": 4.525547445255475e-06, + "loss": 2.2173, + "step": 264 + }, + { + "epoch": 0.29041095890410956, + "grad_norm": 1.6749192476272583, + "learning_rate": 4.543795620437956e-06, + "loss": 2.283, + "step": 265 + }, + { + "epoch": 0.2915068493150685, + "grad_norm": 2.0484230518341064, + "learning_rate": 4.562043795620438e-06, + "loss": 2.2931, + "step": 266 + }, + { + "epoch": 0.2926027397260274, + "grad_norm": 1.9212749004364014, + "learning_rate": 4.580291970802921e-06, + "loss": 2.2661, + "step": 267 + }, + { + "epoch": 0.2936986301369863, + "grad_norm": 1.6729028224945068, + "learning_rate": 4.598540145985402e-06, + "loss": 2.2269, + "step": 268 + }, + { + "epoch": 0.2947945205479452, + "grad_norm": 2.1190619468688965, + "learning_rate": 4.6167883211678835e-06, + "loss": 2.2278, + "step": 269 + }, + { + "epoch": 0.2958904109589041, + "grad_norm": 1.7444697618484497, + "learning_rate": 4.635036496350365e-06, + "loss": 2.2634, + "step": 270 + }, + { + "epoch": 0.2958904109589041, + "eval_loss": 2.3706092834472656, + "eval_runtime": 37.7391, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 270 + }, + { + "epoch": 0.29698630136986304, + "grad_norm": 1.8921526670455933, + "learning_rate": 4.653284671532847e-06, + "loss": 2.2697, + "step": 271 + }, + { + "epoch": 0.2980821917808219, + "grad_norm": 1.7347339391708374, + "learning_rate": 4.671532846715329e-06, + "loss": 2.2406, + "step": 272 + }, + { + "epoch": 0.2991780821917808, + "grad_norm": 2.1617953777313232, + "learning_rate": 4.689781021897811e-06, + "loss": 2.2891, + "step": 273 + }, + { + "epoch": 0.30027397260273975, + "grad_norm": 1.873384714126587, + "learning_rate": 4.708029197080292e-06, + "loss": 2.2554, + "step": 274 + }, + { + "epoch": 0.3013698630136986, + "grad_norm": 1.801356554031372, + "learning_rate": 4.726277372262774e-06, + "loss": 2.2073, + "step": 275 + }, + { + "epoch": 0.30246575342465754, + "grad_norm": 1.9250633716583252, + "learning_rate": 4.744525547445255e-06, + "loss": 2.2501, + "step": 276 + }, + { + "epoch": 0.30356164383561646, + "grad_norm": 1.8360860347747803, + "learning_rate": 4.7627737226277375e-06, + "loss": 2.2561, + "step": 277 + }, + { + "epoch": 0.3046575342465753, + "grad_norm": 1.8711498975753784, + "learning_rate": 4.78102189781022e-06, + "loss": 2.2247, + "step": 278 + }, + { + "epoch": 0.30575342465753425, + "grad_norm": 1.6604810953140259, + "learning_rate": 4.799270072992701e-06, + "loss": 2.2448, + "step": 279 + }, + { + "epoch": 0.30684931506849317, + "grad_norm": 1.689201831817627, + "learning_rate": 4.8175182481751835e-06, + "loss": 2.2494, + "step": 280 + }, + { + "epoch": 0.30684931506849317, + "eval_loss": 2.3682427406311035, + "eval_runtime": 37.7233, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 280 + }, + { + "epoch": 0.30794520547945203, + "grad_norm": 1.720961093902588, + "learning_rate": 4.835766423357665e-06, + "loss": 2.2013, + "step": 281 + }, + { + "epoch": 0.30904109589041096, + "grad_norm": 1.7863153219223022, + "learning_rate": 4.854014598540146e-06, + "loss": 2.2183, + "step": 282 + }, + { + "epoch": 0.3101369863013699, + "grad_norm": 1.8583060503005981, + "learning_rate": 4.872262773722628e-06, + "loss": 2.2013, + "step": 283 + }, + { + "epoch": 0.31123287671232874, + "grad_norm": 1.722428560256958, + "learning_rate": 4.89051094890511e-06, + "loss": 2.2399, + "step": 284 + }, + { + "epoch": 0.31232876712328766, + "grad_norm": 1.8408993482589722, + "learning_rate": 4.9087591240875915e-06, + "loss": 2.2489, + "step": 285 + }, + { + "epoch": 0.3134246575342466, + "grad_norm": 1.887604832649231, + "learning_rate": 4.927007299270074e-06, + "loss": 2.2145, + "step": 286 + }, + { + "epoch": 0.31452054794520545, + "grad_norm": 1.7112189531326294, + "learning_rate": 4.945255474452555e-06, + "loss": 2.2915, + "step": 287 + }, + { + "epoch": 0.3156164383561644, + "grad_norm": 1.8314064741134644, + "learning_rate": 4.963503649635037e-06, + "loss": 2.2786, + "step": 288 + }, + { + "epoch": 0.3167123287671233, + "grad_norm": 1.7001186609268188, + "learning_rate": 4.981751824817518e-06, + "loss": 2.249, + "step": 289 + }, + { + "epoch": 0.3178082191780822, + "grad_norm": 1.5820505619049072, + "learning_rate": 5e-06, + "loss": 2.254, + "step": 290 + }, + { + "epoch": 0.3178082191780822, + "eval_loss": 2.3624653816223145, + "eval_runtime": 37.7283, + "eval_samples_per_second": 26.505, + "eval_steps_per_second": 1.113, + "step": 290 + }, + { + "epoch": 0.3189041095890411, + "grad_norm": 1.8516303300857544, + "learning_rate": 4.997969130787978e-06, + "loss": 2.1894, + "step": 291 + }, + { + "epoch": 0.32, + "grad_norm": 1.8591808080673218, + "learning_rate": 4.995938261575955e-06, + "loss": 2.2537, + "step": 292 + }, + { + "epoch": 0.32109589041095893, + "grad_norm": 1.6822415590286255, + "learning_rate": 4.993907392363932e-06, + "loss": 2.1978, + "step": 293 + }, + { + "epoch": 0.3221917808219178, + "grad_norm": 2.0779759883880615, + "learning_rate": 4.9918765231519095e-06, + "loss": 2.2286, + "step": 294 + }, + { + "epoch": 0.3232876712328767, + "grad_norm": 1.6937508583068848, + "learning_rate": 4.989845653939887e-06, + "loss": 2.2339, + "step": 295 + }, + { + "epoch": 0.32438356164383564, + "grad_norm": 1.9342437982559204, + "learning_rate": 4.987814784727864e-06, + "loss": 2.2145, + "step": 296 + }, + { + "epoch": 0.3254794520547945, + "grad_norm": 1.773258924484253, + "learning_rate": 4.985783915515841e-06, + "loss": 2.2368, + "step": 297 + }, + { + "epoch": 0.3265753424657534, + "grad_norm": 1.8196488618850708, + "learning_rate": 4.983753046303819e-06, + "loss": 2.2224, + "step": 298 + }, + { + "epoch": 0.32767123287671235, + "grad_norm": 1.6786775588989258, + "learning_rate": 4.981722177091796e-06, + "loss": 2.2439, + "step": 299 + }, + { + "epoch": 0.3287671232876712, + "grad_norm": 1.7761427164077759, + "learning_rate": 4.979691307879773e-06, + "loss": 2.236, + "step": 300 + }, + { + "epoch": 0.3287671232876712, + "eval_loss": 2.3614282608032227, + "eval_runtime": 37.7314, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 300 + }, + { + "epoch": 0.32986301369863014, + "grad_norm": 1.8389159440994263, + "learning_rate": 4.9776604386677504e-06, + "loss": 2.2745, + "step": 301 + }, + { + "epoch": 0.33095890410958906, + "grad_norm": 1.7774744033813477, + "learning_rate": 4.975629569455728e-06, + "loss": 2.2122, + "step": 302 + }, + { + "epoch": 0.3320547945205479, + "grad_norm": 1.7862017154693604, + "learning_rate": 4.973598700243705e-06, + "loss": 2.2097, + "step": 303 + }, + { + "epoch": 0.33315068493150685, + "grad_norm": 1.8851722478866577, + "learning_rate": 4.971567831031681e-06, + "loss": 2.118, + "step": 304 + }, + { + "epoch": 0.33424657534246577, + "grad_norm": 2.068608045578003, + "learning_rate": 4.9695369618196595e-06, + "loss": 2.1884, + "step": 305 + }, + { + "epoch": 0.33534246575342463, + "grad_norm": 1.8885847330093384, + "learning_rate": 4.967506092607636e-06, + "loss": 2.2097, + "step": 306 + }, + { + "epoch": 0.33643835616438356, + "grad_norm": 1.9441415071487427, + "learning_rate": 4.965475223395614e-06, + "loss": 2.2068, + "step": 307 + }, + { + "epoch": 0.3375342465753425, + "grad_norm": 2.313577890396118, + "learning_rate": 4.9634443541835905e-06, + "loss": 2.2355, + "step": 308 + }, + { + "epoch": 0.33863013698630134, + "grad_norm": 1.8542472124099731, + "learning_rate": 4.961413484971569e-06, + "loss": 2.2299, + "step": 309 + }, + { + "epoch": 0.33972602739726027, + "grad_norm": 1.9382838010787964, + "learning_rate": 4.959382615759545e-06, + "loss": 2.2072, + "step": 310 + }, + { + "epoch": 0.33972602739726027, + "eval_loss": 2.361767053604126, + "eval_runtime": 37.4603, + "eval_samples_per_second": 26.695, + "eval_steps_per_second": 1.121, + "step": 310 + }, + { + "epoch": 0.3408219178082192, + "grad_norm": 2.081242322921753, + "learning_rate": 4.957351746547523e-06, + "loss": 2.3167, + "step": 311 + }, + { + "epoch": 0.3419178082191781, + "grad_norm": 1.9109065532684326, + "learning_rate": 4.9553208773355e-06, + "loss": 2.2336, + "step": 312 + }, + { + "epoch": 0.343013698630137, + "grad_norm": 2.124897003173828, + "learning_rate": 4.953290008123478e-06, + "loss": 2.184, + "step": 313 + }, + { + "epoch": 0.3441095890410959, + "grad_norm": 1.8820433616638184, + "learning_rate": 4.951259138911454e-06, + "loss": 2.2836, + "step": 314 + }, + { + "epoch": 0.3452054794520548, + "grad_norm": 2.016204595565796, + "learning_rate": 4.9492282696994315e-06, + "loss": 2.2221, + "step": 315 + }, + { + "epoch": 0.3463013698630137, + "grad_norm": 2.0679776668548584, + "learning_rate": 4.947197400487409e-06, + "loss": 2.2113, + "step": 316 + }, + { + "epoch": 0.3473972602739726, + "grad_norm": 1.8040891885757446, + "learning_rate": 4.945166531275386e-06, + "loss": 2.2292, + "step": 317 + }, + { + "epoch": 0.34849315068493153, + "grad_norm": 2.0315871238708496, + "learning_rate": 4.943135662063363e-06, + "loss": 2.2657, + "step": 318 + }, + { + "epoch": 0.3495890410958904, + "grad_norm": 1.71120023727417, + "learning_rate": 4.9411047928513406e-06, + "loss": 2.2517, + "step": 319 + }, + { + "epoch": 0.3506849315068493, + "grad_norm": 2.0443193912506104, + "learning_rate": 4.939073923639318e-06, + "loss": 2.2105, + "step": 320 + }, + { + "epoch": 0.3506849315068493, + "eval_loss": 2.3582265377044678, + "eval_runtime": 37.6749, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 320 + }, + { + "epoch": 0.35178082191780824, + "grad_norm": 1.9363837242126465, + "learning_rate": 4.937043054427295e-06, + "loss": 2.1944, + "step": 321 + }, + { + "epoch": 0.3528767123287671, + "grad_norm": 2.0878167152404785, + "learning_rate": 4.935012185215272e-06, + "loss": 2.2477, + "step": 322 + }, + { + "epoch": 0.353972602739726, + "grad_norm": 1.9059010744094849, + "learning_rate": 4.93298131600325e-06, + "loss": 2.2278, + "step": 323 + }, + { + "epoch": 0.35506849315068495, + "grad_norm": 2.220851421356201, + "learning_rate": 4.930950446791227e-06, + "loss": 2.2391, + "step": 324 + }, + { + "epoch": 0.3561643835616438, + "grad_norm": 1.7555570602416992, + "learning_rate": 4.928919577579204e-06, + "loss": 2.2206, + "step": 325 + }, + { + "epoch": 0.35726027397260274, + "grad_norm": 2.1546502113342285, + "learning_rate": 4.9268887083671815e-06, + "loss": 2.2216, + "step": 326 + }, + { + "epoch": 0.35835616438356166, + "grad_norm": 1.9688911437988281, + "learning_rate": 4.924857839155159e-06, + "loss": 2.2423, + "step": 327 + }, + { + "epoch": 0.3594520547945205, + "grad_norm": 1.983587622642517, + "learning_rate": 4.922826969943136e-06, + "loss": 2.1738, + "step": 328 + }, + { + "epoch": 0.36054794520547945, + "grad_norm": 2.1990034580230713, + "learning_rate": 4.920796100731113e-06, + "loss": 2.2341, + "step": 329 + }, + { + "epoch": 0.36164383561643837, + "grad_norm": 1.9690128564834595, + "learning_rate": 4.918765231519091e-06, + "loss": 2.1914, + "step": 330 + }, + { + "epoch": 0.36164383561643837, + "eval_loss": 2.36171817779541, + "eval_runtime": 37.727, + "eval_samples_per_second": 26.506, + "eval_steps_per_second": 1.113, + "step": 330 + }, + { + "epoch": 0.36273972602739724, + "grad_norm": 2.5950212478637695, + "learning_rate": 4.916734362307068e-06, + "loss": 2.1921, + "step": 331 + }, + { + "epoch": 0.36383561643835616, + "grad_norm": 2.023130416870117, + "learning_rate": 4.914703493095045e-06, + "loss": 2.1932, + "step": 332 + }, + { + "epoch": 0.3649315068493151, + "grad_norm": 2.2486252784729004, + "learning_rate": 4.912672623883022e-06, + "loss": 2.2174, + "step": 333 + }, + { + "epoch": 0.36602739726027395, + "grad_norm": 1.723677635192871, + "learning_rate": 4.910641754671e-06, + "loss": 2.2173, + "step": 334 + }, + { + "epoch": 0.36712328767123287, + "grad_norm": 1.94509756565094, + "learning_rate": 4.908610885458977e-06, + "loss": 2.2539, + "step": 335 + }, + { + "epoch": 0.3682191780821918, + "grad_norm": 1.6934322118759155, + "learning_rate": 4.906580016246954e-06, + "loss": 2.2345, + "step": 336 + }, + { + "epoch": 0.3693150684931507, + "grad_norm": 1.6882652044296265, + "learning_rate": 4.9045491470349315e-06, + "loss": 2.231, + "step": 337 + }, + { + "epoch": 0.3704109589041096, + "grad_norm": 1.6730620861053467, + "learning_rate": 4.902518277822909e-06, + "loss": 2.2394, + "step": 338 + }, + { + "epoch": 0.3715068493150685, + "grad_norm": 1.7261624336242676, + "learning_rate": 4.900487408610886e-06, + "loss": 2.2524, + "step": 339 + }, + { + "epoch": 0.3726027397260274, + "grad_norm": 1.8310920000076294, + "learning_rate": 4.898456539398863e-06, + "loss": 2.3004, + "step": 340 + }, + { + "epoch": 0.3726027397260274, + "eval_loss": 2.3536295890808105, + "eval_runtime": 37.7461, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 340 + }, + { + "epoch": 0.3736986301369863, + "grad_norm": 1.7025809288024902, + "learning_rate": 4.89642567018684e-06, + "loss": 2.1923, + "step": 341 + }, + { + "epoch": 0.3747945205479452, + "grad_norm": 1.7285082340240479, + "learning_rate": 4.894394800974818e-06, + "loss": 2.241, + "step": 342 + }, + { + "epoch": 0.37589041095890413, + "grad_norm": 1.7589807510375977, + "learning_rate": 4.892363931762794e-06, + "loss": 2.2019, + "step": 343 + }, + { + "epoch": 0.376986301369863, + "grad_norm": 1.6664931774139404, + "learning_rate": 4.8903330625507725e-06, + "loss": 2.1822, + "step": 344 + }, + { + "epoch": 0.3780821917808219, + "grad_norm": 1.8308027982711792, + "learning_rate": 4.888302193338749e-06, + "loss": 2.2651, + "step": 345 + }, + { + "epoch": 0.37917808219178084, + "grad_norm": 1.6426894664764404, + "learning_rate": 4.886271324126727e-06, + "loss": 2.2514, + "step": 346 + }, + { + "epoch": 0.3802739726027397, + "grad_norm": 1.9765205383300781, + "learning_rate": 4.8842404549147034e-06, + "loss": 2.226, + "step": 347 + }, + { + "epoch": 0.38136986301369863, + "grad_norm": 1.7872203588485718, + "learning_rate": 4.8822095857026816e-06, + "loss": 2.18, + "step": 348 + }, + { + "epoch": 0.38246575342465755, + "grad_norm": 2.0645813941955566, + "learning_rate": 4.880178716490658e-06, + "loss": 2.2023, + "step": 349 + }, + { + "epoch": 0.3835616438356164, + "grad_norm": 1.9823166131973267, + "learning_rate": 4.878147847278636e-06, + "loss": 2.1801, + "step": 350 + }, + { + "epoch": 0.3835616438356164, + "eval_loss": 2.353776454925537, + "eval_runtime": 37.771, + "eval_samples_per_second": 26.475, + "eval_steps_per_second": 1.112, + "step": 350 + }, + { + "epoch": 0.38465753424657534, + "grad_norm": 1.831152319908142, + "learning_rate": 4.8761169780666125e-06, + "loss": 2.1858, + "step": 351 + }, + { + "epoch": 0.38575342465753426, + "grad_norm": 1.9321790933609009, + "learning_rate": 4.874086108854591e-06, + "loss": 2.1638, + "step": 352 + }, + { + "epoch": 0.3868493150684931, + "grad_norm": 1.9097596406936646, + "learning_rate": 4.872055239642567e-06, + "loss": 2.2206, + "step": 353 + }, + { + "epoch": 0.38794520547945205, + "grad_norm": 1.8821815252304077, + "learning_rate": 4.870024370430544e-06, + "loss": 2.2338, + "step": 354 + }, + { + "epoch": 0.38904109589041097, + "grad_norm": 1.7959749698638916, + "learning_rate": 4.867993501218522e-06, + "loss": 2.1882, + "step": 355 + }, + { + "epoch": 0.39013698630136984, + "grad_norm": 1.625320315361023, + "learning_rate": 4.865962632006499e-06, + "loss": 2.2137, + "step": 356 + }, + { + "epoch": 0.39123287671232876, + "grad_norm": 1.6914536952972412, + "learning_rate": 4.863931762794476e-06, + "loss": 2.2374, + "step": 357 + }, + { + "epoch": 0.3923287671232877, + "grad_norm": 2.0801584720611572, + "learning_rate": 4.8619008935824535e-06, + "loss": 2.2517, + "step": 358 + }, + { + "epoch": 0.3934246575342466, + "grad_norm": 1.6874804496765137, + "learning_rate": 4.859870024370431e-06, + "loss": 2.2621, + "step": 359 + }, + { + "epoch": 0.39452054794520547, + "grad_norm": 1.9347832202911377, + "learning_rate": 4.857839155158408e-06, + "loss": 2.2326, + "step": 360 + }, + { + "epoch": 0.39452054794520547, + "eval_loss": 2.3468236923217773, + "eval_runtime": 37.7184, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.114, + "step": 360 + }, + { + "epoch": 0.3956164383561644, + "grad_norm": 1.6798334121704102, + "learning_rate": 4.855808285946385e-06, + "loss": 2.2222, + "step": 361 + }, + { + "epoch": 0.3967123287671233, + "grad_norm": 1.8993884325027466, + "learning_rate": 4.8537774167343626e-06, + "loss": 2.2178, + "step": 362 + }, + { + "epoch": 0.3978082191780822, + "grad_norm": 1.697892189025879, + "learning_rate": 4.85174654752234e-06, + "loss": 2.2217, + "step": 363 + }, + { + "epoch": 0.3989041095890411, + "grad_norm": 1.8484156131744385, + "learning_rate": 4.849715678310317e-06, + "loss": 2.2069, + "step": 364 + }, + { + "epoch": 0.4, + "grad_norm": 1.8125075101852417, + "learning_rate": 4.847684809098294e-06, + "loss": 2.2285, + "step": 365 + }, + { + "epoch": 0.4010958904109589, + "grad_norm": 1.8215030431747437, + "learning_rate": 4.845653939886272e-06, + "loss": 2.2029, + "step": 366 + }, + { + "epoch": 0.4021917808219178, + "grad_norm": 1.777129054069519, + "learning_rate": 4.843623070674249e-06, + "loss": 2.2512, + "step": 367 + }, + { + "epoch": 0.40328767123287673, + "grad_norm": 1.7327351570129395, + "learning_rate": 4.841592201462226e-06, + "loss": 2.1767, + "step": 368 + }, + { + "epoch": 0.4043835616438356, + "grad_norm": 1.7363202571868896, + "learning_rate": 4.8395613322502035e-06, + "loss": 2.2174, + "step": 369 + }, + { + "epoch": 0.4054794520547945, + "grad_norm": 1.9183942079544067, + "learning_rate": 4.837530463038181e-06, + "loss": 2.2383, + "step": 370 + }, + { + "epoch": 0.4054794520547945, + "eval_loss": 2.345961332321167, + "eval_runtime": 37.722, + "eval_samples_per_second": 26.51, + "eval_steps_per_second": 1.113, + "step": 370 + }, + { + "epoch": 0.40657534246575344, + "grad_norm": 1.7290767431259155, + "learning_rate": 4.835499593826158e-06, + "loss": 2.22, + "step": 371 + }, + { + "epoch": 0.4076712328767123, + "grad_norm": 1.7892694473266602, + "learning_rate": 4.833468724614135e-06, + "loss": 2.1403, + "step": 372 + }, + { + "epoch": 0.40876712328767123, + "grad_norm": 1.6671900749206543, + "learning_rate": 4.831437855402113e-06, + "loss": 2.1989, + "step": 373 + }, + { + "epoch": 0.40986301369863015, + "grad_norm": 1.848751425743103, + "learning_rate": 4.82940698619009e-06, + "loss": 2.245, + "step": 374 + }, + { + "epoch": 0.410958904109589, + "grad_norm": 1.6957710981369019, + "learning_rate": 4.827376116978067e-06, + "loss": 2.2239, + "step": 375 + }, + { + "epoch": 0.41205479452054794, + "grad_norm": 1.650583267211914, + "learning_rate": 4.8253452477660444e-06, + "loss": 2.2101, + "step": 376 + }, + { + "epoch": 0.41315068493150686, + "grad_norm": 1.6510974168777466, + "learning_rate": 4.823314378554022e-06, + "loss": 2.204, + "step": 377 + }, + { + "epoch": 0.41424657534246573, + "grad_norm": 1.9469026327133179, + "learning_rate": 4.821283509341998e-06, + "loss": 2.221, + "step": 378 + }, + { + "epoch": 0.41534246575342465, + "grad_norm": 1.6031194925308228, + "learning_rate": 4.819252640129976e-06, + "loss": 2.1926, + "step": 379 + }, + { + "epoch": 0.41643835616438357, + "grad_norm": 2.120325803756714, + "learning_rate": 4.817221770917953e-06, + "loss": 2.2166, + "step": 380 + }, + { + "epoch": 0.41643835616438357, + "eval_loss": 2.3419017791748047, + "eval_runtime": 37.6854, + "eval_samples_per_second": 26.535, + "eval_steps_per_second": 1.114, + "step": 380 + }, + { + "epoch": 0.41753424657534244, + "grad_norm": 1.5533722639083862, + "learning_rate": 4.815190901705931e-06, + "loss": 2.1981, + "step": 381 + }, + { + "epoch": 0.41863013698630136, + "grad_norm": 1.9650007486343384, + "learning_rate": 4.813160032493907e-06, + "loss": 2.2319, + "step": 382 + }, + { + "epoch": 0.4197260273972603, + "grad_norm": 1.6874849796295166, + "learning_rate": 4.811129163281885e-06, + "loss": 2.2263, + "step": 383 + }, + { + "epoch": 0.4208219178082192, + "grad_norm": 1.787996768951416, + "learning_rate": 4.809098294069862e-06, + "loss": 2.2546, + "step": 384 + }, + { + "epoch": 0.42191780821917807, + "grad_norm": 1.7853440046310425, + "learning_rate": 4.80706742485784e-06, + "loss": 2.1994, + "step": 385 + }, + { + "epoch": 0.423013698630137, + "grad_norm": 1.7882641553878784, + "learning_rate": 4.805036555645816e-06, + "loss": 2.2449, + "step": 386 + }, + { + "epoch": 0.4241095890410959, + "grad_norm": 1.8206003904342651, + "learning_rate": 4.8030056864337945e-06, + "loss": 2.2144, + "step": 387 + }, + { + "epoch": 0.4252054794520548, + "grad_norm": 1.7012749910354614, + "learning_rate": 4.800974817221771e-06, + "loss": 2.2122, + "step": 388 + }, + { + "epoch": 0.4263013698630137, + "grad_norm": 1.8711670637130737, + "learning_rate": 4.798943948009749e-06, + "loss": 2.2132, + "step": 389 + }, + { + "epoch": 0.4273972602739726, + "grad_norm": 1.6880611181259155, + "learning_rate": 4.7969130787977254e-06, + "loss": 2.2336, + "step": 390 + }, + { + "epoch": 0.4273972602739726, + "eval_loss": 2.340393543243408, + "eval_runtime": 37.6885, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 390 + }, + { + "epoch": 0.4284931506849315, + "grad_norm": 2.0543065071105957, + "learning_rate": 4.794882209585703e-06, + "loss": 2.1899, + "step": 391 + }, + { + "epoch": 0.4295890410958904, + "grad_norm": 1.5774480104446411, + "learning_rate": 4.79285134037368e-06, + "loss": 2.1982, + "step": 392 + }, + { + "epoch": 0.43068493150684933, + "grad_norm": 1.7154620885849, + "learning_rate": 4.790820471161657e-06, + "loss": 2.2096, + "step": 393 + }, + { + "epoch": 0.4317808219178082, + "grad_norm": 1.8200151920318604, + "learning_rate": 4.7887896019496345e-06, + "loss": 2.2391, + "step": 394 + }, + { + "epoch": 0.4328767123287671, + "grad_norm": 1.6368298530578613, + "learning_rate": 4.786758732737612e-06, + "loss": 2.1855, + "step": 395 + }, + { + "epoch": 0.43397260273972604, + "grad_norm": 1.6851340532302856, + "learning_rate": 4.784727863525589e-06, + "loss": 2.2114, + "step": 396 + }, + { + "epoch": 0.4350684931506849, + "grad_norm": 1.8293818235397339, + "learning_rate": 4.782696994313566e-06, + "loss": 2.1636, + "step": 397 + }, + { + "epoch": 0.43616438356164383, + "grad_norm": 1.6213513612747192, + "learning_rate": 4.780666125101544e-06, + "loss": 2.1884, + "step": 398 + }, + { + "epoch": 0.43726027397260275, + "grad_norm": 1.7791825532913208, + "learning_rate": 4.778635255889521e-06, + "loss": 2.1977, + "step": 399 + }, + { + "epoch": 0.4383561643835616, + "grad_norm": 1.6627994775772095, + "learning_rate": 4.776604386677498e-06, + "loss": 2.2444, + "step": 400 + }, + { + "epoch": 0.4383561643835616, + "eval_loss": 2.3396637439727783, + "eval_runtime": 37.7069, + "eval_samples_per_second": 26.52, + "eval_steps_per_second": 1.114, + "step": 400 + }, + { + "epoch": 0.43945205479452054, + "grad_norm": 1.7597277164459229, + "learning_rate": 4.7745735174654755e-06, + "loss": 2.2369, + "step": 401 + }, + { + "epoch": 0.44054794520547946, + "grad_norm": 1.7754322290420532, + "learning_rate": 4.772542648253453e-06, + "loss": 2.1859, + "step": 402 + }, + { + "epoch": 0.44164383561643833, + "grad_norm": 1.7897601127624512, + "learning_rate": 4.77051177904143e-06, + "loss": 2.2245, + "step": 403 + }, + { + "epoch": 0.44273972602739725, + "grad_norm": 1.7294282913208008, + "learning_rate": 4.768480909829407e-06, + "loss": 2.1998, + "step": 404 + }, + { + "epoch": 0.4438356164383562, + "grad_norm": 1.8406974077224731, + "learning_rate": 4.766450040617385e-06, + "loss": 2.2353, + "step": 405 + }, + { + "epoch": 0.4449315068493151, + "grad_norm": 1.849493384361267, + "learning_rate": 4.764419171405362e-06, + "loss": 2.2165, + "step": 406 + }, + { + "epoch": 0.44602739726027396, + "grad_norm": 1.5568877458572388, + "learning_rate": 4.762388302193339e-06, + "loss": 2.1989, + "step": 407 + }, + { + "epoch": 0.4471232876712329, + "grad_norm": 1.8816689252853394, + "learning_rate": 4.760357432981316e-06, + "loss": 2.2062, + "step": 408 + }, + { + "epoch": 0.4482191780821918, + "grad_norm": 1.7648332118988037, + "learning_rate": 4.758326563769294e-06, + "loss": 2.2204, + "step": 409 + }, + { + "epoch": 0.44931506849315067, + "grad_norm": 1.753944754600525, + "learning_rate": 4.756295694557271e-06, + "loss": 2.1748, + "step": 410 + }, + { + "epoch": 0.44931506849315067, + "eval_loss": 2.339653253555298, + "eval_runtime": 37.4438, + "eval_samples_per_second": 26.707, + "eval_steps_per_second": 1.122, + "step": 410 + }, + { + "epoch": 0.4504109589041096, + "grad_norm": 1.7343329191207886, + "learning_rate": 4.754264825345248e-06, + "loss": 2.1843, + "step": 411 + }, + { + "epoch": 0.4515068493150685, + "grad_norm": 1.8125407695770264, + "learning_rate": 4.7522339561332255e-06, + "loss": 2.1492, + "step": 412 + }, + { + "epoch": 0.4526027397260274, + "grad_norm": 1.5255184173583984, + "learning_rate": 4.750203086921203e-06, + "loss": 2.2178, + "step": 413 + }, + { + "epoch": 0.4536986301369863, + "grad_norm": 1.6692545413970947, + "learning_rate": 4.74817221770918e-06, + "loss": 2.2516, + "step": 414 + }, + { + "epoch": 0.4547945205479452, + "grad_norm": 1.7804819345474243, + "learning_rate": 4.7461413484971565e-06, + "loss": 2.1592, + "step": 415 + }, + { + "epoch": 0.4558904109589041, + "grad_norm": 1.8048933744430542, + "learning_rate": 4.744110479285135e-06, + "loss": 2.1869, + "step": 416 + }, + { + "epoch": 0.456986301369863, + "grad_norm": 1.6439207792282104, + "learning_rate": 4.742079610073111e-06, + "loss": 2.1995, + "step": 417 + }, + { + "epoch": 0.45808219178082193, + "grad_norm": 1.7375173568725586, + "learning_rate": 4.740048740861089e-06, + "loss": 2.2067, + "step": 418 + }, + { + "epoch": 0.4591780821917808, + "grad_norm": 1.9471659660339355, + "learning_rate": 4.738017871649066e-06, + "loss": 2.2513, + "step": 419 + }, + { + "epoch": 0.4602739726027397, + "grad_norm": 1.6524564027786255, + "learning_rate": 4.735987002437044e-06, + "loss": 2.2643, + "step": 420 + }, + { + "epoch": 0.4602739726027397, + "eval_loss": 2.337153434753418, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 420 + }, + { + "epoch": 0.46136986301369864, + "grad_norm": 1.8574563264846802, + "learning_rate": 4.73395613322502e-06, + "loss": 2.2373, + "step": 421 + }, + { + "epoch": 0.4624657534246575, + "grad_norm": 1.836464524269104, + "learning_rate": 4.731925264012998e-06, + "loss": 2.1769, + "step": 422 + }, + { + "epoch": 0.46356164383561643, + "grad_norm": 1.7012423276901245, + "learning_rate": 4.729894394800975e-06, + "loss": 2.1494, + "step": 423 + }, + { + "epoch": 0.46465753424657535, + "grad_norm": 1.953180193901062, + "learning_rate": 4.727863525588953e-06, + "loss": 2.186, + "step": 424 + }, + { + "epoch": 0.4657534246575342, + "grad_norm": 1.8468470573425293, + "learning_rate": 4.725832656376929e-06, + "loss": 2.2279, + "step": 425 + }, + { + "epoch": 0.46684931506849314, + "grad_norm": 2.005077362060547, + "learning_rate": 4.723801787164907e-06, + "loss": 2.1798, + "step": 426 + }, + { + "epoch": 0.46794520547945206, + "grad_norm": 1.8246952295303345, + "learning_rate": 4.721770917952885e-06, + "loss": 2.1916, + "step": 427 + }, + { + "epoch": 0.469041095890411, + "grad_norm": 1.8558839559555054, + "learning_rate": 4.719740048740861e-06, + "loss": 2.2237, + "step": 428 + }, + { + "epoch": 0.47013698630136985, + "grad_norm": 1.7670999765396118, + "learning_rate": 4.717709179528839e-06, + "loss": 2.2179, + "step": 429 + }, + { + "epoch": 0.4712328767123288, + "grad_norm": 1.8544219732284546, + "learning_rate": 4.715678310316816e-06, + "loss": 2.1811, + "step": 430 + }, + { + "epoch": 0.4712328767123288, + "eval_loss": 2.333791732788086, + "eval_runtime": 37.7254, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 430 + }, + { + "epoch": 0.4723287671232877, + "grad_norm": 1.5792925357818604, + "learning_rate": 4.713647441104794e-06, + "loss": 2.1749, + "step": 431 + }, + { + "epoch": 0.47342465753424656, + "grad_norm": 1.5678222179412842, + "learning_rate": 4.71161657189277e-06, + "loss": 2.2294, + "step": 432 + }, + { + "epoch": 0.4745205479452055, + "grad_norm": 1.9467812776565552, + "learning_rate": 4.709585702680748e-06, + "loss": 2.2684, + "step": 433 + }, + { + "epoch": 0.4756164383561644, + "grad_norm": 1.6647964715957642, + "learning_rate": 4.707554833468725e-06, + "loss": 2.1722, + "step": 434 + }, + { + "epoch": 0.4767123287671233, + "grad_norm": 1.7037622928619385, + "learning_rate": 4.705523964256703e-06, + "loss": 2.1915, + "step": 435 + }, + { + "epoch": 0.4778082191780822, + "grad_norm": 1.5857148170471191, + "learning_rate": 4.703493095044679e-06, + "loss": 2.1466, + "step": 436 + }, + { + "epoch": 0.4789041095890411, + "grad_norm": 1.5792173147201538, + "learning_rate": 4.701462225832657e-06, + "loss": 2.2472, + "step": 437 + }, + { + "epoch": 0.48, + "grad_norm": 1.5998480319976807, + "learning_rate": 4.699431356620634e-06, + "loss": 2.2129, + "step": 438 + }, + { + "epoch": 0.4810958904109589, + "grad_norm": 1.6728194952011108, + "learning_rate": 4.697400487408612e-06, + "loss": 2.2057, + "step": 439 + }, + { + "epoch": 0.4821917808219178, + "grad_norm": 1.5402852296829224, + "learning_rate": 4.695369618196588e-06, + "loss": 2.1903, + "step": 440 + }, + { + "epoch": 0.4821917808219178, + "eval_loss": 2.3325934410095215, + "eval_runtime": 37.7463, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 440 + }, + { + "epoch": 0.4832876712328767, + "grad_norm": 1.5890164375305176, + "learning_rate": 4.693338748984566e-06, + "loss": 2.1584, + "step": 441 + }, + { + "epoch": 0.4843835616438356, + "grad_norm": 1.805923581123352, + "learning_rate": 4.691307879772543e-06, + "loss": 2.2263, + "step": 442 + }, + { + "epoch": 0.48547945205479454, + "grad_norm": 1.5932700634002686, + "learning_rate": 4.68927701056052e-06, + "loss": 2.2109, + "step": 443 + }, + { + "epoch": 0.4865753424657534, + "grad_norm": 1.5611685514450073, + "learning_rate": 4.6872461413484975e-06, + "loss": 2.192, + "step": 444 + }, + { + "epoch": 0.4876712328767123, + "grad_norm": 1.6976194381713867, + "learning_rate": 4.685215272136475e-06, + "loss": 2.1998, + "step": 445 + }, + { + "epoch": 0.48876712328767125, + "grad_norm": 1.7149630784988403, + "learning_rate": 4.683184402924452e-06, + "loss": 2.2599, + "step": 446 + }, + { + "epoch": 0.4898630136986301, + "grad_norm": 1.7597423791885376, + "learning_rate": 4.681153533712429e-06, + "loss": 2.1765, + "step": 447 + }, + { + "epoch": 0.49095890410958903, + "grad_norm": 1.539652943611145, + "learning_rate": 4.679122664500407e-06, + "loss": 2.1635, + "step": 448 + }, + { + "epoch": 0.49205479452054796, + "grad_norm": 1.690402865409851, + "learning_rate": 4.677091795288384e-06, + "loss": 2.2417, + "step": 449 + }, + { + "epoch": 0.4931506849315068, + "grad_norm": 1.7267673015594482, + "learning_rate": 4.675060926076361e-06, + "loss": 2.166, + "step": 450 + }, + { + "epoch": 0.4931506849315068, + "eval_loss": 2.331728935241699, + "eval_runtime": 37.7133, + "eval_samples_per_second": 26.516, + "eval_steps_per_second": 1.114, + "step": 450 + }, + { + "epoch": 0.49424657534246574, + "grad_norm": 1.6383073329925537, + "learning_rate": 4.673030056864338e-06, + "loss": 2.1624, + "step": 451 + }, + { + "epoch": 0.49534246575342467, + "grad_norm": 1.5364140272140503, + "learning_rate": 4.670999187652316e-06, + "loss": 2.1696, + "step": 452 + }, + { + "epoch": 0.4964383561643836, + "grad_norm": 1.591477870941162, + "learning_rate": 4.668968318440293e-06, + "loss": 2.2306, + "step": 453 + }, + { + "epoch": 0.49753424657534245, + "grad_norm": 1.8121449947357178, + "learning_rate": 4.66693744922827e-06, + "loss": 2.1449, + "step": 454 + }, + { + "epoch": 0.4986301369863014, + "grad_norm": 1.515539288520813, + "learning_rate": 4.6649065800162475e-06, + "loss": 2.2108, + "step": 455 + }, + { + "epoch": 0.4997260273972603, + "grad_norm": 1.6521453857421875, + "learning_rate": 4.662875710804225e-06, + "loss": 2.1398, + "step": 456 + }, + { + "epoch": 0.5008219178082192, + "grad_norm": 1.7074251174926758, + "learning_rate": 4.660844841592202e-06, + "loss": 2.2342, + "step": 457 + }, + { + "epoch": 0.5019178082191781, + "grad_norm": 1.7688119411468506, + "learning_rate": 4.658813972380179e-06, + "loss": 2.1287, + "step": 458 + }, + { + "epoch": 0.503013698630137, + "grad_norm": 1.7170679569244385, + "learning_rate": 4.656783103168157e-06, + "loss": 2.2143, + "step": 459 + }, + { + "epoch": 0.5041095890410959, + "grad_norm": 1.7401350736618042, + "learning_rate": 4.654752233956134e-06, + "loss": 2.1855, + "step": 460 + }, + { + "epoch": 0.5041095890410959, + "eval_loss": 2.335416555404663, + "eval_runtime": 37.7115, + "eval_samples_per_second": 26.517, + "eval_steps_per_second": 1.114, + "step": 460 + }, + { + "epoch": 0.5052054794520547, + "grad_norm": 1.6344956159591675, + "learning_rate": 4.652721364744111e-06, + "loss": 2.2205, + "step": 461 + }, + { + "epoch": 0.5063013698630137, + "grad_norm": 1.629184603691101, + "learning_rate": 4.6506904955320885e-06, + "loss": 2.2271, + "step": 462 + }, + { + "epoch": 0.5073972602739726, + "grad_norm": 1.769491195678711, + "learning_rate": 4.648659626320066e-06, + "loss": 2.1945, + "step": 463 + }, + { + "epoch": 0.5084931506849315, + "grad_norm": 1.6513251066207886, + "learning_rate": 4.646628757108043e-06, + "loss": 2.1705, + "step": 464 + }, + { + "epoch": 0.5095890410958904, + "grad_norm": 1.9357609748840332, + "learning_rate": 4.6445978878960194e-06, + "loss": 2.1968, + "step": 465 + }, + { + "epoch": 0.5106849315068493, + "grad_norm": 1.75972318649292, + "learning_rate": 4.6425670186839976e-06, + "loss": 2.2229, + "step": 466 + }, + { + "epoch": 0.5117808219178083, + "grad_norm": 1.686574935913086, + "learning_rate": 4.640536149471974e-06, + "loss": 2.2361, + "step": 467 + }, + { + "epoch": 0.5128767123287671, + "grad_norm": 1.655661702156067, + "learning_rate": 4.638505280259952e-06, + "loss": 2.2064, + "step": 468 + }, + { + "epoch": 0.513972602739726, + "grad_norm": 1.7583094835281372, + "learning_rate": 4.6364744110479285e-06, + "loss": 2.2352, + "step": 469 + }, + { + "epoch": 0.5150684931506849, + "grad_norm": 1.5906383991241455, + "learning_rate": 4.634443541835907e-06, + "loss": 2.187, + "step": 470 + }, + { + "epoch": 0.5150684931506849, + "eval_loss": 2.332923412322998, + "eval_runtime": 37.7496, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 470 + }, + { + "epoch": 0.5161643835616438, + "grad_norm": 1.708019733428955, + "learning_rate": 4.632412672623883e-06, + "loss": 2.1032, + "step": 471 + }, + { + "epoch": 0.5172602739726028, + "grad_norm": 1.6350252628326416, + "learning_rate": 4.630381803411861e-06, + "loss": 2.1976, + "step": 472 + }, + { + "epoch": 0.5183561643835617, + "grad_norm": 1.9037389755249023, + "learning_rate": 4.628350934199838e-06, + "loss": 2.224, + "step": 473 + }, + { + "epoch": 0.5194520547945205, + "grad_norm": 1.7126816511154175, + "learning_rate": 4.626320064987816e-06, + "loss": 2.1925, + "step": 474 + }, + { + "epoch": 0.5205479452054794, + "grad_norm": 1.5828406810760498, + "learning_rate": 4.624289195775792e-06, + "loss": 2.1928, + "step": 475 + }, + { + "epoch": 0.5216438356164383, + "grad_norm": 1.6797102689743042, + "learning_rate": 4.62225832656377e-06, + "loss": 2.1902, + "step": 476 + }, + { + "epoch": 0.5227397260273973, + "grad_norm": 1.6871609687805176, + "learning_rate": 4.620227457351747e-06, + "loss": 2.2024, + "step": 477 + }, + { + "epoch": 0.5238356164383562, + "grad_norm": 1.8785804510116577, + "learning_rate": 4.618196588139724e-06, + "loss": 2.1778, + "step": 478 + }, + { + "epoch": 0.5249315068493151, + "grad_norm": 1.6175488233566284, + "learning_rate": 4.616165718927701e-06, + "loss": 2.2438, + "step": 479 + }, + { + "epoch": 0.5260273972602739, + "grad_norm": 1.7724019289016724, + "learning_rate": 4.6141348497156786e-06, + "loss": 2.2272, + "step": 480 + }, + { + "epoch": 0.5260273972602739, + "eval_loss": 2.328923225402832, + "eval_runtime": 37.7443, + "eval_samples_per_second": 26.494, + "eval_steps_per_second": 1.113, + "step": 480 + }, + { + "epoch": 0.5271232876712328, + "grad_norm": 1.7230926752090454, + "learning_rate": 4.612103980503656e-06, + "loss": 2.2021, + "step": 481 + }, + { + "epoch": 0.5282191780821918, + "grad_norm": 1.7899917364120483, + "learning_rate": 4.610073111291633e-06, + "loss": 2.1854, + "step": 482 + }, + { + "epoch": 0.5293150684931507, + "grad_norm": 1.5864417552947998, + "learning_rate": 4.60804224207961e-06, + "loss": 2.2042, + "step": 483 + }, + { + "epoch": 0.5304109589041096, + "grad_norm": 1.875375509262085, + "learning_rate": 4.606011372867588e-06, + "loss": 2.1844, + "step": 484 + }, + { + "epoch": 0.5315068493150685, + "grad_norm": 1.830597162246704, + "learning_rate": 4.603980503655565e-06, + "loss": 2.1474, + "step": 485 + }, + { + "epoch": 0.5326027397260275, + "grad_norm": 1.8624461889266968, + "learning_rate": 4.601949634443542e-06, + "loss": 2.158, + "step": 486 + }, + { + "epoch": 0.5336986301369863, + "grad_norm": 1.7161756753921509, + "learning_rate": 4.5999187652315195e-06, + "loss": 2.1838, + "step": 487 + }, + { + "epoch": 0.5347945205479452, + "grad_norm": 1.6212890148162842, + "learning_rate": 4.597887896019497e-06, + "loss": 2.1911, + "step": 488 + }, + { + "epoch": 0.5358904109589041, + "grad_norm": 1.8586416244506836, + "learning_rate": 4.595857026807474e-06, + "loss": 2.1794, + "step": 489 + }, + { + "epoch": 0.536986301369863, + "grad_norm": 1.653976559638977, + "learning_rate": 4.593826157595451e-06, + "loss": 2.2428, + "step": 490 + }, + { + "epoch": 0.536986301369863, + "eval_loss": 2.3279106616973877, + "eval_runtime": 37.739, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 490 + }, + { + "epoch": 0.538082191780822, + "grad_norm": 2.054537057876587, + "learning_rate": 4.591795288383429e-06, + "loss": 2.2003, + "step": 491 + }, + { + "epoch": 0.5391780821917809, + "grad_norm": 1.730942964553833, + "learning_rate": 4.589764419171406e-06, + "loss": 2.1816, + "step": 492 + }, + { + "epoch": 0.5402739726027397, + "grad_norm": 1.7264728546142578, + "learning_rate": 4.587733549959383e-06, + "loss": 2.1615, + "step": 493 + }, + { + "epoch": 0.5413698630136986, + "grad_norm": 1.8276281356811523, + "learning_rate": 4.5857026807473604e-06, + "loss": 2.1442, + "step": 494 + }, + { + "epoch": 0.5424657534246575, + "grad_norm": 1.9947556257247925, + "learning_rate": 4.583671811535338e-06, + "loss": 2.2192, + "step": 495 + }, + { + "epoch": 0.5435616438356164, + "grad_norm": 1.770601511001587, + "learning_rate": 4.581640942323315e-06, + "loss": 2.233, + "step": 496 + }, + { + "epoch": 0.5446575342465754, + "grad_norm": 1.8946491479873657, + "learning_rate": 4.579610073111292e-06, + "loss": 2.2249, + "step": 497 + }, + { + "epoch": 0.5457534246575343, + "grad_norm": 1.7839634418487549, + "learning_rate": 4.5775792038992695e-06, + "loss": 2.2469, + "step": 498 + }, + { + "epoch": 0.5468493150684931, + "grad_norm": 1.6298807859420776, + "learning_rate": 4.575548334687247e-06, + "loss": 2.1723, + "step": 499 + }, + { + "epoch": 0.547945205479452, + "grad_norm": 1.9310041666030884, + "learning_rate": 4.573517465475224e-06, + "loss": 2.2156, + "step": 500 + }, + { + "epoch": 0.547945205479452, + "eval_loss": 2.3249237537384033, + "eval_runtime": 37.7505, + "eval_samples_per_second": 26.49, + "eval_steps_per_second": 1.113, + "step": 500 + }, + { + "epoch": 0.549041095890411, + "grad_norm": 1.7901647090911865, + "learning_rate": 4.571486596263201e-06, + "loss": 2.2169, + "step": 501 + }, + { + "epoch": 0.5501369863013699, + "grad_norm": 1.7739574909210205, + "learning_rate": 4.569455727051178e-06, + "loss": 2.2005, + "step": 502 + }, + { + "epoch": 0.5512328767123288, + "grad_norm": 1.7310121059417725, + "learning_rate": 4.567424857839156e-06, + "loss": 2.2369, + "step": 503 + }, + { + "epoch": 0.5523287671232877, + "grad_norm": 2.004469156265259, + "learning_rate": 4.565393988627132e-06, + "loss": 2.2051, + "step": 504 + }, + { + "epoch": 0.5534246575342465, + "grad_norm": 1.7906666994094849, + "learning_rate": 4.5633631194151105e-06, + "loss": 2.1908, + "step": 505 + }, + { + "epoch": 0.5545205479452054, + "grad_norm": 1.7224774360656738, + "learning_rate": 4.561332250203087e-06, + "loss": 2.1529, + "step": 506 + }, + { + "epoch": 0.5556164383561644, + "grad_norm": 1.9950156211853027, + "learning_rate": 4.559301380991065e-06, + "loss": 2.1776, + "step": 507 + }, + { + "epoch": 0.5567123287671233, + "grad_norm": 1.8257590532302856, + "learning_rate": 4.5572705117790414e-06, + "loss": 2.2267, + "step": 508 + }, + { + "epoch": 0.5578082191780822, + "grad_norm": 2.038490056991577, + "learning_rate": 4.5552396425670196e-06, + "loss": 2.1703, + "step": 509 + }, + { + "epoch": 0.5589041095890411, + "grad_norm": 1.7314448356628418, + "learning_rate": 4.553208773354996e-06, + "loss": 2.2409, + "step": 510 + }, + { + "epoch": 0.5589041095890411, + "eval_loss": 2.327078104019165, + "eval_runtime": 37.4614, + "eval_samples_per_second": 26.694, + "eval_steps_per_second": 1.121, + "step": 510 + }, + { + "epoch": 0.56, + "grad_norm": 2.0013606548309326, + "learning_rate": 4.551177904142974e-06, + "loss": 2.1732, + "step": 511 + }, + { + "epoch": 0.5610958904109589, + "grad_norm": 1.7352690696716309, + "learning_rate": 4.5491470349309505e-06, + "loss": 2.2027, + "step": 512 + }, + { + "epoch": 0.5621917808219178, + "grad_norm": 1.9024865627288818, + "learning_rate": 4.547116165718929e-06, + "loss": 2.1806, + "step": 513 + }, + { + "epoch": 0.5632876712328767, + "grad_norm": 1.8870717287063599, + "learning_rate": 4.545085296506905e-06, + "loss": 2.1664, + "step": 514 + }, + { + "epoch": 0.5643835616438356, + "grad_norm": 1.781226634979248, + "learning_rate": 4.543054427294882e-06, + "loss": 2.166, + "step": 515 + }, + { + "epoch": 0.5654794520547946, + "grad_norm": 1.8994221687316895, + "learning_rate": 4.54102355808286e-06, + "loss": 2.1951, + "step": 516 + }, + { + "epoch": 0.5665753424657535, + "grad_norm": 1.6999517679214478, + "learning_rate": 4.538992688870837e-06, + "loss": 2.1786, + "step": 517 + }, + { + "epoch": 0.5676712328767123, + "grad_norm": 1.899590253829956, + "learning_rate": 4.536961819658814e-06, + "loss": 2.2051, + "step": 518 + }, + { + "epoch": 0.5687671232876712, + "grad_norm": 1.8054970502853394, + "learning_rate": 4.5349309504467915e-06, + "loss": 2.1675, + "step": 519 + }, + { + "epoch": 0.5698630136986301, + "grad_norm": 1.6732861995697021, + "learning_rate": 4.532900081234769e-06, + "loss": 2.1997, + "step": 520 + }, + { + "epoch": 0.5698630136986301, + "eval_loss": 2.3242335319519043, + "eval_runtime": 37.6748, + "eval_samples_per_second": 26.543, + "eval_steps_per_second": 1.115, + "step": 520 + }, + { + "epoch": 0.570958904109589, + "grad_norm": 2.0063157081604004, + "learning_rate": 4.530869212022746e-06, + "loss": 2.1909, + "step": 521 + }, + { + "epoch": 0.572054794520548, + "grad_norm": 1.7750036716461182, + "learning_rate": 4.528838342810723e-06, + "loss": 2.1935, + "step": 522 + }, + { + "epoch": 0.5731506849315069, + "grad_norm": 1.9993029832839966, + "learning_rate": 4.526807473598701e-06, + "loss": 2.1753, + "step": 523 + }, + { + "epoch": 0.5742465753424657, + "grad_norm": 1.6742994785308838, + "learning_rate": 4.524776604386678e-06, + "loss": 2.2205, + "step": 524 + }, + { + "epoch": 0.5753424657534246, + "grad_norm": 1.9300428628921509, + "learning_rate": 4.522745735174655e-06, + "loss": 2.2081, + "step": 525 + }, + { + "epoch": 0.5764383561643835, + "grad_norm": 1.8770017623901367, + "learning_rate": 4.520714865962632e-06, + "loss": 2.2074, + "step": 526 + }, + { + "epoch": 0.5775342465753425, + "grad_norm": 1.847765326499939, + "learning_rate": 4.51868399675061e-06, + "loss": 2.1775, + "step": 527 + }, + { + "epoch": 0.5786301369863014, + "grad_norm": 1.905117392539978, + "learning_rate": 4.516653127538587e-06, + "loss": 2.1511, + "step": 528 + }, + { + "epoch": 0.5797260273972603, + "grad_norm": 1.7623670101165771, + "learning_rate": 4.514622258326564e-06, + "loss": 2.188, + "step": 529 + }, + { + "epoch": 0.5808219178082191, + "grad_norm": 1.9232176542282104, + "learning_rate": 4.5125913891145415e-06, + "loss": 2.2109, + "step": 530 + }, + { + "epoch": 0.5808219178082191, + "eval_loss": 2.3252475261688232, + "eval_runtime": 37.6923, + "eval_samples_per_second": 26.531, + "eval_steps_per_second": 1.114, + "step": 530 + }, + { + "epoch": 0.581917808219178, + "grad_norm": 1.6523535251617432, + "learning_rate": 4.510560519902519e-06, + "loss": 2.1499, + "step": 531 + }, + { + "epoch": 0.583013698630137, + "grad_norm": 1.9934184551239014, + "learning_rate": 4.508529650690496e-06, + "loss": 2.1383, + "step": 532 + }, + { + "epoch": 0.5841095890410959, + "grad_norm": 1.704522728919983, + "learning_rate": 4.506498781478473e-06, + "loss": 2.1892, + "step": 533 + }, + { + "epoch": 0.5852054794520548, + "grad_norm": 1.9909989833831787, + "learning_rate": 4.504467912266451e-06, + "loss": 2.2422, + "step": 534 + }, + { + "epoch": 0.5863013698630137, + "grad_norm": 1.7490800619125366, + "learning_rate": 4.502437043054428e-06, + "loss": 2.2008, + "step": 535 + }, + { + "epoch": 0.5873972602739727, + "grad_norm": 1.8620145320892334, + "learning_rate": 4.500406173842405e-06, + "loss": 2.1892, + "step": 536 + }, + { + "epoch": 0.5884931506849315, + "grad_norm": 1.7069931030273438, + "learning_rate": 4.4983753046303824e-06, + "loss": 2.1816, + "step": 537 + }, + { + "epoch": 0.5895890410958904, + "grad_norm": 2.152491807937622, + "learning_rate": 4.49634443541836e-06, + "loss": 2.2331, + "step": 538 + }, + { + "epoch": 0.5906849315068493, + "grad_norm": 1.8526151180267334, + "learning_rate": 4.494313566206336e-06, + "loss": 2.1811, + "step": 539 + }, + { + "epoch": 0.5917808219178082, + "grad_norm": 1.6865259408950806, + "learning_rate": 4.492282696994314e-06, + "loss": 2.1599, + "step": 540 + }, + { + "epoch": 0.5917808219178082, + "eval_loss": 2.3265392780303955, + "eval_runtime": 37.7255, + "eval_samples_per_second": 26.507, + "eval_steps_per_second": 1.113, + "step": 540 + }, + { + "epoch": 0.5928767123287672, + "grad_norm": 1.7710354328155518, + "learning_rate": 4.490251827782291e-06, + "loss": 2.1636, + "step": 541 + }, + { + "epoch": 0.5939726027397261, + "grad_norm": 1.8444210290908813, + "learning_rate": 4.488220958570269e-06, + "loss": 2.1345, + "step": 542 + }, + { + "epoch": 0.5950684931506849, + "grad_norm": 1.9095014333724976, + "learning_rate": 4.486190089358245e-06, + "loss": 2.2236, + "step": 543 + }, + { + "epoch": 0.5961643835616438, + "grad_norm": 1.7406108379364014, + "learning_rate": 4.484159220146223e-06, + "loss": 2.2078, + "step": 544 + }, + { + "epoch": 0.5972602739726027, + "grad_norm": 1.712534785270691, + "learning_rate": 4.4821283509342e-06, + "loss": 2.1809, + "step": 545 + }, + { + "epoch": 0.5983561643835617, + "grad_norm": 1.8921459913253784, + "learning_rate": 4.480097481722178e-06, + "loss": 2.158, + "step": 546 + }, + { + "epoch": 0.5994520547945206, + "grad_norm": 1.805146336555481, + "learning_rate": 4.478066612510154e-06, + "loss": 2.1765, + "step": 547 + }, + { + "epoch": 0.6005479452054795, + "grad_norm": 2.1620516777038574, + "learning_rate": 4.4760357432981325e-06, + "loss": 2.2832, + "step": 548 + }, + { + "epoch": 0.6016438356164383, + "grad_norm": 1.661618947982788, + "learning_rate": 4.474004874086109e-06, + "loss": 2.1397, + "step": 549 + }, + { + "epoch": 0.6027397260273972, + "grad_norm": 1.732483983039856, + "learning_rate": 4.471974004874087e-06, + "loss": 2.225, + "step": 550 + }, + { + "epoch": 0.6027397260273972, + "eval_loss": 2.3276846408843994, + "eval_runtime": 37.7156, + "eval_samples_per_second": 26.514, + "eval_steps_per_second": 1.114, + "step": 550 + }, + { + "epoch": 0.6038356164383561, + "grad_norm": 1.8693495988845825, + "learning_rate": 4.4699431356620635e-06, + "loss": 2.1959, + "step": 551 + }, + { + "epoch": 0.6049315068493151, + "grad_norm": 1.8201690912246704, + "learning_rate": 4.467912266450041e-06, + "loss": 2.1882, + "step": 552 + }, + { + "epoch": 0.606027397260274, + "grad_norm": 1.9229968786239624, + "learning_rate": 4.465881397238018e-06, + "loss": 2.2006, + "step": 553 + }, + { + "epoch": 0.6071232876712329, + "grad_norm": 1.6330770254135132, + "learning_rate": 4.463850528025995e-06, + "loss": 2.209, + "step": 554 + }, + { + "epoch": 0.6082191780821918, + "grad_norm": 1.9654744863510132, + "learning_rate": 4.4618196588139726e-06, + "loss": 2.1963, + "step": 555 + }, + { + "epoch": 0.6093150684931506, + "grad_norm": 1.8808507919311523, + "learning_rate": 4.45978878960195e-06, + "loss": 2.1538, + "step": 556 + }, + { + "epoch": 0.6104109589041096, + "grad_norm": 1.8481040000915527, + "learning_rate": 4.457757920389927e-06, + "loss": 2.1769, + "step": 557 + }, + { + "epoch": 0.6115068493150685, + "grad_norm": 2.114439010620117, + "learning_rate": 4.455727051177904e-06, + "loss": 2.2241, + "step": 558 + }, + { + "epoch": 0.6126027397260274, + "grad_norm": 1.5492331981658936, + "learning_rate": 4.453696181965882e-06, + "loss": 2.206, + "step": 559 + }, + { + "epoch": 0.6136986301369863, + "grad_norm": 2.11787486076355, + "learning_rate": 4.451665312753859e-06, + "loss": 2.2328, + "step": 560 + }, + { + "epoch": 0.6136986301369863, + "eval_loss": 2.3215034008026123, + "eval_runtime": 37.7471, + "eval_samples_per_second": 26.492, + "eval_steps_per_second": 1.113, + "step": 560 + }, + { + "epoch": 0.6147945205479453, + "grad_norm": 1.6561832427978516, + "learning_rate": 4.449634443541836e-06, + "loss": 2.1607, + "step": 561 + }, + { + "epoch": 0.6158904109589041, + "grad_norm": 1.7510966062545776, + "learning_rate": 4.4476035743298135e-06, + "loss": 2.1699, + "step": 562 + }, + { + "epoch": 0.616986301369863, + "grad_norm": 1.9424657821655273, + "learning_rate": 4.445572705117791e-06, + "loss": 2.2298, + "step": 563 + }, + { + "epoch": 0.6180821917808219, + "grad_norm": 1.5836161375045776, + "learning_rate": 4.443541835905768e-06, + "loss": 2.1711, + "step": 564 + }, + { + "epoch": 0.6191780821917808, + "grad_norm": 1.895216464996338, + "learning_rate": 4.441510966693745e-06, + "loss": 2.2252, + "step": 565 + }, + { + "epoch": 0.6202739726027398, + "grad_norm": 2.3365776538848877, + "learning_rate": 4.439480097481723e-06, + "loss": 2.1309, + "step": 566 + }, + { + "epoch": 0.6213698630136987, + "grad_norm": 1.6988815069198608, + "learning_rate": 4.4374492282697e-06, + "loss": 2.2023, + "step": 567 + }, + { + "epoch": 0.6224657534246575, + "grad_norm": 2.228900671005249, + "learning_rate": 4.435418359057677e-06, + "loss": 2.2021, + "step": 568 + }, + { + "epoch": 0.6235616438356164, + "grad_norm": 1.5536632537841797, + "learning_rate": 4.433387489845654e-06, + "loss": 2.1706, + "step": 569 + }, + { + "epoch": 0.6246575342465753, + "grad_norm": 1.8463878631591797, + "learning_rate": 4.431356620633632e-06, + "loss": 2.1715, + "step": 570 + }, + { + "epoch": 0.6246575342465753, + "eval_loss": 2.320664167404175, + "eval_runtime": 37.7539, + "eval_samples_per_second": 26.487, + "eval_steps_per_second": 1.112, + "step": 570 + }, + { + "epoch": 0.6257534246575343, + "grad_norm": 1.924741506576538, + "learning_rate": 4.429325751421609e-06, + "loss": 2.1894, + "step": 571 + }, + { + "epoch": 0.6268493150684932, + "grad_norm": 1.5921828746795654, + "learning_rate": 4.427294882209586e-06, + "loss": 2.1358, + "step": 572 + }, + { + "epoch": 0.6279452054794521, + "grad_norm": 2.4617693424224854, + "learning_rate": 4.4252640129975635e-06, + "loss": 2.1918, + "step": 573 + }, + { + "epoch": 0.6290410958904109, + "grad_norm": 1.6239196062088013, + "learning_rate": 4.423233143785541e-06, + "loss": 2.1996, + "step": 574 + }, + { + "epoch": 0.6301369863013698, + "grad_norm": 1.952112078666687, + "learning_rate": 4.421202274573518e-06, + "loss": 2.112, + "step": 575 + }, + { + "epoch": 0.6312328767123287, + "grad_norm": 1.663409948348999, + "learning_rate": 4.4191714053614945e-06, + "loss": 2.149, + "step": 576 + }, + { + "epoch": 0.6323287671232877, + "grad_norm": 1.831711769104004, + "learning_rate": 4.417140536149473e-06, + "loss": 2.209, + "step": 577 + }, + { + "epoch": 0.6334246575342466, + "grad_norm": 1.9281738996505737, + "learning_rate": 4.415109666937449e-06, + "loss": 2.1637, + "step": 578 + }, + { + "epoch": 0.6345205479452055, + "grad_norm": 1.5661600828170776, + "learning_rate": 4.413078797725427e-06, + "loss": 2.1832, + "step": 579 + }, + { + "epoch": 0.6356164383561644, + "grad_norm": 1.6930522918701172, + "learning_rate": 4.411047928513404e-06, + "loss": 2.196, + "step": 580 + }, + { + "epoch": 0.6356164383561644, + "eval_loss": 2.318861246109009, + "eval_runtime": 37.6994, + "eval_samples_per_second": 26.526, + "eval_steps_per_second": 1.114, + "step": 580 + }, + { + "epoch": 0.6367123287671232, + "grad_norm": 1.74558424949646, + "learning_rate": 4.409017059301382e-06, + "loss": 2.2097, + "step": 581 + }, + { + "epoch": 0.6378082191780822, + "grad_norm": 1.664880633354187, + "learning_rate": 4.406986190089358e-06, + "loss": 2.2076, + "step": 582 + }, + { + "epoch": 0.6389041095890411, + "grad_norm": 1.5981497764587402, + "learning_rate": 4.404955320877336e-06, + "loss": 2.1706, + "step": 583 + }, + { + "epoch": 0.64, + "grad_norm": 1.6035112142562866, + "learning_rate": 4.402924451665313e-06, + "loss": 2.1538, + "step": 584 + }, + { + "epoch": 0.6410958904109589, + "grad_norm": 1.6819255352020264, + "learning_rate": 4.400893582453291e-06, + "loss": 2.1816, + "step": 585 + }, + { + "epoch": 0.6421917808219179, + "grad_norm": 1.7082457542419434, + "learning_rate": 4.398862713241267e-06, + "loss": 2.1804, + "step": 586 + }, + { + "epoch": 0.6432876712328767, + "grad_norm": 1.6893151998519897, + "learning_rate": 4.396831844029245e-06, + "loss": 2.1783, + "step": 587 + }, + { + "epoch": 0.6443835616438356, + "grad_norm": 1.8639923334121704, + "learning_rate": 4.394800974817222e-06, + "loss": 2.1989, + "step": 588 + }, + { + "epoch": 0.6454794520547945, + "grad_norm": 1.6407119035720825, + "learning_rate": 4.392770105605199e-06, + "loss": 2.2013, + "step": 589 + }, + { + "epoch": 0.6465753424657534, + "grad_norm": 1.6435848474502563, + "learning_rate": 4.390739236393176e-06, + "loss": 2.2049, + "step": 590 + }, + { + "epoch": 0.6465753424657534, + "eval_loss": 2.318084478378296, + "eval_runtime": 37.6999, + "eval_samples_per_second": 26.525, + "eval_steps_per_second": 1.114, + "step": 590 + }, + { + "epoch": 0.6476712328767124, + "grad_norm": 1.7974674701690674, + "learning_rate": 4.388708367181154e-06, + "loss": 2.188, + "step": 591 + }, + { + "epoch": 0.6487671232876713, + "grad_norm": 1.8340431451797485, + "learning_rate": 4.386677497969131e-06, + "loss": 2.2305, + "step": 592 + }, + { + "epoch": 0.6498630136986301, + "grad_norm": 1.55428946018219, + "learning_rate": 4.384646628757108e-06, + "loss": 2.1837, + "step": 593 + }, + { + "epoch": 0.650958904109589, + "grad_norm": 1.7258585691452026, + "learning_rate": 4.3826157595450855e-06, + "loss": 2.1426, + "step": 594 + }, + { + "epoch": 0.6520547945205479, + "grad_norm": 1.7347694635391235, + "learning_rate": 4.380584890333063e-06, + "loss": 2.1161, + "step": 595 + }, + { + "epoch": 0.6531506849315069, + "grad_norm": 1.6082992553710938, + "learning_rate": 4.37855402112104e-06, + "loss": 2.2284, + "step": 596 + }, + { + "epoch": 0.6542465753424658, + "grad_norm": 1.9446425437927246, + "learning_rate": 4.376523151909017e-06, + "loss": 2.1571, + "step": 597 + }, + { + "epoch": 0.6553424657534247, + "grad_norm": 1.8597826957702637, + "learning_rate": 4.3744922826969946e-06, + "loss": 2.1982, + "step": 598 + }, + { + "epoch": 0.6564383561643835, + "grad_norm": 1.8589327335357666, + "learning_rate": 4.372461413484972e-06, + "loss": 2.2097, + "step": 599 + }, + { + "epoch": 0.6575342465753424, + "grad_norm": 1.6825448274612427, + "learning_rate": 4.370430544272949e-06, + "loss": 2.2077, + "step": 600 + }, + { + "epoch": 0.6575342465753424, + "eval_loss": 2.315889596939087, + "eval_runtime": 37.6852, + "eval_samples_per_second": 26.536, + "eval_steps_per_second": 1.114, + "step": 600 + }, + { + "epoch": 0.6586301369863014, + "grad_norm": 1.8007057905197144, + "learning_rate": 4.368399675060926e-06, + "loss": 2.1843, + "step": 601 + }, + { + "epoch": 0.6597260273972603, + "grad_norm": 2.057697057723999, + "learning_rate": 4.366368805848904e-06, + "loss": 2.13, + "step": 602 + }, + { + "epoch": 0.6608219178082192, + "grad_norm": 1.5714081525802612, + "learning_rate": 4.364337936636881e-06, + "loss": 2.1879, + "step": 603 + }, + { + "epoch": 0.6619178082191781, + "grad_norm": 1.855185866355896, + "learning_rate": 4.362307067424858e-06, + "loss": 2.1659, + "step": 604 + }, + { + "epoch": 0.663013698630137, + "grad_norm": 1.826479196548462, + "learning_rate": 4.3602761982128355e-06, + "loss": 2.1716, + "step": 605 + }, + { + "epoch": 0.6641095890410958, + "grad_norm": 1.6905351877212524, + "learning_rate": 4.358245329000813e-06, + "loss": 2.1825, + "step": 606 + }, + { + "epoch": 0.6652054794520548, + "grad_norm": 1.8157447576522827, + "learning_rate": 4.35621445978879e-06, + "loss": 2.1811, + "step": 607 + }, + { + "epoch": 0.6663013698630137, + "grad_norm": 1.6311850547790527, + "learning_rate": 4.354183590576767e-06, + "loss": 2.1507, + "step": 608 + }, + { + "epoch": 0.6673972602739726, + "grad_norm": 1.579039454460144, + "learning_rate": 4.352152721364745e-06, + "loss": 2.2031, + "step": 609 + }, + { + "epoch": 0.6684931506849315, + "grad_norm": 1.7773518562316895, + "learning_rate": 4.350121852152722e-06, + "loss": 2.1978, + "step": 610 + }, + { + "epoch": 0.6684931506849315, + "eval_loss": 2.314854383468628, + "eval_runtime": 37.4501, + "eval_samples_per_second": 26.702, + "eval_steps_per_second": 1.121, + "step": 610 + }, + { + "epoch": 0.6695890410958905, + "grad_norm": 1.7606587409973145, + "learning_rate": 4.348090982940699e-06, + "loss": 2.1891, + "step": 611 + }, + { + "epoch": 0.6706849315068493, + "grad_norm": 1.6145644187927246, + "learning_rate": 4.3460601137286764e-06, + "loss": 2.1799, + "step": 612 + }, + { + "epoch": 0.6717808219178082, + "grad_norm": 1.8070226907730103, + "learning_rate": 4.344029244516654e-06, + "loss": 2.241, + "step": 613 + }, + { + "epoch": 0.6728767123287671, + "grad_norm": 1.6185903549194336, + "learning_rate": 4.341998375304631e-06, + "loss": 2.1908, + "step": 614 + }, + { + "epoch": 0.673972602739726, + "grad_norm": 1.6690049171447754, + "learning_rate": 4.339967506092607e-06, + "loss": 2.1804, + "step": 615 + }, + { + "epoch": 0.675068493150685, + "grad_norm": 1.740003228187561, + "learning_rate": 4.3379366368805855e-06, + "loss": 2.1734, + "step": 616 + }, + { + "epoch": 0.6761643835616439, + "grad_norm": 1.7311010360717773, + "learning_rate": 4.335905767668562e-06, + "loss": 2.2015, + "step": 617 + }, + { + "epoch": 0.6772602739726027, + "grad_norm": 1.7491734027862549, + "learning_rate": 4.33387489845654e-06, + "loss": 2.1894, + "step": 618 + }, + { + "epoch": 0.6783561643835616, + "grad_norm": 1.6343051195144653, + "learning_rate": 4.3318440292445165e-06, + "loss": 2.1583, + "step": 619 + }, + { + "epoch": 0.6794520547945205, + "grad_norm": 1.584243893623352, + "learning_rate": 4.329813160032495e-06, + "loss": 2.1396, + "step": 620 + }, + { + "epoch": 0.6794520547945205, + "eval_loss": 2.3157904148101807, + "eval_runtime": 37.6734, + "eval_samples_per_second": 26.544, + "eval_steps_per_second": 1.115, + "step": 620 + }, + { + "epoch": 0.6805479452054795, + "grad_norm": 1.6275745630264282, + "learning_rate": 4.327782290820471e-06, + "loss": 2.2064, + "step": 621 + }, + { + "epoch": 0.6816438356164384, + "grad_norm": 1.5353672504425049, + "learning_rate": 4.325751421608449e-06, + "loss": 2.1686, + "step": 622 + }, + { + "epoch": 0.6827397260273973, + "grad_norm": 1.6155133247375488, + "learning_rate": 4.323720552396426e-06, + "loss": 2.1652, + "step": 623 + }, + { + "epoch": 0.6838356164383562, + "grad_norm": 1.638383150100708, + "learning_rate": 4.321689683184404e-06, + "loss": 2.1747, + "step": 624 + }, + { + "epoch": 0.684931506849315, + "grad_norm": 1.6153279542922974, + "learning_rate": 4.31965881397238e-06, + "loss": 2.1975, + "step": 625 + }, + { + "epoch": 0.686027397260274, + "grad_norm": 2.023699998855591, + "learning_rate": 4.3176279447603574e-06, + "loss": 2.1904, + "step": 626 + }, + { + "epoch": 0.6871232876712329, + "grad_norm": 1.5969340801239014, + "learning_rate": 4.315597075548335e-06, + "loss": 2.2133, + "step": 627 + }, + { + "epoch": 0.6882191780821918, + "grad_norm": 2.138498067855835, + "learning_rate": 4.313566206336312e-06, + "loss": 2.1685, + "step": 628 + }, + { + "epoch": 0.6893150684931507, + "grad_norm": 1.7141013145446777, + "learning_rate": 4.311535337124289e-06, + "loss": 2.1818, + "step": 629 + }, + { + "epoch": 0.6904109589041096, + "grad_norm": 1.9214462041854858, + "learning_rate": 4.3095044679122665e-06, + "loss": 2.191, + "step": 630 + }, + { + "epoch": 0.6904109589041096, + "eval_loss": 2.3150644302368164, + "eval_runtime": 37.723, + "eval_samples_per_second": 26.509, + "eval_steps_per_second": 1.113, + "step": 630 + }, + { + "epoch": 0.6915068493150685, + "grad_norm": 2.03241229057312, + "learning_rate": 4.307473598700244e-06, + "loss": 2.1398, + "step": 631 + }, + { + "epoch": 0.6926027397260274, + "grad_norm": 1.608492374420166, + "learning_rate": 4.305442729488221e-06, + "loss": 2.2113, + "step": 632 + }, + { + "epoch": 0.6936986301369863, + "grad_norm": 2.1223549842834473, + "learning_rate": 4.303411860276198e-06, + "loss": 2.2012, + "step": 633 + }, + { + "epoch": 0.6947945205479452, + "grad_norm": 1.5212339162826538, + "learning_rate": 4.301380991064176e-06, + "loss": 2.2201, + "step": 634 + }, + { + "epoch": 0.6958904109589041, + "grad_norm": 1.9565715789794922, + "learning_rate": 4.299350121852153e-06, + "loss": 2.2155, + "step": 635 + }, + { + "epoch": 0.6969863013698631, + "grad_norm": 1.700822114944458, + "learning_rate": 4.29731925264013e-06, + "loss": 2.1764, + "step": 636 + }, + { + "epoch": 0.6980821917808219, + "grad_norm": 1.5625321865081787, + "learning_rate": 4.2952883834281075e-06, + "loss": 2.1407, + "step": 637 + }, + { + "epoch": 0.6991780821917808, + "grad_norm": 2.3319756984710693, + "learning_rate": 4.293257514216085e-06, + "loss": 2.2058, + "step": 638 + }, + { + "epoch": 0.7002739726027397, + "grad_norm": 1.617644190788269, + "learning_rate": 4.291226645004062e-06, + "loss": 2.2157, + "step": 639 + }, + { + "epoch": 0.7013698630136986, + "grad_norm": 1.9357377290725708, + "learning_rate": 4.289195775792039e-06, + "loss": 2.1736, + "step": 640 + }, + { + "epoch": 0.7013698630136986, + "eval_loss": 2.310267925262451, + "eval_runtime": 37.731, + "eval_samples_per_second": 26.503, + "eval_steps_per_second": 1.113, + "step": 640 + }, + { + "epoch": 0.7024657534246576, + "grad_norm": 1.6908962726593018, + "learning_rate": 4.287164906580017e-06, + "loss": 2.1523, + "step": 641 + }, + { + "epoch": 0.7035616438356165, + "grad_norm": 1.6051634550094604, + "learning_rate": 4.285134037367994e-06, + "loss": 2.1499, + "step": 642 + }, + { + "epoch": 0.7046575342465753, + "grad_norm": 1.760164499282837, + "learning_rate": 4.283103168155971e-06, + "loss": 2.1493, + "step": 643 + }, + { + "epoch": 0.7057534246575342, + "grad_norm": 1.6225790977478027, + "learning_rate": 4.281072298943948e-06, + "loss": 2.1737, + "step": 644 + }, + { + "epoch": 0.7068493150684931, + "grad_norm": 1.8128079175949097, + "learning_rate": 4.279041429731926e-06, + "loss": 2.1492, + "step": 645 + }, + { + "epoch": 0.707945205479452, + "grad_norm": 1.993295431137085, + "learning_rate": 4.277010560519903e-06, + "loss": 2.1702, + "step": 646 + }, + { + "epoch": 0.709041095890411, + "grad_norm": 1.7715190649032593, + "learning_rate": 4.27497969130788e-06, + "loss": 2.1939, + "step": 647 + }, + { + "epoch": 0.7101369863013699, + "grad_norm": 1.921738624572754, + "learning_rate": 4.2729488220958575e-06, + "loss": 2.201, + "step": 648 + }, + { + "epoch": 0.7112328767123288, + "grad_norm": 1.5670042037963867, + "learning_rate": 4.270917952883835e-06, + "loss": 2.152, + "step": 649 + }, + { + "epoch": 0.7123287671232876, + "grad_norm": 1.6405668258666992, + "learning_rate": 4.268887083671812e-06, + "loss": 2.2488, + "step": 650 + }, + { + "epoch": 0.7123287671232876, + "eval_loss": 2.3100948333740234, + "eval_runtime": 37.7393, + "eval_samples_per_second": 26.498, + "eval_steps_per_second": 1.113, + "step": 650 + }, + { + "epoch": 0.7134246575342466, + "grad_norm": 1.79966402053833, + "learning_rate": 4.266856214459789e-06, + "loss": 2.1506, + "step": 651 + }, + { + "epoch": 0.7145205479452055, + "grad_norm": 1.6062074899673462, + "learning_rate": 4.264825345247766e-06, + "loss": 2.1818, + "step": 652 + }, + { + "epoch": 0.7156164383561644, + "grad_norm": 1.61439049243927, + "learning_rate": 4.262794476035744e-06, + "loss": 2.1269, + "step": 653 + }, + { + "epoch": 0.7167123287671233, + "grad_norm": 1.7649872303009033, + "learning_rate": 4.26076360682372e-06, + "loss": 2.1703, + "step": 654 + }, + { + "epoch": 0.7178082191780822, + "grad_norm": 1.6238528490066528, + "learning_rate": 4.2587327376116984e-06, + "loss": 2.1764, + "step": 655 + }, + { + "epoch": 0.718904109589041, + "grad_norm": 1.6847354173660278, + "learning_rate": 4.256701868399675e-06, + "loss": 2.1796, + "step": 656 + }, + { + "epoch": 0.72, + "grad_norm": 1.7208436727523804, + "learning_rate": 4.254670999187653e-06, + "loss": 2.1903, + "step": 657 + }, + { + "epoch": 0.7210958904109589, + "grad_norm": 1.5888099670410156, + "learning_rate": 4.252640129975629e-06, + "loss": 2.1686, + "step": 658 + }, + { + "epoch": 0.7221917808219178, + "grad_norm": 1.742118000984192, + "learning_rate": 4.2506092607636075e-06, + "loss": 2.1287, + "step": 659 + }, + { + "epoch": 0.7232876712328767, + "grad_norm": 1.8450883626937866, + "learning_rate": 4.248578391551584e-06, + "loss": 2.1761, + "step": 660 + }, + { + "epoch": 0.7232876712328767, + "eval_loss": 2.3134732246398926, + "eval_runtime": 37.7568, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 660 + }, + { + "epoch": 0.7243835616438357, + "grad_norm": 1.630985140800476, + "learning_rate": 4.246547522339562e-06, + "loss": 2.1743, + "step": 661 + }, + { + "epoch": 0.7254794520547945, + "grad_norm": 1.6984914541244507, + "learning_rate": 4.2445166531275385e-06, + "loss": 2.1356, + "step": 662 + }, + { + "epoch": 0.7265753424657534, + "grad_norm": 1.6732920408248901, + "learning_rate": 4.242485783915517e-06, + "loss": 2.1377, + "step": 663 + }, + { + "epoch": 0.7276712328767123, + "grad_norm": 1.6210153102874756, + "learning_rate": 4.240454914703493e-06, + "loss": 2.1455, + "step": 664 + }, + { + "epoch": 0.7287671232876712, + "grad_norm": 1.7643382549285889, + "learning_rate": 4.23842404549147e-06, + "loss": 2.1373, + "step": 665 + }, + { + "epoch": 0.7298630136986302, + "grad_norm": 1.6491565704345703, + "learning_rate": 4.236393176279448e-06, + "loss": 2.1897, + "step": 666 + }, + { + "epoch": 0.7309589041095891, + "grad_norm": 1.670323133468628, + "learning_rate": 4.234362307067425e-06, + "loss": 2.2081, + "step": 667 + }, + { + "epoch": 0.7320547945205479, + "grad_norm": 1.6350057125091553, + "learning_rate": 4.232331437855402e-06, + "loss": 2.2002, + "step": 668 + }, + { + "epoch": 0.7331506849315068, + "grad_norm": 1.6191266775131226, + "learning_rate": 4.2303005686433795e-06, + "loss": 2.1744, + "step": 669 + }, + { + "epoch": 0.7342465753424657, + "grad_norm": 1.5511432886123657, + "learning_rate": 4.228269699431357e-06, + "loss": 2.1797, + "step": 670 + }, + { + "epoch": 0.7342465753424657, + "eval_loss": 2.313626527786255, + "eval_runtime": 37.7763, + "eval_samples_per_second": 26.472, + "eval_steps_per_second": 1.112, + "step": 670 + }, + { + "epoch": 0.7353424657534247, + "grad_norm": 1.5128833055496216, + "learning_rate": 4.226238830219334e-06, + "loss": 2.1729, + "step": 671 + }, + { + "epoch": 0.7364383561643836, + "grad_norm": 1.6752032041549683, + "learning_rate": 4.224207961007311e-06, + "loss": 2.1747, + "step": 672 + }, + { + "epoch": 0.7375342465753425, + "grad_norm": 1.8643760681152344, + "learning_rate": 4.2221770917952886e-06, + "loss": 2.174, + "step": 673 + }, + { + "epoch": 0.7386301369863014, + "grad_norm": 1.5659266710281372, + "learning_rate": 4.220146222583266e-06, + "loss": 2.1706, + "step": 674 + }, + { + "epoch": 0.7397260273972602, + "grad_norm": 1.9675986766815186, + "learning_rate": 4.218115353371243e-06, + "loss": 2.1558, + "step": 675 + }, + { + "epoch": 0.7408219178082192, + "grad_norm": 1.6091173887252808, + "learning_rate": 4.21608448415922e-06, + "loss": 2.2126, + "step": 676 + }, + { + "epoch": 0.7419178082191781, + "grad_norm": 1.6801071166992188, + "learning_rate": 4.214053614947198e-06, + "loss": 2.1352, + "step": 677 + }, + { + "epoch": 0.743013698630137, + "grad_norm": 1.6964598894119263, + "learning_rate": 4.212022745735175e-06, + "loss": 2.1863, + "step": 678 + }, + { + "epoch": 0.7441095890410959, + "grad_norm": 1.670597791671753, + "learning_rate": 4.209991876523152e-06, + "loss": 2.1478, + "step": 679 + }, + { + "epoch": 0.7452054794520548, + "grad_norm": 1.6591588258743286, + "learning_rate": 4.2079610073111295e-06, + "loss": 2.2198, + "step": 680 + }, + { + "epoch": 0.7452054794520548, + "eval_loss": 2.3091039657592773, + "eval_runtime": 37.7551, + "eval_samples_per_second": 26.486, + "eval_steps_per_second": 1.112, + "step": 680 + }, + { + "epoch": 0.7463013698630137, + "grad_norm": 1.5147838592529297, + "learning_rate": 4.205930138099107e-06, + "loss": 2.1836, + "step": 681 + }, + { + "epoch": 0.7473972602739726, + "grad_norm": 1.6278952360153198, + "learning_rate": 4.203899268887084e-06, + "loss": 2.196, + "step": 682 + }, + { + "epoch": 0.7484931506849315, + "grad_norm": 1.5964363813400269, + "learning_rate": 4.201868399675061e-06, + "loss": 2.1162, + "step": 683 + }, + { + "epoch": 0.7495890410958904, + "grad_norm": 1.6993104219436646, + "learning_rate": 4.199837530463039e-06, + "loss": 2.1731, + "step": 684 + }, + { + "epoch": 0.7506849315068493, + "grad_norm": 1.6196811199188232, + "learning_rate": 4.197806661251016e-06, + "loss": 2.16, + "step": 685 + }, + { + "epoch": 0.7517808219178083, + "grad_norm": 1.6826317310333252, + "learning_rate": 4.195775792038993e-06, + "loss": 2.1123, + "step": 686 + }, + { + "epoch": 0.7528767123287671, + "grad_norm": 1.7604600191116333, + "learning_rate": 4.19374492282697e-06, + "loss": 2.1962, + "step": 687 + }, + { + "epoch": 0.753972602739726, + "grad_norm": 1.5533616542816162, + "learning_rate": 4.191714053614948e-06, + "loss": 2.1593, + "step": 688 + }, + { + "epoch": 0.7550684931506849, + "grad_norm": 1.8652232885360718, + "learning_rate": 4.189683184402924e-06, + "loss": 2.2067, + "step": 689 + }, + { + "epoch": 0.7561643835616438, + "grad_norm": 1.6330097913742065, + "learning_rate": 4.187652315190902e-06, + "loss": 2.1483, + "step": 690 + }, + { + "epoch": 0.7561643835616438, + "eval_loss": 2.309835195541382, + "eval_runtime": 37.746, + "eval_samples_per_second": 26.493, + "eval_steps_per_second": 1.113, + "step": 690 + }, + { + "epoch": 0.7572602739726028, + "grad_norm": 1.5989009141921997, + "learning_rate": 4.185621445978879e-06, + "loss": 2.1686, + "step": 691 + }, + { + "epoch": 0.7583561643835617, + "grad_norm": 1.7416167259216309, + "learning_rate": 4.183590576766857e-06, + "loss": 2.2111, + "step": 692 + }, + { + "epoch": 0.7594520547945205, + "grad_norm": 1.5840280055999756, + "learning_rate": 4.181559707554833e-06, + "loss": 2.1499, + "step": 693 + }, + { + "epoch": 0.7605479452054794, + "grad_norm": 1.839306354522705, + "learning_rate": 4.179528838342811e-06, + "loss": 2.1492, + "step": 694 + }, + { + "epoch": 0.7616438356164383, + "grad_norm": 1.54163658618927, + "learning_rate": 4.177497969130788e-06, + "loss": 2.2029, + "step": 695 + }, + { + "epoch": 0.7627397260273973, + "grad_norm": 1.8050168752670288, + "learning_rate": 4.175467099918766e-06, + "loss": 2.1615, + "step": 696 + }, + { + "epoch": 0.7638356164383562, + "grad_norm": 1.6971197128295898, + "learning_rate": 4.173436230706742e-06, + "loss": 2.1476, + "step": 697 + }, + { + "epoch": 0.7649315068493151, + "grad_norm": 1.6200546026229858, + "learning_rate": 4.1714053614947205e-06, + "loss": 2.1602, + "step": 698 + }, + { + "epoch": 0.766027397260274, + "grad_norm": 1.7440853118896484, + "learning_rate": 4.169374492282697e-06, + "loss": 2.1949, + "step": 699 + }, + { + "epoch": 0.7671232876712328, + "grad_norm": 1.5590331554412842, + "learning_rate": 4.167343623070675e-06, + "loss": 2.1656, + "step": 700 + }, + { + "epoch": 0.7671232876712328, + "eval_loss": 2.309752941131592, + "eval_runtime": 37.7663, + "eval_samples_per_second": 26.479, + "eval_steps_per_second": 1.112, + "step": 700 + }, + { + "epoch": 0.7682191780821918, + "grad_norm": 1.7314256429672241, + "learning_rate": 4.165312753858652e-06, + "loss": 2.1516, + "step": 701 + }, + { + "epoch": 0.7693150684931507, + "grad_norm": 1.5145938396453857, + "learning_rate": 4.163281884646629e-06, + "loss": 2.1795, + "step": 702 + }, + { + "epoch": 0.7704109589041096, + "grad_norm": 1.5300369262695312, + "learning_rate": 4.161251015434607e-06, + "loss": 2.1356, + "step": 703 + }, + { + "epoch": 0.7715068493150685, + "grad_norm": 1.6113280057907104, + "learning_rate": 4.159220146222583e-06, + "loss": 2.157, + "step": 704 + }, + { + "epoch": 0.7726027397260274, + "grad_norm": 1.5458502769470215, + "learning_rate": 4.157189277010561e-06, + "loss": 2.2017, + "step": 705 + }, + { + "epoch": 0.7736986301369863, + "grad_norm": 1.5456187725067139, + "learning_rate": 4.155158407798538e-06, + "loss": 2.1758, + "step": 706 + }, + { + "epoch": 0.7747945205479452, + "grad_norm": 1.6051454544067383, + "learning_rate": 4.153127538586516e-06, + "loss": 2.183, + "step": 707 + }, + { + "epoch": 0.7758904109589041, + "grad_norm": 1.565788745880127, + "learning_rate": 4.151096669374492e-06, + "loss": 2.1538, + "step": 708 + }, + { + "epoch": 0.776986301369863, + "grad_norm": 1.7484439611434937, + "learning_rate": 4.1490658001624705e-06, + "loss": 2.169, + "step": 709 + }, + { + "epoch": 0.7780821917808219, + "grad_norm": 1.7206509113311768, + "learning_rate": 4.147034930950447e-06, + "loss": 2.1661, + "step": 710 + }, + { + "epoch": 0.7780821917808219, + "eval_loss": 2.307927131652832, + "eval_runtime": 37.5145, + "eval_samples_per_second": 26.656, + "eval_steps_per_second": 1.12, + "step": 710 + }, + { + "epoch": 0.7791780821917809, + "grad_norm": 1.8865350484848022, + "learning_rate": 4.145004061738425e-06, + "loss": 2.093, + "step": 711 + }, + { + "epoch": 0.7802739726027397, + "grad_norm": 1.935110092163086, + "learning_rate": 4.1429731925264015e-06, + "loss": 2.1298, + "step": 712 + }, + { + "epoch": 0.7813698630136986, + "grad_norm": 1.6938892602920532, + "learning_rate": 4.140942323314379e-06, + "loss": 2.1508, + "step": 713 + }, + { + "epoch": 0.7824657534246575, + "grad_norm": 1.5774824619293213, + "learning_rate": 4.138911454102356e-06, + "loss": 2.1808, + "step": 714 + }, + { + "epoch": 0.7835616438356164, + "grad_norm": 1.8637802600860596, + "learning_rate": 4.136880584890333e-06, + "loss": 2.1822, + "step": 715 + }, + { + "epoch": 0.7846575342465754, + "grad_norm": 1.6216785907745361, + "learning_rate": 4.1348497156783106e-06, + "loss": 2.1807, + "step": 716 + }, + { + "epoch": 0.7857534246575343, + "grad_norm": 1.7688151597976685, + "learning_rate": 4.132818846466288e-06, + "loss": 2.1967, + "step": 717 + }, + { + "epoch": 0.7868493150684932, + "grad_norm": 1.6824712753295898, + "learning_rate": 4.130787977254265e-06, + "loss": 2.1952, + "step": 718 + }, + { + "epoch": 0.787945205479452, + "grad_norm": 1.911048412322998, + "learning_rate": 4.128757108042242e-06, + "loss": 2.2408, + "step": 719 + }, + { + "epoch": 0.7890410958904109, + "grad_norm": 1.62872314453125, + "learning_rate": 4.12672623883022e-06, + "loss": 2.1647, + "step": 720 + }, + { + "epoch": 0.7890410958904109, + "eval_loss": 2.3071532249450684, + "eval_runtime": 37.6838, + "eval_samples_per_second": 26.537, + "eval_steps_per_second": 1.115, + "step": 720 + }, + { + "epoch": 0.7901369863013699, + "grad_norm": 2.080674171447754, + "learning_rate": 4.124695369618197e-06, + "loss": 2.1614, + "step": 721 + }, + { + "epoch": 0.7912328767123288, + "grad_norm": 1.7225316762924194, + "learning_rate": 4.122664500406174e-06, + "loss": 2.1784, + "step": 722 + }, + { + "epoch": 0.7923287671232877, + "grad_norm": 1.838957667350769, + "learning_rate": 4.1206336311941515e-06, + "loss": 2.1281, + "step": 723 + }, + { + "epoch": 0.7934246575342466, + "grad_norm": 1.9360777139663696, + "learning_rate": 4.118602761982129e-06, + "loss": 2.1168, + "step": 724 + }, + { + "epoch": 0.7945205479452054, + "grad_norm": 1.7712862491607666, + "learning_rate": 4.116571892770106e-06, + "loss": 2.2004, + "step": 725 + }, + { + "epoch": 0.7956164383561644, + "grad_norm": 1.6759192943572998, + "learning_rate": 4.114541023558083e-06, + "loss": 2.1588, + "step": 726 + }, + { + "epoch": 0.7967123287671233, + "grad_norm": 1.8396296501159668, + "learning_rate": 4.112510154346061e-06, + "loss": 2.1777, + "step": 727 + }, + { + "epoch": 0.7978082191780822, + "grad_norm": 1.6775752305984497, + "learning_rate": 4.110479285134038e-06, + "loss": 2.1923, + "step": 728 + }, + { + "epoch": 0.7989041095890411, + "grad_norm": 1.7111576795578003, + "learning_rate": 4.108448415922015e-06, + "loss": 2.1854, + "step": 729 + }, + { + "epoch": 0.8, + "grad_norm": 1.7400974035263062, + "learning_rate": 4.1064175467099924e-06, + "loss": 2.0897, + "step": 730 + }, + { + "epoch": 0.8, + "eval_loss": 2.3070170879364014, + "eval_runtime": 37.7602, + "eval_samples_per_second": 26.483, + "eval_steps_per_second": 1.112, + "step": 730 + }, + { + "epoch": 0.8010958904109589, + "grad_norm": 1.9166353940963745, + "learning_rate": 4.10438667749797e-06, + "loss": 2.2417, + "step": 731 + }, + { + "epoch": 0.8021917808219178, + "grad_norm": 1.7914725542068481, + "learning_rate": 4.102355808285947e-06, + "loss": 2.1677, + "step": 732 + }, + { + "epoch": 0.8032876712328767, + "grad_norm": 1.7602828741073608, + "learning_rate": 4.100324939073924e-06, + "loss": 2.1623, + "step": 733 + }, + { + "epoch": 0.8043835616438356, + "grad_norm": 1.6864640712738037, + "learning_rate": 4.0982940698619015e-06, + "loss": 2.1748, + "step": 734 + }, + { + "epoch": 0.8054794520547945, + "grad_norm": 1.564070463180542, + "learning_rate": 4.096263200649879e-06, + "loss": 2.1458, + "step": 735 + }, + { + "epoch": 0.8065753424657535, + "grad_norm": 1.9004014730453491, + "learning_rate": 4.094232331437856e-06, + "loss": 2.1536, + "step": 736 + }, + { + "epoch": 0.8076712328767123, + "grad_norm": 1.6670420169830322, + "learning_rate": 4.092201462225833e-06, + "loss": 2.1809, + "step": 737 + }, + { + "epoch": 0.8087671232876712, + "grad_norm": 1.6553101539611816, + "learning_rate": 4.090170593013811e-06, + "loss": 2.1343, + "step": 738 + }, + { + "epoch": 0.8098630136986301, + "grad_norm": 1.9083986282348633, + "learning_rate": 4.088139723801787e-06, + "loss": 2.1455, + "step": 739 + }, + { + "epoch": 0.810958904109589, + "grad_norm": 1.6069375276565552, + "learning_rate": 4.086108854589765e-06, + "loss": 2.1621, + "step": 740 + }, + { + "epoch": 0.810958904109589, + "eval_loss": 2.3045499324798584, + "eval_runtime": 37.7804, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 740 + }, + { + "epoch": 0.812054794520548, + "grad_norm": 1.6774693727493286, + "learning_rate": 4.084077985377742e-06, + "loss": 2.2079, + "step": 741 + }, + { + "epoch": 0.8131506849315069, + "grad_norm": 1.772355556488037, + "learning_rate": 4.08204711616572e-06, + "loss": 2.1276, + "step": 742 + }, + { + "epoch": 0.8142465753424658, + "grad_norm": 1.6998035907745361, + "learning_rate": 4.080016246953696e-06, + "loss": 2.1903, + "step": 743 + }, + { + "epoch": 0.8153424657534246, + "grad_norm": 1.850844383239746, + "learning_rate": 4.077985377741674e-06, + "loss": 2.1965, + "step": 744 + }, + { + "epoch": 0.8164383561643835, + "grad_norm": 1.5916879177093506, + "learning_rate": 4.075954508529651e-06, + "loss": 2.1982, + "step": 745 + }, + { + "epoch": 0.8175342465753425, + "grad_norm": 1.8195812702178955, + "learning_rate": 4.073923639317629e-06, + "loss": 2.1507, + "step": 746 + }, + { + "epoch": 0.8186301369863014, + "grad_norm": 1.5145621299743652, + "learning_rate": 4.071892770105605e-06, + "loss": 2.1236, + "step": 747 + }, + { + "epoch": 0.8197260273972603, + "grad_norm": 1.5750877857208252, + "learning_rate": 4.069861900893583e-06, + "loss": 2.142, + "step": 748 + }, + { + "epoch": 0.8208219178082192, + "grad_norm": 1.708666443824768, + "learning_rate": 4.06783103168156e-06, + "loss": 2.1577, + "step": 749 + }, + { + "epoch": 0.821917808219178, + "grad_norm": 1.5988762378692627, + "learning_rate": 4.065800162469538e-06, + "loss": 2.1795, + "step": 750 + }, + { + "epoch": 0.821917808219178, + "eval_loss": 2.3052196502685547, + "eval_runtime": 37.7983, + "eval_samples_per_second": 26.456, + "eval_steps_per_second": 1.111, + "step": 750 + }, + { + "epoch": 0.823013698630137, + "grad_norm": 1.608738899230957, + "learning_rate": 4.063769293257514e-06, + "loss": 2.1578, + "step": 751 + }, + { + "epoch": 0.8241095890410959, + "grad_norm": 1.5578334331512451, + "learning_rate": 4.061738424045492e-06, + "loss": 2.1724, + "step": 752 + }, + { + "epoch": 0.8252054794520548, + "grad_norm": 1.6957333087921143, + "learning_rate": 4.059707554833469e-06, + "loss": 2.1975, + "step": 753 + }, + { + "epoch": 0.8263013698630137, + "grad_norm": 1.691512107849121, + "learning_rate": 4.057676685621446e-06, + "loss": 2.1705, + "step": 754 + }, + { + "epoch": 0.8273972602739726, + "grad_norm": 1.6061747074127197, + "learning_rate": 4.0556458164094235e-06, + "loss": 2.1031, + "step": 755 + }, + { + "epoch": 0.8284931506849315, + "grad_norm": 1.6074447631835938, + "learning_rate": 4.053614947197401e-06, + "loss": 2.1738, + "step": 756 + }, + { + "epoch": 0.8295890410958904, + "grad_norm": 1.9722917079925537, + "learning_rate": 4.051584077985378e-06, + "loss": 2.2043, + "step": 757 + }, + { + "epoch": 0.8306849315068493, + "grad_norm": 1.597692847251892, + "learning_rate": 4.049553208773355e-06, + "loss": 2.1621, + "step": 758 + }, + { + "epoch": 0.8317808219178082, + "grad_norm": 1.85458242893219, + "learning_rate": 4.047522339561333e-06, + "loss": 2.1504, + "step": 759 + }, + { + "epoch": 0.8328767123287671, + "grad_norm": 1.8328344821929932, + "learning_rate": 4.04549147034931e-06, + "loss": 2.2541, + "step": 760 + }, + { + "epoch": 0.8328767123287671, + "eval_loss": 2.3033809661865234, + "eval_runtime": 37.8088, + "eval_samples_per_second": 26.449, + "eval_steps_per_second": 1.111, + "step": 760 + }, + { + "epoch": 0.8339726027397261, + "grad_norm": 1.6544418334960938, + "learning_rate": 4.043460601137287e-06, + "loss": 2.1605, + "step": 761 + }, + { + "epoch": 0.8350684931506849, + "grad_norm": 1.9443641901016235, + "learning_rate": 4.041429731925264e-06, + "loss": 2.1621, + "step": 762 + }, + { + "epoch": 0.8361643835616438, + "grad_norm": 1.731277346611023, + "learning_rate": 4.039398862713242e-06, + "loss": 2.2168, + "step": 763 + }, + { + "epoch": 0.8372602739726027, + "grad_norm": 1.671072244644165, + "learning_rate": 4.037367993501219e-06, + "loss": 2.1612, + "step": 764 + }, + { + "epoch": 0.8383561643835616, + "grad_norm": 1.8777449131011963, + "learning_rate": 4.035337124289196e-06, + "loss": 2.1339, + "step": 765 + }, + { + "epoch": 0.8394520547945206, + "grad_norm": 1.5941543579101562, + "learning_rate": 4.0333062550771735e-06, + "loss": 2.1545, + "step": 766 + }, + { + "epoch": 0.8405479452054795, + "grad_norm": 1.9245301485061646, + "learning_rate": 4.031275385865151e-06, + "loss": 2.187, + "step": 767 + }, + { + "epoch": 0.8416438356164384, + "grad_norm": 1.69217848777771, + "learning_rate": 4.029244516653128e-06, + "loss": 2.1168, + "step": 768 + }, + { + "epoch": 0.8427397260273972, + "grad_norm": 1.7042242288589478, + "learning_rate": 4.027213647441105e-06, + "loss": 2.1353, + "step": 769 + }, + { + "epoch": 0.8438356164383561, + "grad_norm": 1.6391927003860474, + "learning_rate": 4.025182778229083e-06, + "loss": 2.1568, + "step": 770 + }, + { + "epoch": 0.8438356164383561, + "eval_loss": 2.310056686401367, + "eval_runtime": 37.7573, + "eval_samples_per_second": 26.485, + "eval_steps_per_second": 1.112, + "step": 770 + }, + { + "epoch": 0.8449315068493151, + "grad_norm": 1.7833209037780762, + "learning_rate": 4.02315190901706e-06, + "loss": 2.1664, + "step": 771 + }, + { + "epoch": 0.846027397260274, + "grad_norm": 1.8203891515731812, + "learning_rate": 4.021121039805037e-06, + "loss": 2.2342, + "step": 772 + }, + { + "epoch": 0.8471232876712329, + "grad_norm": 1.572940707206726, + "learning_rate": 4.0190901705930144e-06, + "loss": 2.1418, + "step": 773 + }, + { + "epoch": 0.8482191780821918, + "grad_norm": 1.6871955394744873, + "learning_rate": 4.017059301380992e-06, + "loss": 2.2023, + "step": 774 + }, + { + "epoch": 0.8493150684931506, + "grad_norm": 1.6343133449554443, + "learning_rate": 4.015028432168969e-06, + "loss": 2.1302, + "step": 775 + }, + { + "epoch": 0.8504109589041096, + "grad_norm": 1.4492065906524658, + "learning_rate": 4.012997562956945e-06, + "loss": 2.1427, + "step": 776 + }, + { + "epoch": 0.8515068493150685, + "grad_norm": 1.6599146127700806, + "learning_rate": 4.0109666937449235e-06, + "loss": 2.1479, + "step": 777 + }, + { + "epoch": 0.8526027397260274, + "grad_norm": 1.6836353540420532, + "learning_rate": 4.0089358245329e-06, + "loss": 2.148, + "step": 778 + }, + { + "epoch": 0.8536986301369863, + "grad_norm": 1.5977990627288818, + "learning_rate": 4.006904955320878e-06, + "loss": 2.1946, + "step": 779 + }, + { + "epoch": 0.8547945205479452, + "grad_norm": 1.6259957551956177, + "learning_rate": 4.0048740861088545e-06, + "loss": 2.1761, + "step": 780 + }, + { + "epoch": 0.8547945205479452, + "eval_loss": 2.303910970687866, + "eval_runtime": 37.7728, + "eval_samples_per_second": 26.474, + "eval_steps_per_second": 1.112, + "step": 780 + }, + { + "epoch": 0.8558904109589041, + "grad_norm": 1.6990300416946411, + "learning_rate": 4.002843216896833e-06, + "loss": 2.1309, + "step": 781 + }, + { + "epoch": 0.856986301369863, + "grad_norm": 1.6556029319763184, + "learning_rate": 4.000812347684809e-06, + "loss": 2.151, + "step": 782 + }, + { + "epoch": 0.8580821917808219, + "grad_norm": 1.5659290552139282, + "learning_rate": 3.998781478472787e-06, + "loss": 2.1317, + "step": 783 + }, + { + "epoch": 0.8591780821917808, + "grad_norm": 1.6553704738616943, + "learning_rate": 3.996750609260764e-06, + "loss": 2.2344, + "step": 784 + }, + { + "epoch": 0.8602739726027397, + "grad_norm": 1.6608418226242065, + "learning_rate": 3.994719740048742e-06, + "loss": 2.1807, + "step": 785 + }, + { + "epoch": 0.8613698630136987, + "grad_norm": 1.5862030982971191, + "learning_rate": 3.992688870836718e-06, + "loss": 2.1351, + "step": 786 + }, + { + "epoch": 0.8624657534246576, + "grad_norm": 1.767526388168335, + "learning_rate": 3.990658001624696e-06, + "loss": 2.1856, + "step": 787 + }, + { + "epoch": 0.8635616438356164, + "grad_norm": 1.6902577877044678, + "learning_rate": 3.988627132412673e-06, + "loss": 2.1592, + "step": 788 + }, + { + "epoch": 0.8646575342465753, + "grad_norm": 1.7760623693466187, + "learning_rate": 3.98659626320065e-06, + "loss": 2.1645, + "step": 789 + }, + { + "epoch": 0.8657534246575342, + "grad_norm": 1.6791038513183594, + "learning_rate": 3.984565393988627e-06, + "loss": 2.1611, + "step": 790 + }, + { + "epoch": 0.8657534246575342, + "eval_loss": 2.3027942180633545, + "eval_runtime": 37.736, + "eval_samples_per_second": 26.5, + "eval_steps_per_second": 1.113, + "step": 790 + }, + { + "epoch": 0.8668493150684932, + "grad_norm": 1.6695773601531982, + "learning_rate": 3.9825345247766046e-06, + "loss": 2.1747, + "step": 791 + }, + { + "epoch": 0.8679452054794521, + "grad_norm": 1.8736445903778076, + "learning_rate": 3.980503655564582e-06, + "loss": 2.1806, + "step": 792 + }, + { + "epoch": 0.869041095890411, + "grad_norm": 1.6695518493652344, + "learning_rate": 3.978472786352559e-06, + "loss": 2.1631, + "step": 793 + }, + { + "epoch": 0.8701369863013698, + "grad_norm": 1.6782366037368774, + "learning_rate": 3.976441917140536e-06, + "loss": 2.2175, + "step": 794 + }, + { + "epoch": 0.8712328767123287, + "grad_norm": 1.678619623184204, + "learning_rate": 3.974411047928514e-06, + "loss": 2.1883, + "step": 795 + }, + { + "epoch": 0.8723287671232877, + "grad_norm": 1.6274909973144531, + "learning_rate": 3.972380178716491e-06, + "loss": 2.1866, + "step": 796 + }, + { + "epoch": 0.8734246575342466, + "grad_norm": 1.5609477758407593, + "learning_rate": 3.970349309504468e-06, + "loss": 2.1257, + "step": 797 + }, + { + "epoch": 0.8745205479452055, + "grad_norm": 1.7419973611831665, + "learning_rate": 3.9683184402924455e-06, + "loss": 2.1253, + "step": 798 + }, + { + "epoch": 0.8756164383561644, + "grad_norm": 1.6296330690383911, + "learning_rate": 3.966287571080423e-06, + "loss": 2.1755, + "step": 799 + }, + { + "epoch": 0.8767123287671232, + "grad_norm": 1.5995864868164062, + "learning_rate": 3.9642567018684e-06, + "loss": 2.169, + "step": 800 + }, + { + "epoch": 0.8767123287671232, + "eval_loss": 2.301098108291626, + "eval_runtime": 37.7893, + "eval_samples_per_second": 26.463, + "eval_steps_per_second": 1.111, + "step": 800 + }, + { + "epoch": 0.8778082191780822, + "grad_norm": 1.653707504272461, + "learning_rate": 3.962225832656377e-06, + "loss": 2.115, + "step": 801 + }, + { + "epoch": 0.8789041095890411, + "grad_norm": 1.5908581018447876, + "learning_rate": 3.960194963444355e-06, + "loss": 2.1659, + "step": 802 + }, + { + "epoch": 0.88, + "grad_norm": 1.6865545511245728, + "learning_rate": 3.958164094232332e-06, + "loss": 2.1666, + "step": 803 + }, + { + "epoch": 0.8810958904109589, + "grad_norm": 1.6234707832336426, + "learning_rate": 3.956133225020309e-06, + "loss": 2.1594, + "step": 804 + }, + { + "epoch": 0.8821917808219178, + "grad_norm": 1.7375986576080322, + "learning_rate": 3.954102355808286e-06, + "loss": 2.1596, + "step": 805 + }, + { + "epoch": 0.8832876712328767, + "grad_norm": 1.6529247760772705, + "learning_rate": 3.952071486596264e-06, + "loss": 2.1495, + "step": 806 + }, + { + "epoch": 0.8843835616438356, + "grad_norm": 1.517357349395752, + "learning_rate": 3.950040617384241e-06, + "loss": 2.1394, + "step": 807 + }, + { + "epoch": 0.8854794520547945, + "grad_norm": 1.7367429733276367, + "learning_rate": 3.948009748172218e-06, + "loss": 2.206, + "step": 808 + }, + { + "epoch": 0.8865753424657534, + "grad_norm": 1.5467783212661743, + "learning_rate": 3.9459788789601955e-06, + "loss": 2.1672, + "step": 809 + }, + { + "epoch": 0.8876712328767123, + "grad_norm": 1.5647791624069214, + "learning_rate": 3.943948009748173e-06, + "loss": 2.1695, + "step": 810 + }, + { + "epoch": 0.8876712328767123, + "eval_loss": 2.2975995540618896, + "eval_runtime": 37.4912, + "eval_samples_per_second": 26.673, + "eval_steps_per_second": 1.12, + "step": 810 + }, + { + "epoch": 0.8887671232876713, + "grad_norm": 1.597876787185669, + "learning_rate": 3.94191714053615e-06, + "loss": 2.1808, + "step": 811 + }, + { + "epoch": 0.8898630136986302, + "grad_norm": 1.7645189762115479, + "learning_rate": 3.939886271324127e-06, + "loss": 2.1291, + "step": 812 + }, + { + "epoch": 0.890958904109589, + "grad_norm": 1.5209035873413086, + "learning_rate": 3.937855402112104e-06, + "loss": 2.1269, + "step": 813 + }, + { + "epoch": 0.8920547945205479, + "grad_norm": 1.5804905891418457, + "learning_rate": 3.935824532900082e-06, + "loss": 2.1751, + "step": 814 + }, + { + "epoch": 0.8931506849315068, + "grad_norm": 1.687865972518921, + "learning_rate": 3.933793663688058e-06, + "loss": 2.1917, + "step": 815 + }, + { + "epoch": 0.8942465753424658, + "grad_norm": 1.6767957210540771, + "learning_rate": 3.9317627944760365e-06, + "loss": 2.118, + "step": 816 + }, + { + "epoch": 0.8953424657534247, + "grad_norm": 1.657046914100647, + "learning_rate": 3.929731925264013e-06, + "loss": 2.1479, + "step": 817 + }, + { + "epoch": 0.8964383561643836, + "grad_norm": 1.5638110637664795, + "learning_rate": 3.927701056051991e-06, + "loss": 2.1565, + "step": 818 + }, + { + "epoch": 0.8975342465753424, + "grad_norm": 1.6865980625152588, + "learning_rate": 3.9256701868399674e-06, + "loss": 2.1382, + "step": 819 + }, + { + "epoch": 0.8986301369863013, + "grad_norm": 1.6738817691802979, + "learning_rate": 3.9236393176279456e-06, + "loss": 2.1257, + "step": 820 + }, + { + "epoch": 0.8986301369863013, + "eval_loss": 2.2953474521636963, + "eval_runtime": 37.7016, + "eval_samples_per_second": 26.524, + "eval_steps_per_second": 1.114, + "step": 820 + }, + { + "epoch": 0.8997260273972603, + "grad_norm": 1.568678617477417, + "learning_rate": 3.921608448415922e-06, + "loss": 2.169, + "step": 821 + }, + { + "epoch": 0.9008219178082192, + "grad_norm": 1.6803818941116333, + "learning_rate": 3.9195775792039e-06, + "loss": 2.1489, + "step": 822 + }, + { + "epoch": 0.9019178082191781, + "grad_norm": 1.7010858058929443, + "learning_rate": 3.9175467099918765e-06, + "loss": 2.1846, + "step": 823 + }, + { + "epoch": 0.903013698630137, + "grad_norm": 1.5782372951507568, + "learning_rate": 3.915515840779855e-06, + "loss": 2.1788, + "step": 824 + }, + { + "epoch": 0.9041095890410958, + "grad_norm": 1.7804685831069946, + "learning_rate": 3.913484971567831e-06, + "loss": 2.1058, + "step": 825 + }, + { + "epoch": 0.9052054794520548, + "grad_norm": 1.5909292697906494, + "learning_rate": 3.911454102355808e-06, + "loss": 2.1716, + "step": 826 + }, + { + "epoch": 0.9063013698630137, + "grad_norm": 1.7001307010650635, + "learning_rate": 3.909423233143786e-06, + "loss": 2.1647, + "step": 827 + }, + { + "epoch": 0.9073972602739726, + "grad_norm": 1.5576646327972412, + "learning_rate": 3.907392363931763e-06, + "loss": 2.2174, + "step": 828 + }, + { + "epoch": 0.9084931506849315, + "grad_norm": 1.6621390581130981, + "learning_rate": 3.90536149471974e-06, + "loss": 2.2344, + "step": 829 + }, + { + "epoch": 0.9095890410958904, + "grad_norm": 1.592091679573059, + "learning_rate": 3.9033306255077175e-06, + "loss": 2.2501, + "step": 830 + }, + { + "epoch": 0.9095890410958904, + "eval_loss": 2.294229507446289, + "eval_runtime": 37.7195, + "eval_samples_per_second": 26.512, + "eval_steps_per_second": 1.113, + "step": 830 + }, + { + "epoch": 0.9106849315068493, + "grad_norm": 1.5698535442352295, + "learning_rate": 3.901299756295695e-06, + "loss": 2.1534, + "step": 831 + }, + { + "epoch": 0.9117808219178082, + "grad_norm": 1.7159922122955322, + "learning_rate": 3.899268887083672e-06, + "loss": 2.1601, + "step": 832 + }, + { + "epoch": 0.9128767123287671, + "grad_norm": 1.5601210594177246, + "learning_rate": 3.897238017871649e-06, + "loss": 2.1598, + "step": 833 + }, + { + "epoch": 0.913972602739726, + "grad_norm": 1.5987435579299927, + "learning_rate": 3.8952071486596266e-06, + "loss": 2.1307, + "step": 834 + }, + { + "epoch": 0.915068493150685, + "grad_norm": 1.7568409442901611, + "learning_rate": 3.893176279447604e-06, + "loss": 2.2209, + "step": 835 + }, + { + "epoch": 0.9161643835616439, + "grad_norm": 1.5044634342193604, + "learning_rate": 3.891145410235581e-06, + "loss": 2.1699, + "step": 836 + }, + { + "epoch": 0.9172602739726028, + "grad_norm": 1.7287421226501465, + "learning_rate": 3.889114541023558e-06, + "loss": 2.1489, + "step": 837 + }, + { + "epoch": 0.9183561643835616, + "grad_norm": 1.6595203876495361, + "learning_rate": 3.887083671811536e-06, + "loss": 2.1167, + "step": 838 + }, + { + "epoch": 0.9194520547945205, + "grad_norm": 1.5620248317718506, + "learning_rate": 3.885052802599513e-06, + "loss": 2.2119, + "step": 839 + }, + { + "epoch": 0.9205479452054794, + "grad_norm": 1.6492353677749634, + "learning_rate": 3.88302193338749e-06, + "loss": 2.1725, + "step": 840 + }, + { + "epoch": 0.9205479452054794, + "eval_loss": 2.2956349849700928, + "eval_runtime": 37.779, + "eval_samples_per_second": 26.47, + "eval_steps_per_second": 1.112, + "step": 840 + }, + { + "epoch": 0.9216438356164384, + "grad_norm": 1.6601520776748657, + "learning_rate": 3.8809910641754675e-06, + "loss": 2.1532, + "step": 841 + }, + { + "epoch": 0.9227397260273973, + "grad_norm": 1.6723994016647339, + "learning_rate": 3.878960194963445e-06, + "loss": 2.1931, + "step": 842 + }, + { + "epoch": 0.9238356164383562, + "grad_norm": 1.7594473361968994, + "learning_rate": 3.876929325751422e-06, + "loss": 2.1219, + "step": 843 + }, + { + "epoch": 0.924931506849315, + "grad_norm": 1.546398401260376, + "learning_rate": 3.874898456539399e-06, + "loss": 2.1679, + "step": 844 + }, + { + "epoch": 0.9260273972602739, + "grad_norm": 1.8171741962432861, + "learning_rate": 3.872867587327377e-06, + "loss": 2.1434, + "step": 845 + }, + { + "epoch": 0.9271232876712329, + "grad_norm": 1.703196406364441, + "learning_rate": 3.870836718115354e-06, + "loss": 2.1574, + "step": 846 + }, + { + "epoch": 0.9282191780821918, + "grad_norm": 1.6290628910064697, + "learning_rate": 3.868805848903331e-06, + "loss": 2.1657, + "step": 847 + }, + { + "epoch": 0.9293150684931507, + "grad_norm": 1.5868806838989258, + "learning_rate": 3.8667749796913084e-06, + "loss": 2.1659, + "step": 848 + }, + { + "epoch": 0.9304109589041096, + "grad_norm": 1.551581621170044, + "learning_rate": 3.864744110479286e-06, + "loss": 2.1916, + "step": 849 + }, + { + "epoch": 0.9315068493150684, + "grad_norm": 1.6634963750839233, + "learning_rate": 3.862713241267262e-06, + "loss": 2.2027, + "step": 850 + }, + { + "epoch": 0.9315068493150684, + "eval_loss": 2.2988946437835693, + "eval_runtime": 37.7841, + "eval_samples_per_second": 26.466, + "eval_steps_per_second": 1.112, + "step": 850 + }, + { + "epoch": 0.9326027397260274, + "grad_norm": 1.6138510704040527, + "learning_rate": 3.86068237205524e-06, + "loss": 2.1634, + "step": 851 + }, + { + "epoch": 0.9336986301369863, + "grad_norm": 1.6294615268707275, + "learning_rate": 3.858651502843217e-06, + "loss": 2.1175, + "step": 852 + }, + { + "epoch": 0.9347945205479452, + "grad_norm": 1.7515549659729004, + "learning_rate": 3.856620633631195e-06, + "loss": 2.1562, + "step": 853 + }, + { + "epoch": 0.9358904109589041, + "grad_norm": 1.6304186582565308, + "learning_rate": 3.854589764419171e-06, + "loss": 2.2256, + "step": 854 + }, + { + "epoch": 0.936986301369863, + "grad_norm": 1.6189597845077515, + "learning_rate": 3.852558895207149e-06, + "loss": 2.1742, + "step": 855 + }, + { + "epoch": 0.938082191780822, + "grad_norm": 1.665515661239624, + "learning_rate": 3.850528025995126e-06, + "loss": 2.1481, + "step": 856 + }, + { + "epoch": 0.9391780821917808, + "grad_norm": 1.7066419124603271, + "learning_rate": 3.848497156783104e-06, + "loss": 2.1522, + "step": 857 + }, + { + "epoch": 0.9402739726027397, + "grad_norm": 1.884599208831787, + "learning_rate": 3.84646628757108e-06, + "loss": 2.1493, + "step": 858 + }, + { + "epoch": 0.9413698630136986, + "grad_norm": 1.6224873065948486, + "learning_rate": 3.8444354183590585e-06, + "loss": 2.1461, + "step": 859 + }, + { + "epoch": 0.9424657534246575, + "grad_norm": 1.916416049003601, + "learning_rate": 3.842404549147035e-06, + "loss": 2.1265, + "step": 860 + }, + { + "epoch": 0.9424657534246575, + "eval_loss": 2.297154426574707, + "eval_runtime": 37.77, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 860 + }, + { + "epoch": 0.9435616438356165, + "grad_norm": 1.5281857252120972, + "learning_rate": 3.840373679935013e-06, + "loss": 2.1476, + "step": 861 + }, + { + "epoch": 0.9446575342465754, + "grad_norm": 1.7905207872390747, + "learning_rate": 3.8383428107229894e-06, + "loss": 2.1421, + "step": 862 + }, + { + "epoch": 0.9457534246575342, + "grad_norm": 1.8282849788665771, + "learning_rate": 3.836311941510967e-06, + "loss": 2.1807, + "step": 863 + }, + { + "epoch": 0.9468493150684931, + "grad_norm": 1.7861522436141968, + "learning_rate": 3.834281072298944e-06, + "loss": 2.1337, + "step": 864 + }, + { + "epoch": 0.947945205479452, + "grad_norm": 1.6107250452041626, + "learning_rate": 3.832250203086921e-06, + "loss": 2.1911, + "step": 865 + }, + { + "epoch": 0.949041095890411, + "grad_norm": 1.657909631729126, + "learning_rate": 3.8302193338748985e-06, + "loss": 2.2257, + "step": 866 + }, + { + "epoch": 0.9501369863013699, + "grad_norm": 1.8969722986221313, + "learning_rate": 3.828188464662876e-06, + "loss": 2.1778, + "step": 867 + }, + { + "epoch": 0.9512328767123288, + "grad_norm": 1.6498360633850098, + "learning_rate": 3.826157595450853e-06, + "loss": 2.127, + "step": 868 + }, + { + "epoch": 0.9523287671232876, + "grad_norm": 1.5750585794448853, + "learning_rate": 3.82412672623883e-06, + "loss": 2.1716, + "step": 869 + }, + { + "epoch": 0.9534246575342465, + "grad_norm": 1.6047886610031128, + "learning_rate": 3.822095857026808e-06, + "loss": 2.1354, + "step": 870 + }, + { + "epoch": 0.9534246575342465, + "eval_loss": 2.295785665512085, + "eval_runtime": 37.7419, + "eval_samples_per_second": 26.496, + "eval_steps_per_second": 1.113, + "step": 870 + }, + { + "epoch": 0.9545205479452055, + "grad_norm": 1.6426187753677368, + "learning_rate": 3.820064987814785e-06, + "loss": 2.1375, + "step": 871 + }, + { + "epoch": 0.9556164383561644, + "grad_norm": 1.5118632316589355, + "learning_rate": 3.818034118602762e-06, + "loss": 2.1491, + "step": 872 + }, + { + "epoch": 0.9567123287671233, + "grad_norm": 1.630476474761963, + "learning_rate": 3.8160032493907395e-06, + "loss": 2.1497, + "step": 873 + }, + { + "epoch": 0.9578082191780822, + "grad_norm": 1.6820838451385498, + "learning_rate": 3.813972380178717e-06, + "loss": 2.1746, + "step": 874 + }, + { + "epoch": 0.958904109589041, + "grad_norm": 1.5480482578277588, + "learning_rate": 3.811941510966694e-06, + "loss": 2.1226, + "step": 875 + }, + { + "epoch": 0.96, + "grad_norm": 1.601493239402771, + "learning_rate": 3.809910641754671e-06, + "loss": 2.0934, + "step": 876 + }, + { + "epoch": 0.9610958904109589, + "grad_norm": 1.5756622552871704, + "learning_rate": 3.8078797725426486e-06, + "loss": 2.1679, + "step": 877 + }, + { + "epoch": 0.9621917808219178, + "grad_norm": 1.6811070442199707, + "learning_rate": 3.8058489033306254e-06, + "loss": 2.1407, + "step": 878 + }, + { + "epoch": 0.9632876712328767, + "grad_norm": 1.5748940706253052, + "learning_rate": 3.803818034118603e-06, + "loss": 2.1985, + "step": 879 + }, + { + "epoch": 0.9643835616438357, + "grad_norm": 1.5412342548370361, + "learning_rate": 3.80178716490658e-06, + "loss": 2.1876, + "step": 880 + }, + { + "epoch": 0.9643835616438357, + "eval_loss": 2.291630268096924, + "eval_runtime": 37.7702, + "eval_samples_per_second": 26.476, + "eval_steps_per_second": 1.112, + "step": 880 + }, + { + "epoch": 0.9654794520547946, + "grad_norm": 1.7080684900283813, + "learning_rate": 3.7997562956945577e-06, + "loss": 2.1453, + "step": 881 + }, + { + "epoch": 0.9665753424657534, + "grad_norm": 1.8705110549926758, + "learning_rate": 3.7977254264825345e-06, + "loss": 2.1298, + "step": 882 + }, + { + "epoch": 0.9676712328767123, + "grad_norm": 1.7332537174224854, + "learning_rate": 3.7956945572705122e-06, + "loss": 2.2117, + "step": 883 + }, + { + "epoch": 0.9687671232876712, + "grad_norm": 1.88144052028656, + "learning_rate": 3.793663688058489e-06, + "loss": 2.2005, + "step": 884 + }, + { + "epoch": 0.9698630136986301, + "grad_norm": 1.6263943910598755, + "learning_rate": 3.791632818846467e-06, + "loss": 2.1853, + "step": 885 + }, + { + "epoch": 0.9709589041095891, + "grad_norm": 1.5968067646026611, + "learning_rate": 3.7896019496344436e-06, + "loss": 2.1998, + "step": 886 + }, + { + "epoch": 0.972054794520548, + "grad_norm": 1.7149827480316162, + "learning_rate": 3.787571080422421e-06, + "loss": 2.1855, + "step": 887 + }, + { + "epoch": 0.9731506849315068, + "grad_norm": 1.7375026941299438, + "learning_rate": 3.785540211210398e-06, + "loss": 2.1664, + "step": 888 + }, + { + "epoch": 0.9742465753424657, + "grad_norm": 1.6411423683166504, + "learning_rate": 3.7835093419983755e-06, + "loss": 2.1492, + "step": 889 + }, + { + "epoch": 0.9753424657534246, + "grad_norm": 1.6208266019821167, + "learning_rate": 3.7814784727863527e-06, + "loss": 2.2341, + "step": 890 + }, + { + "epoch": 0.9753424657534246, + "eval_loss": 2.2929985523223877, + "eval_runtime": 37.6693, + "eval_samples_per_second": 26.547, + "eval_steps_per_second": 1.115, + "step": 890 + }, + { + "epoch": 0.9764383561643836, + "grad_norm": 1.6871188879013062, + "learning_rate": 3.77944760357433e-06, + "loss": 2.1985, + "step": 891 + }, + { + "epoch": 0.9775342465753425, + "grad_norm": 1.8475587368011475, + "learning_rate": 3.7774167343623073e-06, + "loss": 2.1423, + "step": 892 + }, + { + "epoch": 0.9786301369863014, + "grad_norm": 1.6115065813064575, + "learning_rate": 3.7753858651502846e-06, + "loss": 2.1944, + "step": 893 + }, + { + "epoch": 0.9797260273972602, + "grad_norm": 1.6232231855392456, + "learning_rate": 3.773354995938262e-06, + "loss": 2.1751, + "step": 894 + }, + { + "epoch": 0.9808219178082191, + "grad_norm": 1.8564099073410034, + "learning_rate": 3.771324126726239e-06, + "loss": 2.2021, + "step": 895 + }, + { + "epoch": 0.9819178082191781, + "grad_norm": 1.7270197868347168, + "learning_rate": 3.7692932575142164e-06, + "loss": 2.1351, + "step": 896 + }, + { + "epoch": 0.983013698630137, + "grad_norm": 1.7187668085098267, + "learning_rate": 3.7672623883021937e-06, + "loss": 2.0856, + "step": 897 + }, + { + "epoch": 0.9841095890410959, + "grad_norm": 1.7766387462615967, + "learning_rate": 3.765231519090171e-06, + "loss": 2.168, + "step": 898 + }, + { + "epoch": 0.9852054794520548, + "grad_norm": 1.8063374757766724, + "learning_rate": 3.7632006498781482e-06, + "loss": 2.1317, + "step": 899 + }, + { + "epoch": 0.9863013698630136, + "grad_norm": 1.7601871490478516, + "learning_rate": 3.761169780666125e-06, + "loss": 2.1742, + "step": 900 + }, + { + "epoch": 0.9863013698630136, + "eval_loss": 2.292574167251587, + "eval_runtime": 37.6888, + "eval_samples_per_second": 26.533, + "eval_steps_per_second": 1.114, + "step": 900 + } + ], + "logging_steps": 1, + "max_steps": 2736, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 2.805578468425728e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/zero_to_fp32.py b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-900/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters)