Upload folder using huggingface_hub

added_tokens.json

@@ -0,0 +1,26 @@
+{
+  "</think>": 151666,
+  "</tool_call>": 151658,
+  "<think>": 151665,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

config.json

@@ -0,0 +1,28 @@
+{
+  "_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.44.2",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.44.2"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7dadc648f42a695e45aad4e7d535478afa740a222c7fa6f23d86a6ccadd00ec3
+size 9979257128

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e99f35631579cb8247afc2c91afb1a6eca8b282369e057dc942a0319c701b883
+size 5252014776

model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15231233024
+  },
+  "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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-00002-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "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-00002-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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-00002-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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.bias": "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"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,223 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,3729 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 5624,
+  "global_step": 16872,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "grad_norm": 64.30284881591797,
+      "learning_rate": 2.9620853080568726e-09,
+      "loss": 3.7905,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 74.81798553466797,
+      "learning_rate": 9.478672985781992e-08,
+      "loss": 4.4137,
+      "step": 32
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 101.4776840209961,
+      "learning_rate": 1.8957345971563984e-07,
+      "loss": 4.2954,
+      "step": 64
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 68.84349822998047,
+      "learning_rate": 2.843601895734597e-07,
+      "loss": 3.376,
+      "step": 96
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 42.47146224975586,
+      "learning_rate": 3.791469194312797e-07,
+      "loss": 1.962,
+      "step": 128
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 46.63043212890625,
+      "learning_rate": 4.7393364928909956e-07,
+      "loss": 1.168,
+      "step": 160
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 2.709139347076416,
+      "learning_rate": 5.687203791469194e-07,
+      "loss": 0.4681,
+      "step": 192
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 3.304001808166504,
+      "learning_rate": 6.635071090047394e-07,
+      "loss": 0.4059,
+      "step": 224
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 2.5251598358154297,
+      "learning_rate": 7.582938388625594e-07,
+      "loss": 0.3999,
+      "step": 256
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 1.9685628414154053,
+      "learning_rate": 8.530805687203792e-07,
+      "loss": 0.3904,
+      "step": 288
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 1.7696113586425781,
+      "learning_rate": 9.478672985781991e-07,
+      "loss": 0.3769,
+      "step": 320
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 1.8694322109222412,
+      "learning_rate": 1.042654028436019e-06,
+      "loss": 0.3718,
+      "step": 352
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 1.7985926866531372,
+      "learning_rate": 1.1374407582938388e-06,
+      "loss": 0.3569,
+      "step": 384
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 1.9462366104125977,
+      "learning_rate": 1.2322274881516587e-06,
+      "loss": 0.3432,
+      "step": 416
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 1.916548490524292,
+      "learning_rate": 1.3270142180094788e-06,
+      "loss": 0.331,
+      "step": 448
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 1.2683026790618896,
+      "learning_rate": 1.4218009478672987e-06,
+      "loss": 0.334,
+      "step": 480
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 1.4642925262451172,
+      "learning_rate": 1.5165876777251187e-06,
+      "loss": 0.3386,
+      "step": 512
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 1.3890858888626099,
+      "learning_rate": 1.6113744075829384e-06,
+      "loss": 0.3073,
+      "step": 544
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 1.4565709829330444,
+      "learning_rate": 1.7061611374407585e-06,
+      "loss": 0.3263,
+      "step": 576
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 1.349706768989563,
+      "learning_rate": 1.8009478672985784e-06,
+      "loss": 0.3165,
+      "step": 608
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 1.7937098741531372,
+      "learning_rate": 1.8957345971563982e-06,
+      "loss": 0.3052,
+      "step": 640
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 1.616604208946228,
+      "learning_rate": 1.990521327014218e-06,
+      "loss": 0.3138,
+      "step": 672
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 1.3185547590255737,
+      "learning_rate": 2.085308056872038e-06,
+      "loss": 0.297,
+      "step": 704
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 1.8201861381530762,
+      "learning_rate": 2.180094786729858e-06,
+      "loss": 0.3072,
+      "step": 736
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 1.1981289386749268,
+      "learning_rate": 2.2748815165876777e-06,
+      "loss": 0.2957,
+      "step": 768
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 1.7006583213806152,
+      "learning_rate": 2.369668246445498e-06,
+      "loss": 0.2941,
+      "step": 800
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 2.1674447059631348,
+      "learning_rate": 2.4644549763033174e-06,
+      "loss": 0.2798,
+      "step": 832
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 1.3364813327789307,
+      "learning_rate": 2.5592417061611373e-06,
+      "loss": 0.2996,
+      "step": 864
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 1.2685174942016602,
+      "learning_rate": 2.6540284360189576e-06,
+      "loss": 0.3027,
+      "step": 896
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 1.3952556848526,
+      "learning_rate": 2.7488151658767775e-06,
+      "loss": 0.2985,
+      "step": 928
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 1.251714825630188,
+      "learning_rate": 2.8436018957345973e-06,
+      "loss": 0.2905,
+      "step": 960
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 1.553152322769165,
+      "learning_rate": 2.938388625592417e-06,
+      "loss": 0.278,
+      "step": 992
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 1.380920171737671,
+      "learning_rate": 3.0331753554502375e-06,
+      "loss": 0.2813,
+      "step": 1024
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 1.5351643562316895,
+      "learning_rate": 3.1279620853080574e-06,
+      "loss": 0.2805,
+      "step": 1056
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 1.4867887496948242,
+      "learning_rate": 3.222748815165877e-06,
+      "loss": 0.2767,
+      "step": 1088
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 1.317229986190796,
+      "learning_rate": 3.3175355450236967e-06,
+      "loss": 0.2859,
+      "step": 1120
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 1.8770791292190552,
+      "learning_rate": 3.412322274881517e-06,
+      "loss": 0.2875,
+      "step": 1152
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 1.4476697444915771,
+      "learning_rate": 3.507109004739337e-06,
+      "loss": 0.2884,
+      "step": 1184
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 1.351965069770813,
+      "learning_rate": 3.6018957345971567e-06,
+      "loss": 0.2802,
+      "step": 1216
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 1.4647209644317627,
+      "learning_rate": 3.6966824644549766e-06,
+      "loss": 0.2703,
+      "step": 1248
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 1.901773452758789,
+      "learning_rate": 3.7914691943127964e-06,
+      "loss": 0.2815,
+      "step": 1280
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 1.139844536781311,
+      "learning_rate": 3.886255924170616e-06,
+      "loss": 0.2658,
+      "step": 1312
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 1.1863261461257935,
+      "learning_rate": 3.981042654028436e-06,
+      "loss": 0.2707,
+      "step": 1344
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 1.2720916271209717,
+      "learning_rate": 4.075829383886256e-06,
+      "loss": 0.2646,
+      "step": 1376
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 1.6161096096038818,
+      "learning_rate": 4.170616113744076e-06,
+      "loss": 0.2748,
+      "step": 1408
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 1.4303381443023682,
+      "learning_rate": 4.265402843601897e-06,
+      "loss": 0.2691,
+      "step": 1440
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 1.401880145072937,
+      "learning_rate": 4.360189573459716e-06,
+      "loss": 0.2699,
+      "step": 1472
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 1.3495467901229858,
+      "learning_rate": 4.4549763033175355e-06,
+      "loss": 0.2772,
+      "step": 1504
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 1.3915464878082275,
+      "learning_rate": 4.549763033175355e-06,
+      "loss": 0.2752,
+      "step": 1536
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 1.3412673473358154,
+      "learning_rate": 4.644549763033176e-06,
+      "loss": 0.2751,
+      "step": 1568
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 1.3777296543121338,
+      "learning_rate": 4.739336492890996e-06,
+      "loss": 0.2717,
+      "step": 1600
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 1.2612873315811157,
+      "learning_rate": 4.834123222748816e-06,
+      "loss": 0.2678,
+      "step": 1632
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 1.2989717721939087,
+      "learning_rate": 4.928909952606635e-06,
+      "loss": 0.2778,
+      "step": 1664
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 1.3525702953338623,
+      "learning_rate": 4.999996575341721e-06,
+      "loss": 0.2719,
+      "step": 1696
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 1.4678899049758911,
+      "learning_rate": 4.999914384012144e-06,
+      "loss": 0.2755,
+      "step": 1728
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 1.2093278169631958,
+      "learning_rate": 4.999722607745944e-06,
+      "loss": 0.2755,
+      "step": 1760
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 1.4915423393249512,
+      "learning_rate": 4.999421254949728e-06,
+      "loss": 0.2686,
+      "step": 1792
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 1.1101861000061035,
+      "learning_rate": 4.999010338833436e-06,
+      "loss": 0.2594,
+      "step": 1824
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 1.3432806730270386,
+      "learning_rate": 4.9984898774097735e-06,
+      "loss": 0.2658,
+      "step": 1856
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 1.2808105945587158,
+      "learning_rate": 4.997859893493414e-06,
+      "loss": 0.2632,
+      "step": 1888
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 1.3815045356750488,
+      "learning_rate": 4.997120414700003e-06,
+      "loss": 0.2557,
+      "step": 1920
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 1.4393643140792847,
+      "learning_rate": 4.996271473444944e-06,
+      "loss": 0.263,
+      "step": 1952
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 1.138375163078308,
+      "learning_rate": 4.995313106941982e-06,
+      "loss": 0.2805,
+      "step": 1984
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 1.6412934064865112,
+      "learning_rate": 4.994245357201568e-06,
+      "loss": 0.2641,
+      "step": 2016
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 1.465922236442566,
+      "learning_rate": 4.9930682710290205e-06,
+      "loss": 0.2637,
+      "step": 2048
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 1.4526797533035278,
+      "learning_rate": 4.991781900022471e-06,
+      "loss": 0.2596,
+      "step": 2080
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 1.504759669303894,
+      "learning_rate": 4.990386300570607e-06,
+      "loss": 0.2633,
+      "step": 2112
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 1.5599263906478882,
+      "learning_rate": 4.988881533850192e-06,
+      "loss": 0.2658,
+      "step": 2144
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 1.1662814617156982,
+      "learning_rate": 4.987267665823392e-06,
+      "loss": 0.2694,
+      "step": 2176
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 1.3952819108963013,
+      "learning_rate": 4.98554476723488e-06,
+      "loss": 0.2449,
+      "step": 2208
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 1.2887946367263794,
+      "learning_rate": 4.983712913608736e-06,
+      "loss": 0.2651,
+      "step": 2240
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 1.5893690586090088,
+      "learning_rate": 4.981772185245135e-06,
+      "loss": 0.2568,
+      "step": 2272
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 1.228550672531128,
+      "learning_rate": 4.979722667216829e-06,
+      "loss": 0.2667,
+      "step": 2304
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 1.2756662368774414,
+      "learning_rate": 4.977564449365415e-06,
+      "loss": 0.2508,
+      "step": 2336
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 1.5225822925567627,
+      "learning_rate": 4.975297626297399e-06,
+      "loss": 0.2691,
+      "step": 2368
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 1.2656946182250977,
+      "learning_rate": 4.972922297380052e-06,
+      "loss": 0.2704,
+      "step": 2400
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 1.3268104791641235,
+      "learning_rate": 4.970438566737043e-06,
+      "loss": 0.2577,
+      "step": 2432
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 1.5536099672317505,
+      "learning_rate": 4.96784654324389e-06,
+      "loss": 0.2578,
+      "step": 2464
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 1.1516194343566895,
+      "learning_rate": 4.965146340523175e-06,
+      "loss": 0.2446,
+      "step": 2496
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 1.1923089027404785,
+      "learning_rate": 4.962338076939569e-06,
+      "loss": 0.2569,
+      "step": 2528
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 1.124197006225586,
+      "learning_rate": 4.959421875594643e-06,
+      "loss": 0.2625,
+      "step": 2560
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 1.680388331413269,
+      "learning_rate": 4.95639786432147e-06,
+      "loss": 0.264,
+      "step": 2592
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 1.3039462566375732,
+      "learning_rate": 4.953266175679023e-06,
+      "loss": 0.2624,
+      "step": 2624
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 1.109054684638977,
+      "learning_rate": 4.9500269469463655e-06,
+      "loss": 0.2548,
+      "step": 2656
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 1.2704750299453735,
+      "learning_rate": 4.94668032011663e-06,
+      "loss": 0.2569,
+      "step": 2688
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 1.1952179670333862,
+      "learning_rate": 4.943226441890794e-06,
+      "loss": 0.2599,
+      "step": 2720
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 1.2229312658309937,
+      "learning_rate": 4.939665463671255e-06,
+      "loss": 0.2577,
+      "step": 2752
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 1.3956924676895142,
+      "learning_rate": 4.935997541555188e-06,
+      "loss": 0.2642,
+      "step": 2784
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 1.116629958152771,
+      "learning_rate": 4.932222836327703e-06,
+      "loss": 0.2587,
+      "step": 2816
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 1.1389435529708862,
+      "learning_rate": 4.928341513454801e-06,
+      "loss": 0.2566,
+      "step": 2848
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 1.3800580501556396,
+      "learning_rate": 4.9243537430761155e-06,
+      "loss": 0.2579,
+      "step": 2880
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 1.3852914571762085,
+      "learning_rate": 4.920259699997461e-06,
+      "loss": 0.2666,
+      "step": 2912
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 1.31257963180542,
+      "learning_rate": 4.916059563683162e-06,
+      "loss": 0.2547,
+      "step": 2944
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 1.599116563796997,
+      "learning_rate": 4.911753518248194e-06,
+      "loss": 0.2612,
+      "step": 2976
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 1.2397140264511108,
+      "learning_rate": 4.907341752450105e-06,
+      "loss": 0.2589,
+      "step": 3008
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 1.3178327083587646,
+      "learning_rate": 4.9028244596807525e-06,
+      "loss": 0.2605,
+      "step": 3040
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 1.7413417100906372,
+      "learning_rate": 4.898201837957811e-06,
+      "loss": 0.2565,
+      "step": 3072
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 1.314085602760315,
+      "learning_rate": 4.893474089916105e-06,
+      "loss": 0.2498,
+      "step": 3104
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 1.1399492025375366,
+      "learning_rate": 4.888641422798719e-06,
+      "loss": 0.2647,
+      "step": 3136
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 1.3332985639572144,
+      "learning_rate": 4.883704048447916e-06,
+      "loss": 0.2594,
+      "step": 3168
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 1.3460063934326172,
+      "learning_rate": 4.87866218329585e-06,
+      "loss": 0.2571,
+      "step": 3200
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 1.5006327629089355,
+      "learning_rate": 4.87351604835508e-06,
+      "loss": 0.2458,
+      "step": 3232
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 1.1781283617019653,
+      "learning_rate": 4.868265869208879e-06,
+      "loss": 0.2452,
+      "step": 3264
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 1.117686152458191,
+      "learning_rate": 4.862911876001348e-06,
+      "loss": 0.2469,
+      "step": 3296
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.9969549775123596,
+      "learning_rate": 4.857454303427328e-06,
+      "loss": 0.2453,
+      "step": 3328
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 1.4894945621490479,
+      "learning_rate": 4.851893390722109e-06,
+      "loss": 0.2457,
+      "step": 3360
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 1.106041431427002,
+      "learning_rate": 4.846229381650946e-06,
+      "loss": 0.2474,
+      "step": 3392
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 1.035601019859314,
+      "learning_rate": 4.840462524498372e-06,
+      "loss": 0.2593,
+      "step": 3424
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 1.7077690362930298,
+      "learning_rate": 4.834593072057313e-06,
+      "loss": 0.2506,
+      "step": 3456
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 1.1017436981201172,
+      "learning_rate": 4.8286212816180124e-06,
+      "loss": 0.2506,
+      "step": 3488
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 1.2720685005187988,
+      "learning_rate": 4.8225474149567434e-06,
+      "loss": 0.2567,
+      "step": 3520
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 1.328189730644226,
+      "learning_rate": 4.816371738324343e-06,
+      "loss": 0.2531,
+      "step": 3552
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 1.2597825527191162,
+      "learning_rate": 4.810094522434534e-06,
+      "loss": 0.246,
+      "step": 3584
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 1.244281530380249,
+      "learning_rate": 4.803716042452063e-06,
+      "loss": 0.2433,
+      "step": 3616
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 1.4658986330032349,
+      "learning_rate": 4.797236577980634e-06,
+      "loss": 0.2496,
+      "step": 3648
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 1.4121670722961426,
+      "learning_rate": 4.7906564130506575e-06,
+      "loss": 0.2531,
+      "step": 3680
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 1.1751240491867065,
+      "learning_rate": 4.783975836106791e-06,
+      "loss": 0.2515,
+      "step": 3712
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 1.2011898756027222,
+      "learning_rate": 4.777195139995308e-06,
+      "loss": 0.2453,
+      "step": 3744
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 1.5764689445495605,
+      "learning_rate": 4.770314621951245e-06,
+      "loss": 0.2496,
+      "step": 3776
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 1.4584077596664429,
+      "learning_rate": 4.763334583585388e-06,
+      "loss": 0.2392,
+      "step": 3808
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 1.0098185539245605,
+      "learning_rate": 4.756255330871039e-06,
+      "loss": 0.2393,
+      "step": 3840
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 1.3514459133148193,
+      "learning_rate": 4.749077174130609e-06,
+      "loss": 0.2572,
+      "step": 3872
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 1.3888107538223267,
+      "learning_rate": 4.741800428022014e-06,
+      "loss": 0.2383,
+      "step": 3904
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 1.3402737379074097,
+      "learning_rate": 4.734425411524884e-06,
+      "loss": 0.2556,
+      "step": 3936
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 1.2175307273864746,
+      "learning_rate": 4.726952447926576e-06,
+      "loss": 0.2555,
+      "step": 3968
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 1.386852502822876,
+      "learning_rate": 4.719381864808005e-06,
+      "loss": 0.2503,
+      "step": 4000
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 1.2774380445480347,
+      "learning_rate": 4.711713994029284e-06,
+      "loss": 0.2503,
+      "step": 4032
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 1.177322268486023,
+      "learning_rate": 4.703949171715179e-06,
+      "loss": 0.2574,
+      "step": 4064
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 1.269942283630371,
+      "learning_rate": 4.69608773824037e-06,
+      "loss": 0.2529,
+      "step": 4096
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 1.2209409475326538,
+      "learning_rate": 4.688130038214534e-06,
+      "loss": 0.2536,
+      "step": 4128
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 1.4368942975997925,
+      "learning_rate": 4.6800764204672385e-06,
+      "loss": 0.2378,
+      "step": 4160
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 1.6493048667907715,
+      "learning_rate": 4.671927238032651e-06,
+      "loss": 0.2538,
+      "step": 4192
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 1.038549542427063,
+      "learning_rate": 4.6636828481340594e-06,
+      "loss": 0.2501,
+      "step": 4224
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 1.343204379081726,
+      "learning_rate": 4.655343612168219e-06,
+      "loss": 0.251,
+      "step": 4256
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 1.4020464420318604,
+      "learning_rate": 4.646909895689508e-06,
+      "loss": 0.2564,
+      "step": 4288
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 1.1331307888031006,
+      "learning_rate": 4.638382068393899e-06,
+      "loss": 0.2505,
+      "step": 4320
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 1.3825620412826538,
+      "learning_rate": 4.629760504102761e-06,
+      "loss": 0.2513,
+      "step": 4352
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 1.310570478439331,
+      "learning_rate": 4.621045580746467e-06,
+      "loss": 0.2464,
+      "step": 4384
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 1.15547776222229,
+      "learning_rate": 4.61223768034783e-06,
+      "loss": 0.2515,
+      "step": 4416
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 1.340010404586792,
+      "learning_rate": 4.603337189005354e-06,
+      "loss": 0.2473,
+      "step": 4448
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 1.2413158416748047,
+      "learning_rate": 4.594344496876313e-06,
+      "loss": 0.2354,
+      "step": 4480
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 1.2394189834594727,
+      "learning_rate": 4.585259998159646e-06,
+      "loss": 0.2512,
+      "step": 4512
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 1.2866027355194092,
+      "learning_rate": 4.576084091078677e-06,
+      "loss": 0.2364,
+      "step": 4544
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 1.1080009937286377,
+      "learning_rate": 4.5668171778636585e-06,
+      "loss": 0.2432,
+      "step": 4576
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 1.2469310760498047,
+      "learning_rate": 4.5574596647341414e-06,
+      "loss": 0.256,
+      "step": 4608
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 1.0387507677078247,
+      "learning_rate": 4.548011961881167e-06,
+      "loss": 0.232,
+      "step": 4640
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 1.2382770776748657,
+      "learning_rate": 4.538474483449286e-06,
+      "loss": 0.2552,
+      "step": 4672
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 1.2282336950302124,
+      "learning_rate": 4.528847647518403e-06,
+      "loss": 0.2525,
+      "step": 4704
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 1.4016482830047607,
+      "learning_rate": 4.5191318760854526e-06,
+      "loss": 0.2582,
+      "step": 4736
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 1.3214083909988403,
+      "learning_rate": 4.509327595045898e-06,
+      "loss": 0.2578,
+      "step": 4768
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.9114232063293457,
+      "learning_rate": 4.499435234175065e-06,
+      "loss": 0.2533,
+      "step": 4800
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 1.172450065612793,
+      "learning_rate": 4.4894552271093e-06,
+      "loss": 0.264,
+      "step": 4832
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 1.249770998954773,
+      "learning_rate": 4.4793880113269595e-06,
+      "loss": 0.2389,
+      "step": 4864
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 1.0912755727767944,
+      "learning_rate": 4.469234028129241e-06,
+      "loss": 0.2456,
+      "step": 4896
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 1.1503956317901611,
+      "learning_rate": 4.458993722620827e-06,
+      "loss": 0.2562,
+      "step": 4928
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 1.1564654111862183,
+      "learning_rate": 4.448667543690384e-06,
+      "loss": 0.25,
+      "step": 4960
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 1.271000862121582,
+      "learning_rate": 4.438255943990879e-06,
+      "loss": 0.243,
+      "step": 4992
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 1.0601048469543457,
+      "learning_rate": 4.427759379919739e-06,
+      "loss": 0.2397,
+      "step": 5024
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 1.214858889579773,
+      "learning_rate": 4.417178311598845e-06,
+      "loss": 0.2442,
+      "step": 5056
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 1.0516908168792725,
+      "learning_rate": 4.406513202854363e-06,
+      "loss": 0.2467,
+      "step": 5088
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 1.326076865196228,
+      "learning_rate": 4.3957645211964065e-06,
+      "loss": 0.2488,
+      "step": 5120
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 1.173823356628418,
+      "learning_rate": 4.384932737798554e-06,
+      "loss": 0.241,
+      "step": 5152
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 1.4526327848434448,
+      "learning_rate": 4.3740183274771845e-06,
+      "loss": 0.2553,
+      "step": 5184
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 1.2346609830856323,
+      "learning_rate": 4.363021768670668e-06,
+      "loss": 0.242,
+      "step": 5216
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.8957495093345642,
+      "learning_rate": 4.351943543418392e-06,
+      "loss": 0.2444,
+      "step": 5248
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 1.097772479057312,
+      "learning_rate": 4.340784137339632e-06,
+      "loss": 0.2531,
+      "step": 5280
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 1.1537779569625854,
+      "learning_rate": 4.329544039612264e-06,
+      "loss": 0.2507,
+      "step": 5312
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 1.1922253370285034,
+      "learning_rate": 4.318223742951321e-06,
+      "loss": 0.2335,
+      "step": 5344
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 1.1036819219589233,
+      "learning_rate": 4.306823743587394e-06,
+      "loss": 0.2465,
+      "step": 5376
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 1.229779839515686,
+      "learning_rate": 4.295344541244879e-06,
+      "loss": 0.2403,
+      "step": 5408
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 1.4036519527435303,
+      "learning_rate": 4.283786639120074e-06,
+      "loss": 0.254,
+      "step": 5440
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.9732062816619873,
+      "learning_rate": 4.272150543859117e-06,
+      "loss": 0.2517,
+      "step": 5472
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 1.3309801816940308,
+      "learning_rate": 4.260436765535784e-06,
+      "loss": 0.25,
+      "step": 5504
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 1.3353493213653564,
+      "learning_rate": 4.2486458176291176e-06,
+      "loss": 0.2482,
+      "step": 5536
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 1.6585358381271362,
+      "learning_rate": 4.236778217000934e-06,
+      "loss": 0.248,
+      "step": 5568
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.9717461466789246,
+      "learning_rate": 4.224834483873152e-06,
+      "loss": 0.2366,
+      "step": 5600
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.9571962356567383,
+      "learning_rate": 4.2128151418049976e-06,
+      "loss": 0.2404,
+      "step": 5632
+    },
+    {
+      "epoch": 1.01,
+      "grad_norm": 1.0692377090454102,
+      "learning_rate": 4.200720717670048e-06,
+      "loss": 0.2135,
+      "step": 5664
+    },
+    {
+      "epoch": 1.01,
+      "grad_norm": 1.1159001588821411,
+      "learning_rate": 4.188551741633144e-06,
+      "loss": 0.1854,
+      "step": 5696
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 1.4514949321746826,
+      "learning_rate": 4.176308747127136e-06,
+      "loss": 0.2095,
+      "step": 5728
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 1.4603676795959473,
+      "learning_rate": 4.1639922708295176e-06,
+      "loss": 0.2015,
+      "step": 5760
+    },
+    {
+      "epoch": 1.03,
+      "grad_norm": 1.1802875995635986,
+      "learning_rate": 4.151602852638888e-06,
+      "loss": 0.222,
+      "step": 5792
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 1.2036052942276,
+      "learning_rate": 4.139141035651288e-06,
+      "loss": 0.2093,
+      "step": 5824
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 1.1690653562545776,
+      "learning_rate": 4.126607366136395e-06,
+      "loss": 0.1925,
+      "step": 5856
+    },
+    {
+      "epoch": 1.05,
+      "grad_norm": 0.9996016621589661,
+      "learning_rate": 4.114002393513577e-06,
+      "loss": 0.206,
+      "step": 5888
+    },
+    {
+      "epoch": 1.05,
+      "grad_norm": 1.1670773029327393,
+      "learning_rate": 4.101326670327807e-06,
+      "loss": 0.2097,
+      "step": 5920
+    },
+    {
+      "epoch": 1.06,
+      "grad_norm": 0.8733654022216797,
+      "learning_rate": 4.0885807522254435e-06,
+      "loss": 0.2015,
+      "step": 5952
+    },
+    {
+      "epoch": 1.06,
+      "grad_norm": 1.2280749082565308,
+      "learning_rate": 4.075765197929872e-06,
+      "loss": 0.2108,
+      "step": 5984
+    },
+    {
+      "epoch": 1.07,
+      "grad_norm": 1.1926356554031372,
+      "learning_rate": 4.0628805692170105e-06,
+      "loss": 0.2047,
+      "step": 6016
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 1.0048396587371826,
+      "learning_rate": 4.049927430890693e-06,
+      "loss": 0.2077,
+      "step": 6048
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 1.026442050933838,
+      "learning_rate": 4.0369063507578995e-06,
+      "loss": 0.2051,
+      "step": 6080
+    },
+    {
+      "epoch": 1.09,
+      "grad_norm": 1.1310842037200928,
+      "learning_rate": 4.023817899603875e-06,
+      "loss": 0.2055,
+      "step": 6112
+    },
+    {
+      "epoch": 1.09,
+      "grad_norm": 1.1275712251663208,
+      "learning_rate": 4.010662651167106e-06,
+      "loss": 0.1965,
+      "step": 6144
+    },
+    {
+      "epoch": 1.1,
+      "grad_norm": 1.1789113283157349,
+      "learning_rate": 3.997441182114164e-06,
+      "loss": 0.2118,
+      "step": 6176
+    },
+    {
+      "epoch": 1.1,
+      "grad_norm": 1.3836599588394165,
+      "learning_rate": 3.984154072014438e-06,
+      "loss": 0.2056,
+      "step": 6208
+    },
+    {
+      "epoch": 1.11,
+      "grad_norm": 1.1013050079345703,
+      "learning_rate": 3.970801903314722e-06,
+      "loss": 0.2109,
+      "step": 6240
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 1.1018249988555908,
+      "learning_rate": 3.957385261313685e-06,
+      "loss": 0.202,
+      "step": 6272
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 1.3906185626983643,
+      "learning_rate": 3.943904734136213e-06,
+      "loss": 0.2065,
+      "step": 6304
+    },
+    {
+      "epoch": 1.13,
+      "grad_norm": 1.2197610139846802,
+      "learning_rate": 3.930360912707632e-06,
+      "loss": 0.2096,
+      "step": 6336
+    },
+    {
+      "epoch": 1.13,
+      "grad_norm": 1.0342845916748047,
+      "learning_rate": 3.916754390727795e-06,
+      "loss": 0.2024,
+      "step": 6368
+    },
+    {
+      "epoch": 1.14,
+      "grad_norm": 1.236260175704956,
+      "learning_rate": 3.90308576464507e-06,
+      "loss": 0.216,
+      "step": 6400
+    },
+    {
+      "epoch": 1.14,
+      "grad_norm": 1.3906086683273315,
+      "learning_rate": 3.889355633630186e-06,
+      "loss": 0.2153,
+      "step": 6432
+    },
+    {
+      "epoch": 1.15,
+      "grad_norm": 1.2441129684448242,
+      "learning_rate": 3.875564599549968e-06,
+      "loss": 0.2092,
+      "step": 6464
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 1.2320338487625122,
+      "learning_rate": 3.861713266940959e-06,
+      "loss": 0.2038,
+      "step": 6496
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 1.6422646045684814,
+      "learning_rate": 3.847802242982915e-06,
+      "loss": 0.205,
+      "step": 6528
+    },
+    {
+      "epoch": 1.17,
+      "grad_norm": 1.1179068088531494,
+      "learning_rate": 3.83383213747219e-06,
+      "loss": 0.2162,
+      "step": 6560
+    },
+    {
+      "epoch": 1.17,
+      "grad_norm": 1.0986745357513428,
+      "learning_rate": 3.8198035627950084e-06,
+      "loss": 0.1956,
+      "step": 6592
+    },
+    {
+      "epoch": 1.18,
+      "grad_norm": 1.340859055519104,
+      "learning_rate": 3.8057171339006138e-06,
+      "loss": 0.2093,
+      "step": 6624
+    },
+    {
+      "epoch": 1.18,
+      "grad_norm": 1.7803446054458618,
+      "learning_rate": 3.791573468274323e-06,
+      "loss": 0.2133,
+      "step": 6656
+    },
+    {
+      "epoch": 1.19,
+      "grad_norm": 1.022388219833374,
+      "learning_rate": 3.777373185910448e-06,
+      "loss": 0.2182,
+      "step": 6688
+    },
+    {
+      "epoch": 1.19,
+      "grad_norm": 1.0795223712921143,
+      "learning_rate": 3.7631169092851226e-06,
+      "loss": 0.2051,
+      "step": 6720
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 1.0785856246948242,
+      "learning_rate": 3.7488052633290174e-06,
+      "loss": 0.2047,
+      "step": 6752
+    },
+    {
+      "epoch": 1.21,
+      "grad_norm": 1.0391508340835571,
+      "learning_rate": 3.7344388753999434e-06,
+      "loss": 0.2081,
+      "step": 6784
+    },
+    {
+      "epoch": 1.21,
+      "grad_norm": 1.3015925884246826,
+      "learning_rate": 3.720018375255352e-06,
+      "loss": 0.2013,
+      "step": 6816
+    },
+    {
+      "epoch": 1.22,
+      "grad_norm": 1.2382066249847412,
+      "learning_rate": 3.7055443950247276e-06,
+      "loss": 0.2037,
+      "step": 6848
+    },
+    {
+      "epoch": 1.22,
+      "grad_norm": 1.1386123895645142,
+      "learning_rate": 3.691017569181882e-06,
+      "loss": 0.2046,
+      "step": 6880
+    },
+    {
+      "epoch": 1.23,
+      "grad_norm": 0.9857081770896912,
+      "learning_rate": 3.6764385345171393e-06,
+      "loss": 0.207,
+      "step": 6912
+    },
+    {
+      "epoch": 1.23,
+      "grad_norm": 1.1276394128799438,
+      "learning_rate": 3.661807930109422e-06,
+      "loss": 0.2134,
+      "step": 6944
+    },
+    {
+      "epoch": 1.24,
+      "grad_norm": 1.1821982860565186,
+      "learning_rate": 3.647126397298234e-06,
+      "loss": 0.2162,
+      "step": 6976
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 1.218800663948059,
+      "learning_rate": 3.632394579655555e-06,
+      "loss": 0.2023,
+      "step": 7008
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 1.083310842514038,
+      "learning_rate": 3.6176131229576193e-06,
+      "loss": 0.1999,
+      "step": 7040
+    },
+    {
+      "epoch": 1.26,
+      "grad_norm": 1.0640002489089966,
+      "learning_rate": 3.602782675156617e-06,
+      "loss": 0.2125,
+      "step": 7072
+    },
+    {
+      "epoch": 1.26,
+      "grad_norm": 1.1672149896621704,
+      "learning_rate": 3.5879038863522843e-06,
+      "loss": 0.2157,
+      "step": 7104
+    },
+    {
+      "epoch": 1.27,
+      "grad_norm": 1.1732845306396484,
+      "learning_rate": 3.572977408763407e-06,
+      "loss": 0.2082,
+      "step": 7136
+    },
+    {
+      "epoch": 1.27,
+      "grad_norm": 1.1544941663742065,
+      "learning_rate": 3.5580038966992344e-06,
+      "loss": 0.2067,
+      "step": 7168
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 1.2914546728134155,
+      "learning_rate": 3.5429840065307924e-06,
+      "loss": 0.2019,
+      "step": 7200
+    },
+    {
+      "epoch": 1.29,
+      "grad_norm": 1.0473650693893433,
+      "learning_rate": 3.527918396662115e-06,
+      "loss": 0.1952,
+      "step": 7232
+    },
+    {
+      "epoch": 1.29,
+      "grad_norm": 1.2211614847183228,
+      "learning_rate": 3.512807727501379e-06,
+      "loss": 0.2093,
+      "step": 7264
+    },
+    {
+      "epoch": 1.3,
+      "grad_norm": 1.1035760641098022,
+      "learning_rate": 3.4976526614319573e-06,
+      "loss": 0.2007,
+      "step": 7296
+    },
+    {
+      "epoch": 1.3,
+      "grad_norm": 1.2120308876037598,
+      "learning_rate": 3.4824538627833825e-06,
+      "loss": 0.2205,
+      "step": 7328
+    },
+    {
+      "epoch": 1.31,
+      "grad_norm": 0.8647122979164124,
+      "learning_rate": 3.4672119978022277e-06,
+      "loss": 0.2063,
+      "step": 7360
+    },
+    {
+      "epoch": 1.31,
+      "grad_norm": 1.1142189502716064,
+      "learning_rate": 3.4519277346228953e-06,
+      "loss": 0.2075,
+      "step": 7392
+    },
+    {
+      "epoch": 1.32,
+      "grad_norm": 1.3183207511901855,
+      "learning_rate": 3.436601743238335e-06,
+      "loss": 0.2094,
+      "step": 7424
+    },
+    {
+      "epoch": 1.33,
+      "grad_norm": 1.0320820808410645,
+      "learning_rate": 3.421234695470673e-06,
+      "loss": 0.2029,
+      "step": 7456
+    },
+    {
+      "epoch": 1.33,
+      "grad_norm": 1.3065481185913086,
+      "learning_rate": 3.4058272649417607e-06,
+      "loss": 0.2127,
+      "step": 7488
+    },
+    {
+      "epoch": 1.34,
+      "grad_norm": 1.209372639656067,
+      "learning_rate": 3.3903801270436465e-06,
+      "loss": 0.2015,
+      "step": 7520
+    },
+    {
+      "epoch": 1.34,
+      "grad_norm": 1.0177247524261475,
+      "learning_rate": 3.374893958908971e-06,
+      "loss": 0.2075,
+      "step": 7552
+    },
+    {
+      "epoch": 1.35,
+      "grad_norm": 1.4457709789276123,
+      "learning_rate": 3.3593694393812827e-06,
+      "loss": 0.2098,
+      "step": 7584
+    },
+    {
+      "epoch": 1.35,
+      "grad_norm": 1.1711078882217407,
+      "learning_rate": 3.3438072489852837e-06,
+      "loss": 0.2088,
+      "step": 7616
+    },
+    {
+      "epoch": 1.36,
+      "grad_norm": 1.1928409337997437,
+      "learning_rate": 3.3282080698969953e-06,
+      "loss": 0.1918,
+      "step": 7648
+    },
+    {
+      "epoch": 1.37,
+      "grad_norm": 0.9215808510780334,
+      "learning_rate": 3.3125725859138548e-06,
+      "loss": 0.2106,
+      "step": 7680
+    },
+    {
+      "epoch": 1.37,
+      "grad_norm": 1.3021633625030518,
+      "learning_rate": 3.2969014824247436e-06,
+      "loss": 0.2018,
+      "step": 7712
+    },
+    {
+      "epoch": 1.38,
+      "grad_norm": 1.1597398519515991,
+      "learning_rate": 3.28119544637994e-06,
+      "loss": 0.2035,
+      "step": 7744
+    },
+    {
+      "epoch": 1.38,
+      "grad_norm": 1.2015706300735474,
+      "learning_rate": 3.265455166261009e-06,
+      "loss": 0.2027,
+      "step": 7776
+    },
+    {
+      "epoch": 1.39,
+      "grad_norm": 1.1449264287948608,
+      "learning_rate": 3.2496813320506183e-06,
+      "loss": 0.2165,
+      "step": 7808
+    },
+    {
+      "epoch": 1.39,
+      "grad_norm": 1.1332265138626099,
+      "learning_rate": 3.2338746352022965e-06,
+      "loss": 0.2006,
+      "step": 7840
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 1.430891990661621,
+      "learning_rate": 3.2180357686101226e-06,
+      "loss": 0.2102,
+      "step": 7872
+    },
+    {
+      "epoch": 1.41,
+      "grad_norm": 1.4063985347747803,
+      "learning_rate": 3.2021654265783505e-06,
+      "loss": 0.196,
+      "step": 7904
+    },
+    {
+      "epoch": 1.41,
+      "grad_norm": 1.3970558643341064,
+      "learning_rate": 3.1862643047909746e-06,
+      "loss": 0.2161,
+      "step": 7936
+    },
+    {
+      "epoch": 1.42,
+      "grad_norm": 1.3233983516693115,
+      "learning_rate": 3.170333100281236e-06,
+      "loss": 0.1921,
+      "step": 7968
+    },
+    {
+      "epoch": 1.42,
+      "grad_norm": 1.2325806617736816,
+      "learning_rate": 3.154372511401064e-06,
+      "loss": 0.2042,
+      "step": 8000
+    },
+    {
+      "epoch": 1.43,
+      "grad_norm": 1.0558186769485474,
+      "learning_rate": 3.1383832377904676e-06,
+      "loss": 0.2056,
+      "step": 8032
+    },
+    {
+      "epoch": 1.43,
+      "grad_norm": 1.189503788948059,
+      "learning_rate": 3.1223659803468653e-06,
+      "loss": 0.203,
+      "step": 8064
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 1.1646627187728882,
+      "learning_rate": 3.1063214411943576e-06,
+      "loss": 0.2088,
+      "step": 8096
+    },
+    {
+      "epoch": 1.45,
+      "grad_norm": 1.1149977445602417,
+      "learning_rate": 3.0902503236529533e-06,
+      "loss": 0.2081,
+      "step": 8128
+    },
+    {
+      "epoch": 1.45,
+      "grad_norm": 1.5566740036010742,
+      "learning_rate": 3.074153332207738e-06,
+      "loss": 0.2141,
+      "step": 8160
+    },
+    {
+      "epoch": 1.46,
+      "grad_norm": 1.304262638092041,
+      "learning_rate": 3.058031172477992e-06,
+      "loss": 0.2006,
+      "step": 8192
+    },
+    {
+      "epoch": 1.46,
+      "grad_norm": 1.1247010231018066,
+      "learning_rate": 3.041884551186258e-06,
+      "loss": 0.2109,
+      "step": 8224
+    },
+    {
+      "epoch": 1.47,
+      "grad_norm": 1.4587311744689941,
+      "learning_rate": 3.0257141761273627e-06,
+      "loss": 0.2016,
+      "step": 8256
+    },
+    {
+      "epoch": 1.47,
+      "grad_norm": 1.1545838117599487,
+      "learning_rate": 3.0095207561373935e-06,
+      "loss": 0.183,
+      "step": 8288
+    },
+    {
+      "epoch": 1.48,
+      "grad_norm": 1.3221790790557861,
+      "learning_rate": 2.9933050010626208e-06,
+      "loss": 0.1985,
+      "step": 8320
+    },
+    {
+      "epoch": 1.49,
+      "grad_norm": 1.1717700958251953,
+      "learning_rate": 2.9770676217283844e-06,
+      "loss": 0.2113,
+      "step": 8352
+    },
+    {
+      "epoch": 1.49,
+      "grad_norm": 1.0709022283554077,
+      "learning_rate": 2.960809329907934e-06,
+      "loss": 0.2012,
+      "step": 8384
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 1.4594495296478271,
+      "learning_rate": 2.944530838291229e-06,
+      "loss": 0.2039,
+      "step": 8416
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 1.3135212659835815,
+      "learning_rate": 2.928232860453694e-06,
+      "loss": 0.206,
+      "step": 8448
+    },
+    {
+      "epoch": 1.51,
+      "grad_norm": 1.316394329071045,
+      "learning_rate": 2.911916110824945e-06,
+      "loss": 0.212,
+      "step": 8480
+    },
+    {
+      "epoch": 1.51,
+      "grad_norm": 1.3602452278137207,
+      "learning_rate": 2.895581304657465e-06,
+      "loss": 0.2068,
+      "step": 8512
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 1.2073897123336792,
+      "learning_rate": 2.8792291579952553e-06,
+      "loss": 0.2098,
+      "step": 8544
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 1.2983072996139526,
+      "learning_rate": 2.8628603876424467e-06,
+      "loss": 0.2086,
+      "step": 8576
+    },
+    {
+      "epoch": 1.53,
+      "grad_norm": 1.0781196355819702,
+      "learning_rate": 2.846475711131877e-06,
+      "loss": 0.201,
+      "step": 8608
+    },
+    {
+      "epoch": 1.54,
+      "grad_norm": 1.1917251348495483,
+      "learning_rate": 2.8300758466936366e-06,
+      "loss": 0.1982,
+      "step": 8640
+    },
+    {
+      "epoch": 1.54,
+      "grad_norm": 1.2894983291625977,
+      "learning_rate": 2.813661513223588e-06,
+      "loss": 0.1943,
+      "step": 8672
+    },
+    {
+      "epoch": 1.55,
+      "grad_norm": 1.2249202728271484,
+      "learning_rate": 2.7972334302518504e-06,
+      "loss": 0.2145,
+      "step": 8704
+    },
+    {
+      "epoch": 1.55,
+      "grad_norm": 1.1947064399719238,
+      "learning_rate": 2.7807923179112576e-06,
+      "loss": 0.2003,
+      "step": 8736
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 1.0660251379013062,
+      "learning_rate": 2.764338896905792e-06,
+      "loss": 0.1984,
+      "step": 8768
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 1.0243247747421265,
+      "learning_rate": 2.7478738884789934e-06,
+      "loss": 0.2036,
+      "step": 8800
+    },
+    {
+      "epoch": 1.57,
+      "grad_norm": 1.286199927330017,
+      "learning_rate": 2.731398014382341e-06,
+      "loss": 0.2027,
+      "step": 8832
+    },
+    {
+      "epoch": 1.58,
+      "grad_norm": 1.1617448329925537,
+      "learning_rate": 2.714911996843617e-06,
+      "loss": 0.2162,
+      "step": 8864
+    },
+    {
+      "epoch": 1.58,
+      "grad_norm": 1.1921496391296387,
+      "learning_rate": 2.6984165585352435e-06,
+      "loss": 0.2124,
+      "step": 8896
+    },
+    {
+      "epoch": 1.59,
+      "grad_norm": 1.2066140174865723,
+      "learning_rate": 2.6819124225426085e-06,
+      "loss": 0.199,
+      "step": 8928
+    },
+    {
+      "epoch": 1.59,
+      "grad_norm": 1.0459320545196533,
+      "learning_rate": 2.665400312332368e-06,
+      "loss": 0.2072,
+      "step": 8960
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 1.2983636856079102,
+      "learning_rate": 2.648880951720729e-06,
+      "loss": 0.2024,
+      "step": 8992
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 1.0876768827438354,
+      "learning_rate": 2.6323550648417267e-06,
+      "loss": 0.2143,
+      "step": 9024
+    },
+    {
+      "epoch": 1.61,
+      "grad_norm": 1.0047022104263306,
+      "learning_rate": 2.6158233761154744e-06,
+      "loss": 0.2043,
+      "step": 9056
+    },
+    {
+      "epoch": 1.62,
+      "grad_norm": 0.9878237247467041,
+      "learning_rate": 2.5992866102164146e-06,
+      "loss": 0.1991,
+      "step": 9088
+    },
+    {
+      "epoch": 1.62,
+      "grad_norm": 0.9894827604293823,
+      "learning_rate": 2.58274549204155e-06,
+      "loss": 0.1979,
+      "step": 9120
+    },
+    {
+      "epoch": 1.63,
+      "grad_norm": 0.9374232292175293,
+      "learning_rate": 2.5662007466786674e-06,
+      "loss": 0.2055,
+      "step": 9152
+    },
+    {
+      "epoch": 1.63,
+      "grad_norm": 1.259948492050171,
+      "learning_rate": 2.5496530993745518e-06,
+      "loss": 0.2057,
+      "step": 9184
+    },
+    {
+      "epoch": 1.64,
+      "grad_norm": 0.958737850189209,
+      "learning_rate": 2.533103275503197e-06,
+      "loss": 0.2029,
+      "step": 9216
+    },
+    {
+      "epoch": 1.64,
+      "grad_norm": 1.079717755317688,
+      "learning_rate": 2.5165520005340082e-06,
+      "loss": 0.2049,
+      "step": 9248
+    },
+    {
+      "epoch": 1.65,
+      "grad_norm": 1.1001982688903809,
+      "learning_rate": 2.5e-06,
+      "loss": 0.211,
+      "step": 9280
+    },
+    {
+      "epoch": 1.66,
+      "grad_norm": 1.2408779859542847,
+      "learning_rate": 2.4834479994659926e-06,
+      "loss": 0.2028,
+      "step": 9312
+    },
+    {
+      "epoch": 1.66,
+      "grad_norm": 1.0395313501358032,
+      "learning_rate": 2.4668967244968035e-06,
+      "loss": 0.1988,
+      "step": 9344
+    },
+    {
+      "epoch": 1.67,
+      "grad_norm": 1.0080056190490723,
+      "learning_rate": 2.4503469006254487e-06,
+      "loss": 0.1988,
+      "step": 9376
+    },
+    {
+      "epoch": 1.67,
+      "grad_norm": 1.4669593572616577,
+      "learning_rate": 2.4337992533213334e-06,
+      "loss": 0.1942,
+      "step": 9408
+    },
+    {
+      "epoch": 1.68,
+      "grad_norm": 1.3393524885177612,
+      "learning_rate": 2.4172545079584508e-06,
+      "loss": 0.1964,
+      "step": 9440
+    },
+    {
+      "epoch": 1.68,
+      "grad_norm": 0.9786149263381958,
+      "learning_rate": 2.4007133897835863e-06,
+      "loss": 0.1984,
+      "step": 9472
+    },
+    {
+      "epoch": 1.69,
+      "grad_norm": 1.1999776363372803,
+      "learning_rate": 2.3841766238845264e-06,
+      "loss": 0.2102,
+      "step": 9504
+    },
+    {
+      "epoch": 1.7,
+      "grad_norm": 1.3276174068450928,
+      "learning_rate": 2.367644935158274e-06,
+      "loss": 0.1941,
+      "step": 9536
+    },
+    {
+      "epoch": 1.7,
+      "grad_norm": 1.0124472379684448,
+      "learning_rate": 2.3511190482792713e-06,
+      "loss": 0.199,
+      "step": 9568
+    },
+    {
+      "epoch": 1.71,
+      "grad_norm": 1.258489966392517,
+      "learning_rate": 2.3345996876676334e-06,
+      "loss": 0.2008,
+      "step": 9600
+    },
+    {
+      "epoch": 1.71,
+      "grad_norm": 1.1993016004562378,
+      "learning_rate": 2.318087577457392e-06,
+      "loss": 0.2154,
+      "step": 9632
+    },
+    {
+      "epoch": 1.72,
+      "grad_norm": 1.250908613204956,
+      "learning_rate": 2.3015834414647573e-06,
+      "loss": 0.2068,
+      "step": 9664
+    },
+    {
+      "epoch": 1.72,
+      "grad_norm": 1.211915373802185,
+      "learning_rate": 2.2850880031563845e-06,
+      "loss": 0.1946,
+      "step": 9696
+    },
+    {
+      "epoch": 1.73,
+      "grad_norm": 1.0278340578079224,
+      "learning_rate": 2.26860198561766e-06,
+      "loss": 0.1948,
+      "step": 9728
+    },
+    {
+      "epoch": 1.74,
+      "grad_norm": 1.2455780506134033,
+      "learning_rate": 2.2521261115210074e-06,
+      "loss": 0.197,
+      "step": 9760
+    },
+    {
+      "epoch": 1.74,
+      "grad_norm": 1.2321908473968506,
+      "learning_rate": 2.2356611030942084e-06,
+      "loss": 0.2075,
+      "step": 9792
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 1.0618436336517334,
+      "learning_rate": 2.219207682088743e-06,
+      "loss": 0.1931,
+      "step": 9824
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 1.36842942237854,
+      "learning_rate": 2.20276656974815e-06,
+      "loss": 0.1999,
+      "step": 9856
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 1.033603310585022,
+      "learning_rate": 2.186338486776412e-06,
+      "loss": 0.2028,
+      "step": 9888
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 1.303781270980835,
+      "learning_rate": 2.169924153306363e-06,
+      "loss": 0.214,
+      "step": 9920
+    },
+    {
+      "epoch": 1.77,
+      "grad_norm": 1.2051355838775635,
+      "learning_rate": 2.153524288868124e-06,
+      "loss": 0.2091,
+      "step": 9952
+    },
+    {
+      "epoch": 1.78,
+      "grad_norm": 0.9946267604827881,
+      "learning_rate": 2.137139612357554e-06,
+      "loss": 0.1942,
+      "step": 9984
+    },
+    {
+      "epoch": 1.78,
+      "grad_norm": 1.283492088317871,
+      "learning_rate": 2.120770842004746e-06,
+      "loss": 0.1971,
+      "step": 10016
+    },
+    {
+      "epoch": 1.79,
+      "grad_norm": 0.9329233169555664,
+      "learning_rate": 2.1044186953425358e-06,
+      "loss": 0.203,
+      "step": 10048
+    },
+    {
+      "epoch": 1.79,
+      "grad_norm": 1.082767367362976,
+      "learning_rate": 2.0880838891750553e-06,
+      "loss": 0.2012,
+      "step": 10080
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 1.1007740497589111,
+      "learning_rate": 2.0717671395463063e-06,
+      "loss": 0.2028,
+      "step": 10112
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 1.2502014636993408,
+      "learning_rate": 2.0554691617087725e-06,
+      "loss": 0.2121,
+      "step": 10144
+    },
+    {
+      "epoch": 1.81,
+      "grad_norm": 1.073034405708313,
+      "learning_rate": 2.0391906700920667e-06,
+      "loss": 0.1994,
+      "step": 10176
+    },
+    {
+      "epoch": 1.82,
+      "grad_norm": 0.9409189820289612,
+      "learning_rate": 2.0229323782716156e-06,
+      "loss": 0.2054,
+      "step": 10208
+    },
+    {
+      "epoch": 1.82,
+      "grad_norm": 1.197383999824524,
+      "learning_rate": 2.0066949989373797e-06,
+      "loss": 0.1946,
+      "step": 10240
+    },
+    {
+      "epoch": 1.83,
+      "grad_norm": 1.261612892150879,
+      "learning_rate": 1.9904792438626074e-06,
+      "loss": 0.2038,
+      "step": 10272
+    },
+    {
+      "epoch": 1.83,
+      "grad_norm": 1.4839472770690918,
+      "learning_rate": 1.9742858238726377e-06,
+      "loss": 0.2067,
+      "step": 10304
+    },
+    {
+      "epoch": 1.84,
+      "grad_norm": 1.0103521347045898,
+      "learning_rate": 1.9581154488137425e-06,
+      "loss": 0.2104,
+      "step": 10336
+    },
+    {
+      "epoch": 1.84,
+      "grad_norm": 1.2776283025741577,
+      "learning_rate": 1.9419688275220085e-06,
+      "loss": 0.196,
+      "step": 10368
+    },
+    {
+      "epoch": 1.85,
+      "grad_norm": 1.0784910917282104,
+      "learning_rate": 1.9258466677922624e-06,
+      "loss": 0.1975,
+      "step": 10400
+    },
+    {
+      "epoch": 1.85,
+      "grad_norm": 0.9643808007240295,
+      "learning_rate": 1.909749676347047e-06,
+      "loss": 0.2111,
+      "step": 10432
+    },
+    {
+      "epoch": 1.86,
+      "grad_norm": 1.3432437181472778,
+      "learning_rate": 1.8936785588056428e-06,
+      "loss": 0.1923,
+      "step": 10464
+    },
+    {
+      "epoch": 1.87,
+      "grad_norm": 1.133470892906189,
+      "learning_rate": 1.8776340196531351e-06,
+      "loss": 0.2016,
+      "step": 10496
+    },
+    {
+      "epoch": 1.87,
+      "grad_norm": 1.0897003412246704,
+      "learning_rate": 1.8616167622095328e-06,
+      "loss": 0.193,
+      "step": 10528
+    },
+    {
+      "epoch": 1.88,
+      "grad_norm": 0.9629374146461487,
+      "learning_rate": 1.8456274885989374e-06,
+      "loss": 0.1937,
+      "step": 10560
+    },
+    {
+      "epoch": 1.88,
+      "grad_norm": 1.406630039215088,
+      "learning_rate": 1.829666899718765e-06,
+      "loss": 0.1997,
+      "step": 10592
+    },
+    {
+      "epoch": 1.89,
+      "grad_norm": 1.165366291999817,
+      "learning_rate": 1.8137356952090258e-06,
+      "loss": 0.1976,
+      "step": 10624
+    },
+    {
+      "epoch": 1.89,
+      "grad_norm": 1.2674609422683716,
+      "learning_rate": 1.7978345734216502e-06,
+      "loss": 0.1908,
+      "step": 10656
+    },
+    {
+      "epoch": 1.9,
+      "grad_norm": 1.2211626768112183,
+      "learning_rate": 1.7819642313898783e-06,
+      "loss": 0.1984,
+      "step": 10688
+    },
+    {
+      "epoch": 1.91,
+      "grad_norm": 1.1066653728485107,
+      "learning_rate": 1.766125364797704e-06,
+      "loss": 0.2035,
+      "step": 10720
+    },
+    {
+      "epoch": 1.91,
+      "grad_norm": 1.539157748222351,
+      "learning_rate": 1.7503186679493821e-06,
+      "loss": 0.201,
+      "step": 10752
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 1.1897879838943481,
+      "learning_rate": 1.7345448337389918e-06,
+      "loss": 0.194,
+      "step": 10784
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 1.162571668624878,
+      "learning_rate": 1.7188045536200604e-06,
+      "loss": 0.1899,
+      "step": 10816
+    },
+    {
+      "epoch": 1.93,
+      "grad_norm": 1.0616639852523804,
+      "learning_rate": 1.7030985175752574e-06,
+      "loss": 0.1978,
+      "step": 10848
+    },
+    {
+      "epoch": 1.93,
+      "grad_norm": 1.0391641855239868,
+      "learning_rate": 1.687427414086146e-06,
+      "loss": 0.2017,
+      "step": 10880
+    },
+    {
+      "epoch": 1.94,
+      "grad_norm": 1.0870113372802734,
+      "learning_rate": 1.6717919301030055e-06,
+      "loss": 0.2012,
+      "step": 10912
+    },
+    {
+      "epoch": 1.95,
+      "grad_norm": 1.143446922302246,
+      "learning_rate": 1.6561927510147172e-06,
+      "loss": 0.1911,
+      "step": 10944
+    },
+    {
+      "epoch": 1.95,
+      "grad_norm": 1.015080213546753,
+      "learning_rate": 1.6406305606187183e-06,
+      "loss": 0.198,
+      "step": 10976
+    },
+    {
+      "epoch": 1.96,
+      "grad_norm": 0.9722278714179993,
+      "learning_rate": 1.6251060410910301e-06,
+      "loss": 0.1862,
+      "step": 11008
+    },
+    {
+      "epoch": 1.96,
+      "grad_norm": 0.9311158061027527,
+      "learning_rate": 1.6096198729563539e-06,
+      "loss": 0.198,
+      "step": 11040
+    },
+    {
+      "epoch": 1.97,
+      "grad_norm": 1.3127020597457886,
+      "learning_rate": 1.5941727350582399e-06,
+      "loss": 0.2,
+      "step": 11072
+    },
+    {
+      "epoch": 1.97,
+      "grad_norm": 0.9450055956840515,
+      "learning_rate": 1.5787653045293278e-06,
+      "loss": 0.2015,
+      "step": 11104
+    },
+    {
+      "epoch": 1.98,
+      "grad_norm": 1.1553057432174683,
+      "learning_rate": 1.5633982567616657e-06,
+      "loss": 0.2068,
+      "step": 11136
+    },
+    {
+      "epoch": 1.99,
+      "grad_norm": 0.8617095351219177,
+      "learning_rate": 1.548072265377105e-06,
+      "loss": 0.2014,
+      "step": 11168
+    },
+    {
+      "epoch": 1.99,
+      "grad_norm": 0.9857013821601868,
+      "learning_rate": 1.532788002197773e-06,
+      "loss": 0.2031,
+      "step": 11200
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.0158861875534058,
+      "learning_rate": 1.5175461372166177e-06,
+      "loss": 0.1941,
+      "step": 11232
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.1121022701263428,
+      "learning_rate": 1.5023473385680438e-06,
+      "loss": 0.1708,
+      "step": 11264
+    },
+    {
+      "epoch": 2.01,
+      "grad_norm": 1.1300593614578247,
+      "learning_rate": 1.4871922724986215e-06,
+      "loss": 0.1504,
+      "step": 11296
+    },
+    {
+      "epoch": 2.01,
+      "grad_norm": 1.232246994972229,
+      "learning_rate": 1.4720816033378856e-06,
+      "loss": 0.151,
+      "step": 11328
+    },
+    {
+      "epoch": 2.02,
+      "grad_norm": 1.2618398666381836,
+      "learning_rate": 1.4570159934692085e-06,
+      "loss": 0.1421,
+      "step": 11360
+    },
+    {
+      "epoch": 2.03,
+      "grad_norm": 1.275038242340088,
+      "learning_rate": 1.4419961033007669e-06,
+      "loss": 0.1457,
+      "step": 11392
+    },
+    {
+      "epoch": 2.03,
+      "grad_norm": 1.079405426979065,
+      "learning_rate": 1.427022591236594e-06,
+      "loss": 0.144,
+      "step": 11424
+    },
+    {
+      "epoch": 2.04,
+      "grad_norm": 1.1695400476455688,
+      "learning_rate": 1.4120961136477168e-06,
+      "loss": 0.1531,
+      "step": 11456
+    },
+    {
+      "epoch": 2.04,
+      "grad_norm": 1.008547306060791,
+      "learning_rate": 1.3972173248433832e-06,
+      "loss": 0.1453,
+      "step": 11488
+    },
+    {
+      "epoch": 2.05,
+      "grad_norm": 1.0746265649795532,
+      "learning_rate": 1.3823868770423815e-06,
+      "loss": 0.1446,
+      "step": 11520
+    },
+    {
+      "epoch": 2.05,
+      "grad_norm": 1.1596614122390747,
+      "learning_rate": 1.3676054203444462e-06,
+      "loss": 0.1477,
+      "step": 11552
+    },
+    {
+      "epoch": 2.06,
+      "grad_norm": 1.1029706001281738,
+      "learning_rate": 1.3528736027017663e-06,
+      "loss": 0.1477,
+      "step": 11584
+    },
+    {
+      "epoch": 2.07,
+      "grad_norm": 0.9398396015167236,
+      "learning_rate": 1.3381920698905788e-06,
+      "loss": 0.1477,
+      "step": 11616
+    },
+    {
+      "epoch": 2.07,
+      "grad_norm": 1.0209776163101196,
+      "learning_rate": 1.3235614654828604e-06,
+      "loss": 0.1448,
+      "step": 11648
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 0.9415841102600098,
+      "learning_rate": 1.3089824308181187e-06,
+      "loss": 0.1481,
+      "step": 11680
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 1.088069200515747,
+      "learning_rate": 1.2944556049752726e-06,
+      "loss": 0.149,
+      "step": 11712
+    },
+    {
+      "epoch": 2.09,
+      "grad_norm": 1.3269786834716797,
+      "learning_rate": 1.2799816247446494e-06,
+      "loss": 0.1497,
+      "step": 11744
+    },
+    {
+      "epoch": 2.09,
+      "grad_norm": 0.9119545817375183,
+      "learning_rate": 1.265561124600057e-06,
+      "loss": 0.1467,
+      "step": 11776
+    },
+    {
+      "epoch": 2.1,
+      "grad_norm": 1.0677683353424072,
+      "learning_rate": 1.251194736670983e-06,
+      "loss": 0.1448,
+      "step": 11808
+    },
+    {
+      "epoch": 2.11,
+      "grad_norm": 1.0756884813308716,
+      "learning_rate": 1.2368830907148778e-06,
+      "loss": 0.1363,
+      "step": 11840
+    },
+    {
+      "epoch": 2.11,
+      "grad_norm": 1.0578961372375488,
+      "learning_rate": 1.2226268140895528e-06,
+      "loss": 0.1527,
+      "step": 11872
+    },
+    {
+      "epoch": 2.12,
+      "grad_norm": 1.0562700033187866,
+      "learning_rate": 1.2084265317256772e-06,
+      "loss": 0.1449,
+      "step": 11904
+    },
+    {
+      "epoch": 2.12,
+      "grad_norm": 1.0958082675933838,
+      "learning_rate": 1.1942828660993869e-06,
+      "loss": 0.1474,
+      "step": 11936
+    },
+    {
+      "epoch": 2.13,
+      "grad_norm": 0.9672511219978333,
+      "learning_rate": 1.1801964372049932e-06,
+      "loss": 0.1459,
+      "step": 11968
+    },
+    {
+      "epoch": 2.13,
+      "grad_norm": 1.1125974655151367,
+      "learning_rate": 1.1661678625278106e-06,
+      "loss": 0.1483,
+      "step": 12000
+    },
+    {
+      "epoch": 2.14,
+      "grad_norm": 1.227283239364624,
+      "learning_rate": 1.152197757017086e-06,
+      "loss": 0.1453,
+      "step": 12032
+    },
+    {
+      "epoch": 2.15,
+      "grad_norm": 1.3206896781921387,
+      "learning_rate": 1.1382867330590414e-06,
+      "loss": 0.1425,
+      "step": 12064
+    },
+    {
+      "epoch": 2.15,
+      "grad_norm": 0.9912922978401184,
+      "learning_rate": 1.1244354004500335e-06,
+      "loss": 0.1529,
+      "step": 12096
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 0.8996345400810242,
+      "learning_rate": 1.110644366369815e-06,
+      "loss": 0.1437,
+      "step": 12128
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 0.9117119312286377,
+      "learning_rate": 1.0969142353549315e-06,
+      "loss": 0.1429,
+      "step": 12160
+    },
+    {
+      "epoch": 2.17,
+      "grad_norm": 1.1916334629058838,
+      "learning_rate": 1.0832456092722063e-06,
+      "loss": 0.1509,
+      "step": 12192
+    },
+    {
+      "epoch": 2.17,
+      "grad_norm": 1.1345574855804443,
+      "learning_rate": 1.0696390872923696e-06,
+      "loss": 0.1547,
+      "step": 12224
+    },
+    {
+      "epoch": 2.18,
+      "grad_norm": 1.3311399221420288,
+      "learning_rate": 1.0560952658637869e-06,
+      "loss": 0.1428,
+      "step": 12256
+    },
+    {
+      "epoch": 2.18,
+      "grad_norm": 1.0195939540863037,
+      "learning_rate": 1.042614738686315e-06,
+      "loss": 0.1447,
+      "step": 12288
+    },
+    {
+      "epoch": 2.19,
+      "grad_norm": 1.1453065872192383,
+      "learning_rate": 1.029198096685278e-06,
+      "loss": 0.1384,
+      "step": 12320
+    },
+    {
+      "epoch": 2.2,
+      "grad_norm": 1.1899516582489014,
+      "learning_rate": 1.0158459279855632e-06,
+      "loss": 0.1433,
+      "step": 12352
+    },
+    {
+      "epoch": 2.2,
+      "grad_norm": 1.060065507888794,
+      "learning_rate": 1.0025588178858372e-06,
+      "loss": 0.1456,
+      "step": 12384
+    },
+    {
+      "epoch": 2.21,
+      "grad_norm": 1.1489146947860718,
+      "learning_rate": 9.893373488328953e-07,
+      "loss": 0.1433,
+      "step": 12416
+    },
+    {
+      "epoch": 2.21,
+      "grad_norm": 1.3114417791366577,
+      "learning_rate": 9.761821003961246e-07,
+      "loss": 0.1467,
+      "step": 12448
+    },
+    {
+      "epoch": 2.22,
+      "grad_norm": 1.3255183696746826,
+      "learning_rate": 9.630936492421005e-07,
+      "loss": 0.1463,
+      "step": 12480
+    },
+    {
+      "epoch": 2.22,
+      "grad_norm": 1.2642742395401,
+      "learning_rate": 9.500725691093085e-07,
+      "loss": 0.1525,
+      "step": 12512
+    },
+    {
+      "epoch": 2.23,
+      "grad_norm": 1.2281138896942139,
+      "learning_rate": 9.371194307829895e-07,
+      "loss": 0.1383,
+      "step": 12544
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 0.9627875089645386,
+      "learning_rate": 9.242348020701295e-07,
+      "loss": 0.1642,
+      "step": 12576
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 1.1187249422073364,
+      "learning_rate": 9.114192477745568e-07,
+      "loss": 0.1439,
+      "step": 12608
+    },
+    {
+      "epoch": 2.25,
+      "grad_norm": 1.0410840511322021,
+      "learning_rate": 8.986733296721931e-07,
+      "loss": 0.142,
+      "step": 12640
+    },
+    {
+      "epoch": 2.25,
+      "grad_norm": 1.2345024347305298,
+      "learning_rate": 8.859976064864235e-07,
+      "loss": 0.1512,
+      "step": 12672
+    },
+    {
+      "epoch": 2.26,
+      "grad_norm": 1.6558443307876587,
+      "learning_rate": 8.733926338636056e-07,
+      "loss": 0.1363,
+      "step": 12704
+    },
+    {
+      "epoch": 2.26,
+      "grad_norm": 1.3118538856506348,
+      "learning_rate": 8.608589643487128e-07,
+      "loss": 0.1471,
+      "step": 12736
+    },
+    {
+      "epoch": 2.27,
+      "grad_norm": 1.1155567169189453,
+      "learning_rate": 8.483971473611133e-07,
+      "loss": 0.1396,
+      "step": 12768
+    },
+    {
+      "epoch": 2.28,
+      "grad_norm": 1.0880179405212402,
+      "learning_rate": 8.360077291704821e-07,
+      "loss": 0.1413,
+      "step": 12800
+    },
+    {
+      "epoch": 2.28,
+      "grad_norm": 0.9752321839332581,
+      "learning_rate": 8.236912528728647e-07,
+      "loss": 0.146,
+      "step": 12832
+    },
+    {
+      "epoch": 2.29,
+      "grad_norm": 0.9778379201889038,
+      "learning_rate": 8.114482583668576e-07,
+      "loss": 0.1403,
+      "step": 12864
+    },
+    {
+      "epoch": 2.29,
+      "grad_norm": 0.8760839700698853,
+      "learning_rate": 7.99279282329952e-07,
+      "loss": 0.148,
+      "step": 12896
+    },
+    {
+      "epoch": 2.3,
+      "grad_norm": 1.187658667564392,
+      "learning_rate": 7.871848581950039e-07,
+      "loss": 0.132,
+      "step": 12928
+    },
+    {
+      "epoch": 2.3,
+      "grad_norm": 0.9668059349060059,
+      "learning_rate": 7.751655161268481e-07,
+      "loss": 0.1424,
+      "step": 12960
+    },
+    {
+      "epoch": 2.31,
+      "grad_norm": 1.1318392753601074,
+      "learning_rate": 7.632217829990668e-07,
+      "loss": 0.1516,
+      "step": 12992
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 1.3520994186401367,
+      "learning_rate": 7.513541823708828e-07,
+      "loss": 0.1495,
+      "step": 13024
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 1.3352413177490234,
+      "learning_rate": 7.395632344642173e-07,
+      "loss": 0.1446,
+      "step": 13056
+    },
+    {
+      "epoch": 2.33,
+      "grad_norm": 1.0273305177688599,
+      "learning_rate": 7.278494561408833e-07,
+      "loss": 0.1391,
+      "step": 13088
+    },
+    {
+      "epoch": 2.33,
+      "grad_norm": 1.2872681617736816,
+      "learning_rate": 7.162133608799271e-07,
+      "loss": 0.1391,
+      "step": 13120
+    },
+    {
+      "epoch": 2.34,
+      "grad_norm": 1.0563528537750244,
+      "learning_rate": 7.046554587551216e-07,
+      "loss": 0.1521,
+      "step": 13152
+    },
+    {
+      "epoch": 2.34,
+      "grad_norm": 1.1487845182418823,
+      "learning_rate": 6.931762564126074e-07,
+      "loss": 0.1411,
+      "step": 13184
+    },
+    {
+      "epoch": 2.35,
+      "grad_norm": 1.058159351348877,
+      "learning_rate": 6.817762570486791e-07,
+      "loss": 0.1424,
+      "step": 13216
+    },
+    {
+      "epoch": 2.36,
+      "grad_norm": 1.249377727508545,
+      "learning_rate": 6.704559603877367e-07,
+      "loss": 0.1448,
+      "step": 13248
+    },
+    {
+      "epoch": 2.36,
+      "grad_norm": 0.9334893226623535,
+      "learning_rate": 6.592158626603689e-07,
+      "loss": 0.1384,
+      "step": 13280
+    },
+    {
+      "epoch": 2.37,
+      "grad_norm": 1.5639148950576782,
+      "learning_rate": 6.480564565816091e-07,
+      "loss": 0.1426,
+      "step": 13312
+    },
+    {
+      "epoch": 2.37,
+      "grad_norm": 1.0596867799758911,
+      "learning_rate": 6.369782313293335e-07,
+      "loss": 0.1358,
+      "step": 13344
+    },
+    {
+      "epoch": 2.38,
+      "grad_norm": 0.8567415475845337,
+      "learning_rate": 6.259816725228158e-07,
+      "loss": 0.1465,
+      "step": 13376
+    },
+    {
+      "epoch": 2.38,
+      "grad_norm": 1.1086764335632324,
+      "learning_rate": 6.150672622014459e-07,
+      "loss": 0.1538,
+      "step": 13408
+    },
+    {
+      "epoch": 2.39,
+      "grad_norm": 1.1636631488800049,
+      "learning_rate": 6.042354788035943e-07,
+      "loss": 0.1389,
+      "step": 13440
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 1.274596929550171,
+      "learning_rate": 5.934867971456384e-07,
+      "loss": 0.1464,
+      "step": 13472
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 1.173563003540039,
+      "learning_rate": 5.828216884011553e-07,
+      "loss": 0.1435,
+      "step": 13504
+    },
+    {
+      "epoch": 2.41,
+      "grad_norm": 1.0788921117782593,
+      "learning_rate": 5.722406200802613e-07,
+      "loss": 0.145,
+      "step": 13536
+    },
+    {
+      "epoch": 2.41,
+      "grad_norm": 1.1613490581512451,
+      "learning_rate": 5.617440560091212e-07,
+      "loss": 0.1474,
+      "step": 13568
+    },
+    {
+      "epoch": 2.42,
+      "grad_norm": 0.9075080156326294,
+      "learning_rate": 5.513324563096167e-07,
+      "loss": 0.1423,
+      "step": 13600
+    },
+    {
+      "epoch": 2.42,
+      "grad_norm": 1.1296495199203491,
+      "learning_rate": 5.41006277379173e-07,
+      "loss": 0.1506,
+      "step": 13632
+    },
+    {
+      "epoch": 2.43,
+      "grad_norm": 1.2199699878692627,
+      "learning_rate": 5.307659718707603e-07,
+      "loss": 0.1459,
+      "step": 13664
+    },
+    {
+      "epoch": 2.44,
+      "grad_norm": 1.2415364980697632,
+      "learning_rate": 5.20611988673041e-07,
+      "loss": 0.1459,
+      "step": 13696
+    },
+    {
+      "epoch": 2.44,
+      "grad_norm": 0.9814534187316895,
+      "learning_rate": 5.105447728907012e-07,
+      "loss": 0.1405,
+      "step": 13728
+    },
+    {
+      "epoch": 2.45,
+      "grad_norm": 1.1437889337539673,
+      "learning_rate": 5.00564765824936e-07,
+      "loss": 0.147,
+      "step": 13760
+    },
+    {
+      "epoch": 2.45,
+      "grad_norm": 1.1459242105484009,
+      "learning_rate": 4.906724049541023e-07,
+      "loss": 0.1454,
+      "step": 13792
+    },
+    {
+      "epoch": 2.46,
+      "grad_norm": 1.093807578086853,
+      "learning_rate": 4.808681239145479e-07,
+      "loss": 0.1448,
+      "step": 13824
+    },
+    {
+      "epoch": 2.46,
+      "grad_norm": 1.1457182168960571,
+      "learning_rate": 4.711523524815978e-07,
+      "loss": 0.1391,
+      "step": 13856
+    },
+    {
+      "epoch": 2.47,
+      "grad_norm": 1.0422513484954834,
+      "learning_rate": 4.615255165507146e-07,
+      "loss": 0.1435,
+      "step": 13888
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 1.1171213388442993,
+      "learning_rate": 4.5198803811883326e-07,
+      "loss": 0.1545,
+      "step": 13920
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 1.3410863876342773,
+      "learning_rate": 4.4254033526585917e-07,
+      "loss": 0.1526,
+      "step": 13952
+    },
+    {
+      "epoch": 2.49,
+      "grad_norm": 0.9821498394012451,
+      "learning_rate": 4.331828221363424e-07,
+      "loss": 0.1407,
+      "step": 13984
+    },
+    {
+      "epoch": 2.49,
+      "grad_norm": 1.2533886432647705,
+      "learning_rate": 4.239159089213246e-07,
+      "loss": 0.1358,
+      "step": 14016
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.9848787784576416,
+      "learning_rate": 4.147400018403544e-07,
+      "loss": 0.1449,
+      "step": 14048
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.9092288613319397,
+      "learning_rate": 4.056555031236878e-07,
+      "loss": 0.1433,
+      "step": 14080
+    },
+    {
+      "epoch": 2.51,
+      "grad_norm": 1.3839354515075684,
+      "learning_rate": 3.966628109946469e-07,
+      "loss": 0.1494,
+      "step": 14112
+    },
+    {
+      "epoch": 2.51,
+      "grad_norm": 0.9744381904602051,
+      "learning_rate": 3.877623196521707e-07,
+      "loss": 0.1426,
+      "step": 14144
+    },
+    {
+      "epoch": 2.52,
+      "grad_norm": 1.165624976158142,
+      "learning_rate": 3.7895441925353356e-07,
+      "loss": 0.1418,
+      "step": 14176
+    },
+    {
+      "epoch": 2.53,
+      "grad_norm": 0.9751477241516113,
+      "learning_rate": 3.702394958972391e-07,
+      "loss": 0.1479,
+      "step": 14208
+    },
+    {
+      "epoch": 2.53,
+      "grad_norm": 1.0645439624786377,
+      "learning_rate": 3.616179316061011e-07,
+      "loss": 0.1373,
+      "step": 14240
+    },
+    {
+      "epoch": 2.54,
+      "grad_norm": 0.9858296513557434,
+      "learning_rate": 3.5309010431049284e-07,
+      "loss": 0.1367,
+      "step": 14272
+    },
+    {
+      "epoch": 2.54,
+      "grad_norm": 1.0521584749221802,
+      "learning_rate": 3.44656387831781e-07,
+      "loss": 0.1421,
+      "step": 14304
+    },
+    {
+      "epoch": 2.55,
+      "grad_norm": 1.2546510696411133,
+      "learning_rate": 3.363171518659408e-07,
+      "loss": 0.1384,
+      "step": 14336
+    },
+    {
+      "epoch": 2.55,
+      "grad_norm": 1.0795034170150757,
+      "learning_rate": 3.280727619673496e-07,
+      "loss": 0.1463,
+      "step": 14368
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 1.1764500141143799,
+      "learning_rate": 3.199235795327615e-07,
+      "loss": 0.1499,
+      "step": 14400
+    },
+    {
+      "epoch": 2.57,
+      "grad_norm": 1.1234067678451538,
+      "learning_rate": 3.1186996178546674e-07,
+      "loss": 0.1497,
+      "step": 14432
+    },
+    {
+      "epoch": 2.57,
+      "grad_norm": 0.9825364947319031,
+      "learning_rate": 3.039122617596302e-07,
+      "loss": 0.1514,
+      "step": 14464
+    },
+    {
+      "epoch": 2.58,
+      "grad_norm": 1.263085961341858,
+      "learning_rate": 2.960508282848215e-07,
+      "loss": 0.1476,
+      "step": 14496
+    },
+    {
+      "epoch": 2.58,
+      "grad_norm": 1.084181308746338,
+      "learning_rate": 2.8828600597071597e-07,
+      "loss": 0.1308,
+      "step": 14528
+    },
+    {
+      "epoch": 2.59,
+      "grad_norm": 1.1697498559951782,
+      "learning_rate": 2.8061813519199536e-07,
+      "loss": 0.1348,
+      "step": 14560
+    },
+    {
+      "epoch": 2.59,
+      "grad_norm": 1.3982306718826294,
+      "learning_rate": 2.7304755207342467e-07,
+      "loss": 0.1455,
+      "step": 14592
+    },
+    {
+      "epoch": 2.6,
+      "grad_norm": 1.1802705526351929,
+      "learning_rate": 2.655745884751157e-07,
+      "loss": 0.1437,
+      "step": 14624
+    },
+    {
+      "epoch": 2.61,
+      "grad_norm": 1.0200531482696533,
+      "learning_rate": 2.581995719779856e-07,
+      "loss": 0.1394,
+      "step": 14656
+    },
+    {
+      "epoch": 2.61,
+      "grad_norm": 1.1693042516708374,
+      "learning_rate": 2.5092282586939187e-07,
+      "loss": 0.151,
+      "step": 14688
+    },
+    {
+      "epoch": 2.62,
+      "grad_norm": 1.116024374961853,
+      "learning_rate": 2.437446691289616e-07,
+      "loss": 0.1478,
+      "step": 14720
+    },
+    {
+      "epoch": 2.62,
+      "grad_norm": 1.05259370803833,
+      "learning_rate": 2.3666541641461231e-07,
+      "loss": 0.1436,
+      "step": 14752
+    },
+    {
+      "epoch": 2.63,
+      "grad_norm": 1.0545703172683716,
+      "learning_rate": 2.2968537804875485e-07,
+      "loss": 0.1379,
+      "step": 14784
+    },
+    {
+      "epoch": 2.63,
+      "grad_norm": 1.0618197917938232,
+      "learning_rate": 2.228048600046928e-07,
+      "loss": 0.1409,
+      "step": 14816
+    },
+    {
+      "epoch": 2.64,
+      "grad_norm": 1.389092206954956,
+      "learning_rate": 2.1602416389320922e-07,
+      "loss": 0.1499,
+      "step": 14848
+    },
+    {
+      "epoch": 2.65,
+      "grad_norm": 1.0467108488082886,
+      "learning_rate": 2.0934358694934347e-07,
+      "loss": 0.1406,
+      "step": 14880
+    },
+    {
+      "epoch": 2.65,
+      "grad_norm": 1.1932706832885742,
+      "learning_rate": 2.0276342201936637e-07,
+      "loss": 0.1468,
+      "step": 14912
+    },
+    {
+      "epoch": 2.66,
+      "grad_norm": 1.2286850214004517,
+      "learning_rate": 1.9628395754793777e-07,
+      "loss": 0.1457,
+      "step": 14944
+    },
+    {
+      "epoch": 2.66,
+      "grad_norm": 0.9705607891082764,
+      "learning_rate": 1.899054775654663e-07,
+      "loss": 0.1439,
+      "step": 14976
+    },
+    {
+      "epoch": 2.67,
+      "grad_norm": 0.9110348224639893,
+      "learning_rate": 1.8362826167565796e-07,
+      "loss": 0.1439,
+      "step": 15008
+    },
+    {
+      "epoch": 2.67,
+      "grad_norm": 0.9858996272087097,
+      "learning_rate": 1.774525850432568e-07,
+      "loss": 0.1528,
+      "step": 15040
+    },
+    {
+      "epoch": 2.68,
+      "grad_norm": 1.1253962516784668,
+      "learning_rate": 1.7137871838198817e-07,
+      "loss": 0.1408,
+      "step": 15072
+    },
+    {
+      "epoch": 2.69,
+      "grad_norm": 1.5971510410308838,
+      "learning_rate": 1.654069279426873e-07,
+      "loss": 0.1497,
+      "step": 15104
+    },
+    {
+      "epoch": 2.69,
+      "grad_norm": 0.8475412130355835,
+      "learning_rate": 1.5953747550162907e-07,
+      "loss": 0.1456,
+      "step": 15136
+    },
+    {
+      "epoch": 2.7,
+      "grad_norm": 0.9866968989372253,
+      "learning_rate": 1.537706183490545e-07,
+      "loss": 0.1349,
+      "step": 15168
+    },
+    {
+      "epoch": 2.7,
+      "grad_norm": 1.1067461967468262,
+      "learning_rate": 1.481066092778913e-07,
+      "loss": 0.1457,
+      "step": 15200
+    },
+    {
+      "epoch": 2.71,
+      "grad_norm": 1.1080329418182373,
+      "learning_rate": 1.4254569657267235e-07,
+      "loss": 0.146,
+      "step": 15232
+    },
+    {
+      "epoch": 2.71,
+      "grad_norm": 0.992157518863678,
+      "learning_rate": 1.370881239986524e-07,
+      "loss": 0.1439,
+      "step": 15264
+    },
+    {
+      "epoch": 2.72,
+      "grad_norm": 1.032788872718811,
+      "learning_rate": 1.3173413079112128e-07,
+      "loss": 0.1369,
+      "step": 15296
+    },
+    {
+      "epoch": 2.73,
+      "grad_norm": 0.9706469774246216,
+      "learning_rate": 1.264839516449204e-07,
+      "loss": 0.136,
+      "step": 15328
+    },
+    {
+      "epoch": 2.73,
+      "grad_norm": 1.1187324523925781,
+      "learning_rate": 1.2133781670415013e-07,
+      "loss": 0.1359,
+      "step": 15360
+    },
+    {
+      "epoch": 2.74,
+      "grad_norm": 1.1595239639282227,
+      "learning_rate": 1.1629595155208424e-07,
+      "loss": 0.1401,
+      "step": 15392
+    },
+    {
+      "epoch": 2.74,
+      "grad_norm": 1.087785243988037,
+      "learning_rate": 1.1135857720128151e-07,
+      "loss": 0.1358,
+      "step": 15424
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 1.0765306949615479,
+      "learning_rate": 1.0652591008389557e-07,
+      "loss": 0.1438,
+      "step": 15456
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 1.1016918420791626,
+      "learning_rate": 1.0179816204218928e-07,
+      "loss": 0.1373,
+      "step": 15488
+    },
+    {
+      "epoch": 2.76,
+      "grad_norm": 1.0536975860595703,
+      "learning_rate": 9.717554031924842e-08,
+      "loss": 0.1349,
+      "step": 15520
+    },
+    {
+      "epoch": 2.77,
+      "grad_norm": 0.8933613300323486,
+      "learning_rate": 9.265824754989467e-08,
+      "loss": 0.1316,
+      "step": 15552
+    },
+    {
+      "epoch": 2.77,
+      "grad_norm": 0.9983497858047485,
+      "learning_rate": 8.824648175180722e-08,
+      "loss": 0.1346,
+      "step": 15584
+    },
+    {
+      "epoch": 2.78,
+      "grad_norm": 1.0600674152374268,
+      "learning_rate": 8.394043631683862e-08,
+      "loss": 0.1533,
+      "step": 15616
+    },
+    {
+      "epoch": 2.78,
+      "grad_norm": 1.0515772104263306,
+      "learning_rate": 7.974030000253986e-08,
+      "loss": 0.139,
+      "step": 15648
+    },
+    {
+      "epoch": 2.79,
+      "grad_norm": 1.4163565635681152,
+      "learning_rate": 7.564625692388499e-08,
+      "loss": 0.1323,
+      "step": 15680
+    },
+    {
+      "epoch": 2.79,
+      "grad_norm": 1.0619480609893799,
+      "learning_rate": 7.165848654519969e-08,
+      "loss": 0.1373,
+      "step": 15712
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 0.9783567786216736,
+      "learning_rate": 6.777716367229764e-08,
+      "loss": 0.1525,
+      "step": 15744
+    },
+    {
+      "epoch": 2.81,
+      "grad_norm": 1.0433095693588257,
+      "learning_rate": 6.400245844481262e-08,
+      "loss": 0.1409,
+      "step": 15776
+    },
+    {
+      "epoch": 2.81,
+      "grad_norm": 1.258354663848877,
+      "learning_rate": 6.033453632874498e-08,
+      "loss": 0.1402,
+      "step": 15808
+    },
+    {
+      "epoch": 2.82,
+      "grad_norm": 1.1823972463607788,
+      "learning_rate": 5.677355810920604e-08,
+      "loss": 0.1418,
+      "step": 15840
+    },
+    {
+      "epoch": 2.82,
+      "grad_norm": 1.2745051383972168,
+      "learning_rate": 5.3319679883370724e-08,
+      "loss": 0.1471,
+      "step": 15872
+    },
+    {
+      "epoch": 2.83,
+      "grad_norm": 1.238215684890747,
+      "learning_rate": 4.9973053053634365e-08,
+      "loss": 0.1426,
+      "step": 15904
+    },
+    {
+      "epoch": 2.83,
+      "grad_norm": 1.0394669771194458,
+      "learning_rate": 4.6733824320976674e-08,
+      "loss": 0.1335,
+      "step": 15936
+    },
+    {
+      "epoch": 2.84,
+      "grad_norm": 1.2872110605239868,
+      "learning_rate": 4.360213567853072e-08,
+      "loss": 0.1544,
+      "step": 15968
+    },
+    {
+      "epoch": 2.84,
+      "grad_norm": 1.2786180973052979,
+      "learning_rate": 4.057812440535797e-08,
+      "loss": 0.1461,
+      "step": 16000
+    },
+    {
+      "epoch": 2.85,
+      "grad_norm": 1.169412612915039,
+      "learning_rate": 3.766192306043165e-08,
+      "loss": 0.1413,
+      "step": 16032
+    },
+    {
+      "epoch": 2.86,
+      "grad_norm": 1.2436180114746094,
+      "learning_rate": 3.485365947682562e-08,
+      "loss": 0.1357,
+      "step": 16064
+    },
+    {
+      "epoch": 2.86,
+      "grad_norm": 1.1065387725830078,
+      "learning_rate": 3.215345675611076e-08,
+      "loss": 0.1472,
+      "step": 16096
+    },
+    {
+      "epoch": 2.87,
+      "grad_norm": 1.00310218334198,
+      "learning_rate": 2.9561433262957072e-08,
+      "loss": 0.1499,
+      "step": 16128
+    },
+    {
+      "epoch": 2.87,
+      "grad_norm": 0.9328859448432922,
+      "learning_rate": 2.7077702619948963e-08,
+      "loss": 0.1376,
+      "step": 16160
+    },
+    {
+      "epoch": 2.88,
+      "grad_norm": 1.122558832168579,
+      "learning_rate": 2.4702373702600868e-08,
+      "loss": 0.1461,
+      "step": 16192
+    },
+    {
+      "epoch": 2.88,
+      "grad_norm": 1.1374387741088867,
+      "learning_rate": 2.2435550634585522e-08,
+      "loss": 0.1427,
+      "step": 16224
+    },
+    {
+      "epoch": 2.89,
+      "grad_norm": 1.102001428604126,
+      "learning_rate": 2.027733278317151e-08,
+      "loss": 0.1402,
+      "step": 16256
+    },
+    {
+      "epoch": 2.9,
+      "grad_norm": 1.3255332708358765,
+      "learning_rate": 1.822781475486507e-08,
+      "loss": 0.1427,
+      "step": 16288
+    },
+    {
+      "epoch": 2.9,
+      "grad_norm": 1.1129345893859863,
+      "learning_rate": 1.628708639126425e-08,
+      "loss": 0.1443,
+      "step": 16320
+    },
+    {
+      "epoch": 2.91,
+      "grad_norm": 0.9628246426582336,
+      "learning_rate": 1.4455232765120397e-08,
+      "loss": 0.1425,
+      "step": 16352
+    },
+    {
+      "epoch": 2.91,
+      "grad_norm": 1.2334058284759521,
+      "learning_rate": 1.273233417660863e-08,
+      "loss": 0.134,
+      "step": 16384
+    },
+    {
+      "epoch": 2.92,
+      "grad_norm": 1.1855486631393433,
+      "learning_rate": 1.1118466149808994e-08,
+      "loss": 0.1403,
+      "step": 16416
+    },
+    {
+      "epoch": 2.92,
+      "grad_norm": 1.0412129163742065,
+      "learning_rate": 9.61369942939383e-09,
+      "loss": 0.1369,
+      "step": 16448
+    },
+    {
+      "epoch": 2.93,
+      "grad_norm": 1.2178360223770142,
+      "learning_rate": 8.218099977528871e-09,
+      "loss": 0.1346,
+      "step": 16480
+    },
+    {
+      "epoch": 2.94,
+      "grad_norm": 1.1358833312988281,
+      "learning_rate": 6.9317289709799896e-09,
+      "loss": 0.1504,
+      "step": 16512
+    },
+    {
+      "epoch": 2.94,
+      "grad_norm": 1.1772416830062866,
+      "learning_rate": 5.754642798432297e-09,
+      "loss": 0.144,
+      "step": 16544
+    },
+    {
+      "epoch": 2.95,
+      "grad_norm": 1.4765156507492065,
+      "learning_rate": 4.686893058018227e-09,
+      "loss": 0.1531,
+      "step": 16576
+    },
+    {
+      "epoch": 2.95,
+      "grad_norm": 1.0203588008880615,
+      "learning_rate": 3.728526555056289e-09,
+      "loss": 0.1439,
+      "step": 16608
+    },
+    {
+      "epoch": 2.96,
+      "grad_norm": 1.053261637687683,
+      "learning_rate": 2.879585299997434e-09,
+      "loss": 0.1438,
+      "step": 16640
+    },
+    {
+      "epoch": 2.96,
+      "grad_norm": 1.2388105392456055,
+      "learning_rate": 2.1401065065859704e-09,
+      "loss": 0.145,
+      "step": 16672
+    },
+    {
+      "epoch": 2.97,
+      "grad_norm": 0.9954524040222168,
+      "learning_rate": 1.5101225902267036e-09,
+      "loss": 0.147,
+      "step": 16704
+    },
+    {
+      "epoch": 2.98,
+      "grad_norm": 1.2384732961654663,
+      "learning_rate": 9.89661166564404e-10,
+      "loss": 0.1492,
+      "step": 16736
+    },
+    {
+      "epoch": 2.98,
+      "grad_norm": 0.8787427544593811,
+      "learning_rate": 5.787450502728331e-10,
+      "loss": 0.1299,
+      "step": 16768
+    },
+    {
+      "epoch": 2.99,
+      "grad_norm": 0.9824705719947815,
+      "learning_rate": 2.7739225405609694e-10,
+      "loss": 0.1428,
+      "step": 16800
+    },
+    {
+      "epoch": 2.99,
+      "grad_norm": 1.0306531190872192,
+      "learning_rate": 8.561598785705727e-11,
+      "loss": 0.1434,
+      "step": 16832
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.0343343019485474,
+      "learning_rate": 3.424658279460591e-12,
+      "loss": 0.1521,
+      "step": 16864
+    }
+  ],
+  "logging_steps": 32,
+  "max_steps": 16872,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 5624,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf3b1dc61204b974a8ab5a5bd8ff339ab3ec4e4c342560ccdb08e638c7543120
+size 7032