Upload folder using huggingface_hub

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<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 +1,31 @@
+{
+  "_name_or_path": "/apdcephfs/share_302520718/share_info/llm_models/Qwen2.5-Math-7B-YARN",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 128245,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 4096,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.46.3",
+  "use_cache": false,
+  "use_mrope": false,
+  "use_sliding_window": false,
+  "vocab_size": 151665
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.46.3"
+}

latest

@@ -0,0 +1 @@
+global_step412

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8797fadfbd9b015266268e019b68e77c2fe58f13a73536af7c927a123d58d2c
+size 4874800744

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79b3deafa8d5f812fd996c7eab826f19c3508950dd14bb99ede0e6cd43106adb
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff2baa6ba8a5d47ba243846664d432853c0ec013bdb5363c1feb07a58cad4547
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b60b28e72aa862ccf9c93df847fe652d5562991eb91ad56faab7880262d90b21
+size 1087134848

model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15225512960
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00003-of-00004.safetensors"
+  }
+}

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93e79425782fbaba0185c063597dbfabf93e3b6db832c2d36f3ce456c42dd551
+size 15984

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50abb0a1a7651cbc3691325bc5ab5d4be5a0c7d24f1c12e69bd014a3a24c33c9
+size 15984

rng_state_10.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe3eb6f7a9241c6d7a241dfbad7d58e440ce116034ca23b64968ee350c69658e
+size 15997

rng_state_11.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9165266801c4638bcb7bcdc168a19865c1ec3c695758e65f7e84e14a12f697b0
+size 15997

rng_state_12.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:211697c3693a277551875a723c000a4cd1106b6f7a2f400a28c00614944bbbb0
+size 15997

rng_state_13.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0be0f7913679b07ce8d9bc067e134c537e9fa8f4252c0b2bf59a1eb3f5529473
+size 15997

rng_state_14.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1118f2e5d57689b708621afa8e43f721ebbfd15a34386caad646b5ca0c0ff774
+size 15997

rng_state_15.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f9387225ddf513b349ba48cc743b8719f4a02bef764ddaa728ba7d3b066cc8c
+size 15997

rng_state_16.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bed24c7f478571961306a7678d7ed6b060266eeaf1375f8d1ba8f4335c1c5f84
+size 15997

rng_state_17.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:864ffd528f45bdd9ae3dfe9a83328a6f6ba171270654d4ff7a4d2b3ab651ffdd
+size 15997

rng_state_18.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4eb2843bf9eb20d51ee449f5164b3951f287f93ab0b5ff41e6238166e708577f
+size 15997

rng_state_19.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f88f68dda284d1a80633a6f91eece016fd3c6660fc2aa0f19604f2d22c9199e
+size 15997

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15400e019b0249d510ea9f3bff5371a3743a5d5998136695edcc1575a13612fb
+size 15984

rng_state_20.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:391c22c90584acd3f5301f7880958cd35dfaaf0b478ffa2af3affb3904f28f17
+size 15997

rng_state_21.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a46d303a45c7a2a3811e5ea69a3585b3c7131b1d4574d729993079daa27f612
+size 15997

rng_state_22.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d77a627a78f5c2a3ab8c2138b501e31ac2e94fdf84679f7be1689d6946dc286
+size 15997

rng_state_23.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5aa17bdd85dae0d5f65be0e6f8f7b640b12685faffc737f633e82ff7d5e16fa0
+size 15997

rng_state_24.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d473f8a99d19889747f319d4c9c57305776298b2f128ec96eea8785a71b5935e
+size 15997

rng_state_25.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7663813662ee05775f5d92ac2650c60bbbfa8d23d67c32bf25a843c93ae5b8d2
+size 15997

rng_state_26.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a207218fcc684e6c7347802271e3026116e21272033c5baa26f3d29f64018ae4
+size 15997

rng_state_27.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a2c2f0dae2950c62e254db441454a5f5dd9791f079dc6bc995082af288b82c6
+size 15997

rng_state_28.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf39ec1a13eaa914868f2a92a3f3dd4e0aaace4cb7691ef99483377aaaaff184
+size 15997

rng_state_29.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf4eb2f894986284e7b5d8fe753548030a81685b70fb92de3953a1219d0b1932
+size 15997

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51d2ccb318422ff9de081328429a7f015048cde8ad6e03f07a090a8c2393490d
+size 15984

rng_state_30.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:942bf687cd6ac492281255e0aa8ee57201b86e4a1bdbaf4e28713cec757dbe7a
+size 15997

rng_state_31.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c730254bf904e6a7038ad01219c912d3c2a1865d8cbe3961ca59bb1b3a3c68ed
+size 15997

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60948dee683036263219dc9f7164ca65aa40884bd4b2bfa2ad76e5a6d6ceab4b
+size 15984

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76e3b4364935bca9ba941ba321de8ae6fad942732c39696c8f3b77e1d4dd238c
+size 15984

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3339c1057bebad97f8d0c021d2c1f5c7a22dad29151e2ae1208a1badf271a2d
+size 15984

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf711b875d01abb8597e49a64a852588ddad3e847b4d814a7e0aa65a71f6375c
+size 15984

rng_state_8.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6715633e22f0ddb19985bc5dbb1fee0dc7bcee43dee18d69bb9fc8f99e6449e
+size 15984

rng_state_9.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5366ab951913d88bb366d8f7b8392120bbc1ae8784b3cd71f434eee308eb5613
+size 15984

special_tokens_map.json

@@ -0,0 +1,45 @@
+{
+  "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": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "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
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,225 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "128244": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "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
+    }
+  },
+  "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": "<s>",
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'Please reason step by step, and put your final answer within \\\\boxed{}.' }}\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\\nPlease reason step by step, and put your final answer within \\\\boxed{}.<|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": 8000,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "truncation_side": "left",
+  "unk_token": "<unk>"
+}

trainer_state.json

@@ -0,0 +1,3773 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 2000.0,
+  "global_step": 412,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.009708737864077669,
+      "grad_norm": 5.575473016079204,
+      "kl": 0.0,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 0.4845,
+      "step": 1,
+      "step_loss": 0.490234375
+    },
+    {
+      "epoch": 0.019417475728155338,
+      "grad_norm": 9.7143780359321,
+      "kl": 0.0020599365234375,
+      "learning_rate": 2.438044511330269e-06,
+      "loss": 0.4845,
+      "step": 2,
+      "step_loss": 0.48046875
+    },
+    {
+      "epoch": 0.02912621359223301,
+      "grad_norm": 7.3775865577265884,
+      "kl": 0.002471923828125,
+      "learning_rate": 3.5717278751869343e-06,
+      "loss": 0.4945,
+      "step": 3,
+      "step_loss": 0.546875
+    },
+    {
+      "epoch": 0.038834951456310676,
+      "grad_norm": 15.596262146563706,
+      "kl": 0.002349853515625,
+      "learning_rate": 4.376089022660538e-06,
+      "loss": 0.4586,
+      "step": 4,
+      "step_loss": 0.32421875
+    },
+    {
+      "epoch": 0.04854368932038835,
+      "grad_norm": 6.642139360876633,
+      "kl": 0.021240234375,
+      "learning_rate": 5e-06,
+      "loss": 0.4483,
+      "step": 5,
+      "step_loss": 0.71875
+    },
+    {
+      "epoch": 0.05825242718446602,
+      "grad_norm": 10.4028109346367,
+      "kl": 0.037353515625,
+      "learning_rate": 4.999957311534636e-06,
+      "loss": 0.4224,
+      "step": 6,
+      "step_loss": 0.306640625
+    },
+    {
+      "epoch": 0.06796116504854369,
+      "grad_norm": 8.965364568204704,
+      "kl": 0.043212890625,
+      "learning_rate": 4.99982924775837e-06,
+      "loss": 0.3771,
+      "step": 7,
+      "step_loss": 0.361328125
+    },
+    {
+      "epoch": 0.07766990291262135,
+      "grad_norm": 5.86329979229663,
+      "kl": 0.06396484375,
+      "learning_rate": 4.999615813530619e-06,
+      "loss": 0.372,
+      "step": 8,
+      "step_loss": 0.345703125
+    },
+    {
+      "epoch": 0.08737864077669903,
+      "grad_norm": 7.680252752913332,
+      "kl": 0.07763671875,
+      "learning_rate": 4.999317016950212e-06,
+      "loss": 0.4717,
+      "step": 9,
+      "step_loss": 0.482421875
+    },
+    {
+      "epoch": 0.0970873786407767,
+      "grad_norm": 7.28382768968269,
+      "kl": 0.12060546875,
+      "learning_rate": 4.998932869355074e-06,
+      "loss": 0.3988,
+      "step": 10,
+      "step_loss": 0.259765625
+    },
+    {
+      "epoch": 0.10679611650485436,
+      "grad_norm": 3.742924436507147,
+      "kl": 0.07958984375,
+      "learning_rate": 4.998463385321802e-06,
+      "loss": 0.3597,
+      "step": 11,
+      "step_loss": 0.42578125
+    },
+    {
+      "epoch": 0.11650485436893204,
+      "grad_norm": 2.077338507259391,
+      "kl": 0.0986328125,
+      "learning_rate": 4.997908582665111e-06,
+      "loss": 0.3828,
+      "step": 12,
+      "step_loss": 0.2197265625
+    },
+    {
+      "epoch": 0.1262135922330097,
+      "grad_norm": 2.0936762099287187,
+      "kl": 0.2578125,
+      "learning_rate": 4.997268482437153e-06,
+      "loss": 0.3686,
+      "step": 13,
+      "step_loss": 0.40234375
+    },
+    {
+      "epoch": 0.13592233009708737,
+      "grad_norm": 1.5539690974487896,
+      "kl": 0.11572265625,
+      "learning_rate": 4.9965431089267265e-06,
+      "loss": 0.3481,
+      "step": 14,
+      "step_loss": 0.318359375
+    },
+    {
+      "epoch": 0.14563106796116504,
+      "grad_norm": 11.763990149324268,
+      "kl": 0.19921875,
+      "learning_rate": 4.9957324896583495e-06,
+      "loss": 0.3518,
+      "step": 15,
+      "step_loss": 0.419921875
+    },
+    {
+      "epoch": 0.1553398058252427,
+      "grad_norm": 1.6021280918226926,
+      "kl": 0.2080078125,
+      "learning_rate": 4.9948366553912146e-06,
+      "loss": 0.3468,
+      "step": 16,
+      "step_loss": 0.498046875
+    },
+    {
+      "epoch": 0.1650485436893204,
+      "grad_norm": 2.243950658979944,
+      "kl": 0.1279296875,
+      "learning_rate": 4.993855640118024e-06,
+      "loss": 0.3336,
+      "step": 17,
+      "step_loss": 0.267578125
+    },
+    {
+      "epoch": 0.17475728155339806,
+      "grad_norm": 1.5344029876580454,
+      "kl": 0.2216796875,
+      "learning_rate": 4.992789481063699e-06,
+      "loss": 0.3456,
+      "step": 18,
+      "step_loss": 0.3671875
+    },
+    {
+      "epoch": 0.18446601941747573,
+      "grad_norm": 1.7390327115176747,
+      "kl": 0.1259765625,
+      "learning_rate": 4.9916382186839665e-06,
+      "loss": 0.3427,
+      "step": 19,
+      "step_loss": 0.298828125
+    },
+    {
+      "epoch": 0.1941747572815534,
+      "grad_norm": 5.2081774675570855,
+      "kl": 0.1591796875,
+      "learning_rate": 4.990401896663829e-06,
+      "loss": 0.2956,
+      "step": 20,
+      "step_loss": 0.2353515625
+    },
+    {
+      "epoch": 0.20388349514563106,
+      "grad_norm": 2.087334097477202,
+      "kl": 0.169921875,
+      "learning_rate": 4.989080561915895e-06,
+      "loss": 0.3336,
+      "step": 21,
+      "step_loss": 0.2109375
+    },
+    {
+      "epoch": 0.21359223300970873,
+      "grad_norm": 1.8894907173358744,
+      "kl": 0.1845703125,
+      "learning_rate": 4.987674264578615e-06,
+      "loss": 0.2974,
+      "step": 22,
+      "step_loss": 0.2314453125
+    },
+    {
+      "epoch": 0.22330097087378642,
+      "grad_norm": 3.672752113284032,
+      "kl": 0.1748046875,
+      "learning_rate": 4.9861830580143665e-06,
+      "loss": 0.3722,
+      "step": 23,
+      "step_loss": 0.255859375
+    },
+    {
+      "epoch": 0.23300970873786409,
+      "grad_norm": 1.0485056233907877,
+      "kl": 0.1494140625,
+      "learning_rate": 4.984606998807432e-06,
+      "loss": 0.2997,
+      "step": 24,
+      "step_loss": 0.396484375
+    },
+    {
+      "epoch": 0.24271844660194175,
+      "grad_norm": 1.3694786710161069,
+      "kl": 0.11474609375,
+      "learning_rate": 4.982946146761856e-06,
+      "loss": 0.2984,
+      "step": 25,
+      "step_loss": 0.1396484375
+    },
+    {
+      "epoch": 0.2524271844660194,
+      "grad_norm": 1.0149380047661418,
+      "kl": 0.2314453125,
+      "learning_rate": 4.981200564899172e-06,
+      "loss": 0.2926,
+      "step": 26,
+      "step_loss": 0.34375
+    },
+    {
+      "epoch": 0.2621359223300971,
+      "grad_norm": 3.3346882609517627,
+      "kl": 0.1630859375,
+      "learning_rate": 4.979370319456011e-06,
+      "loss": 0.325,
+      "step": 27,
+      "step_loss": 0.2197265625
+    },
+    {
+      "epoch": 0.27184466019417475,
+      "grad_norm": 1.1026809101068904,
+      "kl": 0.1171875,
+      "learning_rate": 4.977455479881591e-06,
+      "loss": 0.3124,
+      "step": 28,
+      "step_loss": 0.42578125
+    },
+    {
+      "epoch": 0.2815533980582524,
+      "grad_norm": 1.4388631908391791,
+      "kl": 0.2470703125,
+      "learning_rate": 4.975456118835079e-06,
+      "loss": 0.2766,
+      "step": 29,
+      "step_loss": 0.2275390625
+    },
+    {
+      "epoch": 0.2912621359223301,
+      "grad_norm": 1.654229929536635,
+      "kl": 0.23828125,
+      "learning_rate": 4.973372312182835e-06,
+      "loss": 0.3192,
+      "step": 30,
+      "step_loss": 0.5546875
+    },
+    {
+      "epoch": 0.30097087378640774,
+      "grad_norm": 1.0784118934051907,
+      "kl": 0.138671875,
+      "learning_rate": 4.971204138995531e-06,
+      "loss": 0.2877,
+      "step": 31,
+      "step_loss": 0.26171875
+    },
+    {
+      "epoch": 0.3106796116504854,
+      "grad_norm": 1.789773991131577,
+      "kl": 0.2138671875,
+      "learning_rate": 4.968951681545156e-06,
+      "loss": 0.3597,
+      "step": 32,
+      "step_loss": 0.23828125
+    },
+    {
+      "epoch": 0.32038834951456313,
+      "grad_norm": 1.3349882623963718,
+      "kl": 0.1240234375,
+      "learning_rate": 4.96661502530189e-06,
+      "loss": 0.2909,
+      "step": 33,
+      "step_loss": 0.2431640625
+    },
+    {
+      "epoch": 0.3300970873786408,
+      "grad_norm": 0.8694048190676366,
+      "kl": 0.12109375,
+      "learning_rate": 4.96419425893086e-06,
+      "loss": 0.2964,
+      "step": 34,
+      "step_loss": 0.25
+    },
+    {
+      "epoch": 0.33980582524271846,
+      "grad_norm": 1.247182149510005,
+      "kl": 0.1650390625,
+      "learning_rate": 4.96168947428878e-06,
+      "loss": 0.3136,
+      "step": 35,
+      "step_loss": 0.25
+    },
+    {
+      "epoch": 0.34951456310679613,
+      "grad_norm": 0.9312530287110918,
+      "kl": 0.1630859375,
+      "learning_rate": 4.959100766420458e-06,
+      "loss": 0.3164,
+      "step": 36,
+      "step_loss": 0.306640625
+    },
+    {
+      "epoch": 0.3592233009708738,
+      "grad_norm": 0.9175318225986004,
+      "kl": 0.22265625,
+      "learning_rate": 4.9564282335552e-06,
+      "loss": 0.2954,
+      "step": 37,
+      "step_loss": 0.259765625
+    },
+    {
+      "epoch": 0.36893203883495146,
+      "grad_norm": 0.993550543291835,
+      "kl": 0.193359375,
+      "learning_rate": 4.953671977103074e-06,
+      "loss": 0.2802,
+      "step": 38,
+      "step_loss": 0.28125
+    },
+    {
+      "epoch": 0.3786407766990291,
+      "grad_norm": 1.0512485528634326,
+      "kl": 0.7578125,
+      "learning_rate": 4.950832101651063e-06,
+      "loss": 0.2843,
+      "step": 39,
+      "step_loss": 0.58203125
+    },
+    {
+      "epoch": 0.3883495145631068,
+      "grad_norm": 0.6686286343760834,
+      "kl": 0.11376953125,
+      "learning_rate": 4.947908714959102e-06,
+      "loss": 0.2512,
+      "step": 40,
+      "step_loss": 0.142578125
+    },
+    {
+      "epoch": 0.39805825242718446,
+      "grad_norm": 0.8469829082363365,
+      "kl": 0.1650390625,
+      "learning_rate": 4.944901927955983e-06,
+      "loss": 0.3058,
+      "step": 41,
+      "step_loss": 0.5546875
+    },
+    {
+      "epoch": 0.4077669902912621,
+      "grad_norm": 0.7813325456880021,
+      "kl": 0.2421875,
+      "learning_rate": 4.941811854735148e-06,
+      "loss": 0.2823,
+      "step": 42,
+      "step_loss": 0.193359375
+    },
+    {
+      "epoch": 0.4174757281553398,
+      "grad_norm": 0.9078321786346385,
+      "kl": 0.11865234375,
+      "learning_rate": 4.938638612550361e-06,
+      "loss": 0.2902,
+      "step": 43,
+      "step_loss": 0.22265625
+    },
+    {
+      "epoch": 0.42718446601941745,
+      "grad_norm": 0.8942167570500859,
+      "kl": 0.138671875,
+      "learning_rate": 4.935382321811256e-06,
+      "loss": 0.3012,
+      "step": 44,
+      "step_loss": 0.3828125
+    },
+    {
+      "epoch": 0.4368932038834951,
+      "grad_norm": 0.8140639878037927,
+      "kl": 0.166015625,
+      "learning_rate": 4.932043106078772e-06,
+      "loss": 0.2861,
+      "step": 45,
+      "step_loss": 0.197265625
+    },
+    {
+      "epoch": 0.44660194174757284,
+      "grad_norm": 1.4604431784813239,
+      "kl": 0.302734375,
+      "learning_rate": 4.928621092060457e-06,
+      "loss": 0.3029,
+      "step": 46,
+      "step_loss": 0.49609375
+    },
+    {
+      "epoch": 0.4563106796116505,
+      "grad_norm": 0.7529571567774104,
+      "kl": 0.1318359375,
+      "learning_rate": 4.925116409605672e-06,
+      "loss": 0.2834,
+      "step": 47,
+      "step_loss": 0.275390625
+    },
+    {
+      "epoch": 0.46601941747572817,
+      "grad_norm": 0.7826578425297724,
+      "kl": 0.267578125,
+      "learning_rate": 4.92152919170065e-06,
+      "loss": 0.3034,
+      "step": 48,
+      "step_loss": 0.5234375
+    },
+    {
+      "epoch": 0.47572815533980584,
+      "grad_norm": 1.066438164425578,
+      "kl": 0.2109375,
+      "learning_rate": 4.917859574463462e-06,
+      "loss": 0.3071,
+      "step": 49,
+      "step_loss": 0.2236328125
+    },
+    {
+      "epoch": 0.4854368932038835,
+      "grad_norm": 1.7914744281397004,
+      "kl": 0.1728515625,
+      "learning_rate": 4.9141076971388435e-06,
+      "loss": 0.289,
+      "step": 50,
+      "step_loss": 0.3046875
+    },
+    {
+      "epoch": 0.49514563106796117,
+      "grad_norm": 0.7820108904555855,
+      "kl": 0.126953125,
+      "learning_rate": 4.9102737020929135e-06,
+      "loss": 0.2661,
+      "step": 51,
+      "step_loss": 0.21875
+    },
+    {
+      "epoch": 0.5048543689320388,
+      "grad_norm": 0.8477559536427011,
+      "kl": 0.1875,
+      "learning_rate": 4.906357734807776e-06,
+      "loss": 0.3048,
+      "step": 52,
+      "step_loss": 0.2578125
+    },
+    {
+      "epoch": 0.5145631067961165,
+      "grad_norm": 0.781267699994807,
+      "kl": 0.150390625,
+      "learning_rate": 4.902359943875992e-06,
+      "loss": 0.2837,
+      "step": 53,
+      "step_loss": 0.2734375
+    },
+    {
+      "epoch": 0.5242718446601942,
+      "grad_norm": 0.7373706510227717,
+      "kl": 0.146484375,
+      "learning_rate": 4.89828048099495e-06,
+      "loss": 0.2632,
+      "step": 54,
+      "step_loss": 0.16015625
+    },
+    {
+      "epoch": 0.5339805825242718,
+      "grad_norm": 0.7396378743294025,
+      "kl": 0.2421875,
+      "learning_rate": 4.894119500961103e-06,
+      "loss": 0.2848,
+      "step": 55,
+      "step_loss": 0.458984375
+    },
+    {
+      "epoch": 0.5436893203883495,
+      "grad_norm": 0.8384171545494072,
+      "kl": 0.1728515625,
+      "learning_rate": 4.889877161664096e-06,
+      "loss": 0.311,
+      "step": 56,
+      "step_loss": 0.4609375
+    },
+    {
+      "epoch": 0.5533980582524272,
+      "grad_norm": 0.7847312422227414,
+      "kl": 0.205078125,
+      "learning_rate": 4.885553624080778e-06,
+      "loss": 0.2971,
+      "step": 57,
+      "step_loss": 0.203125
+    },
+    {
+      "epoch": 0.5631067961165048,
+      "grad_norm": 0.791715637164775,
+      "kl": 0.1494140625,
+      "learning_rate": 4.881149052269091e-06,
+      "loss": 0.2983,
+      "step": 58,
+      "step_loss": 0.197265625
+    },
+    {
+      "epoch": 0.5728155339805825,
+      "grad_norm": 0.6776733370951615,
+      "kl": 0.158203125,
+      "learning_rate": 4.876663613361844e-06,
+      "loss": 0.2871,
+      "step": 59,
+      "step_loss": 0.2490234375
+    },
+    {
+      "epoch": 0.5825242718446602,
+      "grad_norm": 0.6974646233573695,
+      "kl": 0.20703125,
+      "learning_rate": 4.8720974775603745e-06,
+      "loss": 0.2983,
+      "step": 60,
+      "step_loss": 0.203125
+    },
+    {
+      "epoch": 0.5922330097087378,
+      "grad_norm": 1.1133484937202898,
+      "kl": 1.1171875,
+      "learning_rate": 4.867450818128086e-06,
+      "loss": 0.2969,
+      "step": 61,
+      "step_loss": 0.89453125
+    },
+    {
+      "epoch": 0.6019417475728155,
+      "grad_norm": 0.7147826288493739,
+      "kl": 0.1279296875,
+      "learning_rate": 4.862723811383878e-06,
+      "loss": 0.2819,
+      "step": 62,
+      "step_loss": 0.341796875
+    },
+    {
+      "epoch": 0.6116504854368932,
+      "grad_norm": 0.762521447990066,
+      "kl": 0.1591796875,
+      "learning_rate": 4.857916636695449e-06,
+      "loss": 0.2812,
+      "step": 63,
+      "step_loss": 0.1533203125
+    },
+    {
+      "epoch": 0.6213592233009708,
+      "grad_norm": 0.6945241251902946,
+      "kl": 0.2294921875,
+      "learning_rate": 4.853029476472499e-06,
+      "loss": 0.2986,
+      "step": 64,
+      "step_loss": 0.275390625
+    },
+    {
+      "epoch": 0.6310679611650486,
+      "grad_norm": 0.7038147827740143,
+      "kl": 0.1708984375,
+      "learning_rate": 4.848062516159801e-06,
+      "loss": 0.3015,
+      "step": 65,
+      "step_loss": 0.400390625
+    },
+    {
+      "epoch": 0.6407766990291263,
+      "grad_norm": 0.8016050985516795,
+      "kl": 0.19140625,
+      "learning_rate": 4.843015944230166e-06,
+      "loss": 0.2873,
+      "step": 66,
+      "step_loss": 0.279296875
+    },
+    {
+      "epoch": 0.6504854368932039,
+      "grad_norm": 0.7769984336135177,
+      "kl": 0.2451171875,
+      "learning_rate": 4.837889952177294e-06,
+      "loss": 0.2899,
+      "step": 67,
+      "step_loss": 0.56640625
+    },
+    {
+      "epoch": 0.6601941747572816,
+      "grad_norm": 0.6998891333931025,
+      "kl": 0.138671875,
+      "learning_rate": 4.832684734508502e-06,
+      "loss": 0.2643,
+      "step": 68,
+      "step_loss": 0.330078125
+    },
+    {
+      "epoch": 0.6699029126213593,
+      "grad_norm": 1.0392663511926639,
+      "kl": 0.1640625,
+      "learning_rate": 4.827400488737351e-06,
+      "loss": 0.2863,
+      "step": 69,
+      "step_loss": 0.287109375
+    },
+    {
+      "epoch": 0.6796116504854369,
+      "grad_norm": 0.8630658049176575,
+      "kl": 0.40234375,
+      "learning_rate": 4.822037415376147e-06,
+      "loss": 0.2894,
+      "step": 70,
+      "step_loss": 0.435546875
+    },
+    {
+      "epoch": 0.6893203883495146,
+      "grad_norm": 0.7810137963853107,
+      "kl": 0.271484375,
+      "learning_rate": 4.816595717928327e-06,
+      "loss": 0.2842,
+      "step": 71,
+      "step_loss": 0.2451171875
+    },
+    {
+      "epoch": 0.6990291262135923,
+      "grad_norm": 0.7611975818745967,
+      "kl": 0.181640625,
+      "learning_rate": 4.8110756028807506e-06,
+      "loss": 0.2618,
+      "step": 72,
+      "step_loss": 0.3359375
+    },
+    {
+      "epoch": 0.7087378640776699,
+      "grad_norm": 0.9039894718510777,
+      "kl": 0.17578125,
+      "learning_rate": 4.805477279695852e-06,
+      "loss": 0.304,
+      "step": 73,
+      "step_loss": 0.369140625
+    },
+    {
+      "epoch": 0.7184466019417476,
+      "grad_norm": 0.6848452637621099,
+      "kl": 0.154296875,
+      "learning_rate": 4.799800960803699e-06,
+      "loss": 0.2583,
+      "step": 74,
+      "step_loss": 0.265625
+    },
+    {
+      "epoch": 0.7281553398058253,
+      "grad_norm": 0.7706941057886834,
+      "kl": 0.11962890625,
+      "learning_rate": 4.794046861593929e-06,
+      "loss": 0.2913,
+      "step": 75,
+      "step_loss": 0.19921875
+    },
+    {
+      "epoch": 0.7378640776699029,
+      "grad_norm": 0.7340458477237756,
+      "kl": 0.1953125,
+      "learning_rate": 4.788215200407576e-06,
+      "loss": 0.2725,
+      "step": 76,
+      "step_loss": 0.357421875
+    },
+    {
+      "epoch": 0.7475728155339806,
+      "grad_norm": 1.0806357960026565,
+      "kl": 0.23046875,
+      "learning_rate": 4.7823061985287906e-06,
+      "loss": 0.2781,
+      "step": 77,
+      "step_loss": 0.193359375
+    },
+    {
+      "epoch": 0.7572815533980582,
+      "grad_norm": 0.7704916059097427,
+      "kl": 0.34375,
+      "learning_rate": 4.776320080176434e-06,
+      "loss": 0.2742,
+      "step": 78,
+      "step_loss": 0.3203125
+    },
+    {
+      "epoch": 0.7669902912621359,
+      "grad_norm": 0.7492040543213823,
+      "kl": 0.18359375,
+      "learning_rate": 4.770257072495581e-06,
+      "loss": 0.2776,
+      "step": 79,
+      "step_loss": 0.2109375
+    },
+    {
+      "epoch": 0.7766990291262136,
+      "grad_norm": 0.7467369782034328,
+      "kl": 0.51171875,
+      "learning_rate": 4.764117405548891e-06,
+      "loss": 0.2377,
+      "step": 80,
+      "step_loss": 0.365234375
+    },
+    {
+      "epoch": 0.7864077669902912,
+      "grad_norm": 0.9406968288941178,
+      "kl": 0.1513671875,
+      "learning_rate": 4.757901312307882e-06,
+      "loss": 0.2708,
+      "step": 81,
+      "step_loss": 0.162109375
+    },
+    {
+      "epoch": 0.7961165048543689,
+      "grad_norm": 0.7660157121486119,
+      "kl": 0.2333984375,
+      "learning_rate": 4.751609028644097e-06,
+      "loss": 0.2877,
+      "step": 82,
+      "step_loss": 0.30078125
+    },
+    {
+      "epoch": 0.8058252427184466,
+      "grad_norm": 0.6957183682004972,
+      "kl": 0.1806640625,
+      "learning_rate": 4.7452407933201395e-06,
+      "loss": 0.2403,
+      "step": 83,
+      "step_loss": 0.265625
+    },
+    {
+      "epoch": 0.8155339805825242,
+      "grad_norm": 0.7043740093288546,
+      "kl": 0.1767578125,
+      "learning_rate": 4.738796847980627e-06,
+      "loss": 0.2735,
+      "step": 84,
+      "step_loss": 0.3359375
+    },
+    {
+      "epoch": 0.8252427184466019,
+      "grad_norm": 0.6782132306690141,
+      "kl": 0.1572265625,
+      "learning_rate": 4.732277437143015e-06,
+      "loss": 0.2517,
+      "step": 85,
+      "step_loss": 0.2119140625
+    },
+    {
+      "epoch": 0.8349514563106796,
+      "grad_norm": 0.7479330827900228,
+      "kl": 0.21484375,
+      "learning_rate": 4.725682808188321e-06,
+      "loss": 0.2641,
+      "step": 86,
+      "step_loss": 0.232421875
+    },
+    {
+      "epoch": 0.8446601941747572,
+      "grad_norm": 0.7620154952078262,
+      "kl": 0.10791015625,
+      "learning_rate": 4.719013211351734e-06,
+      "loss": 0.2572,
+      "step": 87,
+      "step_loss": 0.146484375
+    },
+    {
+      "epoch": 0.8543689320388349,
+      "grad_norm": 0.6678822109983321,
+      "kl": 0.2060546875,
+      "learning_rate": 4.712268899713125e-06,
+      "loss": 0.2878,
+      "step": 88,
+      "step_loss": 0.275390625
+    },
+    {
+      "epoch": 0.8640776699029126,
+      "grad_norm": 0.7456708178708048,
+      "kl": 0.1298828125,
+      "learning_rate": 4.705450129187439e-06,
+      "loss": 0.2746,
+      "step": 89,
+      "step_loss": 0.2275390625
+    },
+    {
+      "epoch": 0.8737864077669902,
+      "grad_norm": 0.6177231600838415,
+      "kl": 0.2236328125,
+      "learning_rate": 4.698557158514988e-06,
+      "loss": 0.2491,
+      "step": 90,
+      "step_loss": 0.412109375
+    },
+    {
+      "epoch": 0.883495145631068,
+      "grad_norm": 0.8797248927244058,
+      "kl": 0.1513671875,
+      "learning_rate": 4.691590249251627e-06,
+      "loss": 0.2745,
+      "step": 91,
+      "step_loss": 0.12353515625
+    },
+    {
+      "epoch": 0.8932038834951457,
+      "grad_norm": 0.7015383302055267,
+      "kl": 0.1162109375,
+      "learning_rate": 4.684549665758839e-06,
+      "loss": 0.2778,
+      "step": 92,
+      "step_loss": 0.189453125
+    },
+    {
+      "epoch": 0.9029126213592233,
+      "grad_norm": 0.7073373711435095,
+      "kl": 0.2490234375,
+      "learning_rate": 4.677435675193692e-06,
+      "loss": 0.2862,
+      "step": 93,
+      "step_loss": 0.26171875
+    },
+    {
+      "epoch": 0.912621359223301,
+      "grad_norm": 0.6637944746473083,
+      "kl": 0.14453125,
+      "learning_rate": 4.670248547498712e-06,
+      "loss": 0.2734,
+      "step": 94,
+      "step_loss": 0.1376953125
+    },
+    {
+      "epoch": 0.9223300970873787,
+      "grad_norm": 0.720119043217982,
+      "kl": 0.1865234375,
+      "learning_rate": 4.662988555391632e-06,
+      "loss": 0.2899,
+      "step": 95,
+      "step_loss": 0.423828125
+    },
+    {
+      "epoch": 0.9320388349514563,
+      "grad_norm": 0.6938324080314517,
+      "kl": 0.16796875,
+      "learning_rate": 4.655655974355051e-06,
+      "loss": 0.2714,
+      "step": 96,
+      "step_loss": 0.255859375
+    },
+    {
+      "epoch": 0.941747572815534,
+      "grad_norm": 0.6610518480589428,
+      "kl": 0.1630859375,
+      "learning_rate": 4.648251082625975e-06,
+      "loss": 0.2837,
+      "step": 97,
+      "step_loss": 0.29296875
+    },
+    {
+      "epoch": 0.9514563106796117,
+      "grad_norm": 0.6877843962165877,
+      "kl": 0.1748046875,
+      "learning_rate": 4.640774161185259e-06,
+      "loss": 0.3036,
+      "step": 98,
+      "step_loss": 0.515625
+    },
+    {
+      "epoch": 0.9611650485436893,
+      "grad_norm": 0.6332816751225963,
+      "kl": 0.1533203125,
+      "learning_rate": 4.633225493746951e-06,
+      "loss": 0.2489,
+      "step": 99,
+      "step_loss": 0.1826171875
+    },
+    {
+      "epoch": 0.970873786407767,
+      "grad_norm": 0.7362793924005384,
+      "kl": 0.123046875,
+      "learning_rate": 4.625605366747519e-06,
+      "loss": 0.2622,
+      "step": 100,
+      "step_loss": 0.12890625
+    },
+    {
+      "epoch": 0.9805825242718447,
+      "grad_norm": 0.6742913243561754,
+      "kl": 0.2119140625,
+      "learning_rate": 4.617914069334989e-06,
+      "loss": 0.2544,
+      "step": 101,
+      "step_loss": 0.2578125
+    },
+    {
+      "epoch": 0.9902912621359223,
+      "grad_norm": 0.6047332480224725,
+      "kl": 0.259765625,
+      "learning_rate": 4.610151893357968e-06,
+      "loss": 0.2546,
+      "step": 102,
+      "step_loss": 0.37890625
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.6546589091756503,
+      "kl": 0.2138671875,
+      "learning_rate": 4.602319133354571e-06,
+      "loss": 0.2566,
+      "step": 103,
+      "step_loss": 0.2177734375
+    },
+    {
+      "epoch": 1.0,
+      "eval_test_transformed.json_loss": NaN,
+      "eval_test_transformed.json_runtime": 8.816,
+      "eval_test_transformed.json_samples_per_second": 56.715,
+      "eval_test_transformed.json_steps_per_second": 1.815,
+      "step": 103
+    },
+    {
+      "epoch": 1.0097087378640777,
+      "grad_norm": 0.639387377398209,
+      "kl": NaN,
+      "learning_rate": 4.594416086541248e-06,
+      "loss": 0.2493,
+      "step": 104,
+      "step_loss": NaN
+    },
+    {
+      "epoch": 1.0194174757281553,
+      "grad_norm": 0.7392491046589192,
+      "kl": 0.1640625,
+      "learning_rate": 4.5864430528014996e-06,
+      "loss": 0.2413,
+      "step": 105,
+      "step_loss": 0.1396484375
+    },
+    {
+      "epoch": 1.029126213592233,
+      "grad_norm": 0.5649339416591039,
+      "kl": 0.2001953125,
+      "learning_rate": 4.578400334674503e-06,
+      "loss": 0.2223,
+      "step": 106,
+      "step_loss": 0.318359375
+    },
+    {
+      "epoch": 1.0388349514563107,
+      "grad_norm": 0.6022576613202164,
+      "kl": 0.1708984375,
+      "learning_rate": 4.570288237343632e-06,
+      "loss": 0.2368,
+      "step": 107,
+      "step_loss": 0.24609375
+    },
+    {
+      "epoch": 1.0485436893203883,
+      "grad_norm": 0.6973100963269709,
+      "kl": 0.474609375,
+      "learning_rate": 4.562107068624874e-06,
+      "loss": 0.2363,
+      "step": 108,
+      "step_loss": 0.255859375
+    },
+    {
+      "epoch": 1.058252427184466,
+      "grad_norm": 0.6281000265214681,
+      "kl": 0.16796875,
+      "learning_rate": 4.55385713895515e-06,
+      "loss": 0.2271,
+      "step": 109,
+      "step_loss": 0.2578125
+    },
+    {
+      "epoch": 1.0679611650485437,
+      "grad_norm": 0.6251998581820843,
+      "kl": 0.19921875,
+      "learning_rate": 4.545538761380539e-06,
+      "loss": 0.2444,
+      "step": 110,
+      "step_loss": 0.1533203125
+    },
+    {
+      "epoch": 1.0776699029126213,
+      "grad_norm": 0.6499920789529428,
+      "kl": 0.140625,
+      "learning_rate": 4.537152251544394e-06,
+      "loss": 0.2575,
+      "step": 111,
+      "step_loss": 0.21484375
+    },
+    {
+      "epoch": 1.087378640776699,
+      "grad_norm": 0.6196063483230071,
+      "kl": 0.34765625,
+      "learning_rate": 4.5286979276753675e-06,
+      "loss": 0.2216,
+      "step": 112,
+      "step_loss": 0.24609375
+    },
+    {
+      "epoch": 1.0970873786407767,
+      "grad_norm": 0.6519876067572186,
+      "kl": 0.2265625,
+      "learning_rate": 4.520176110575338e-06,
+      "loss": 0.2209,
+      "step": 113,
+      "step_loss": 0.1748046875
+    },
+    {
+      "epoch": 1.1067961165048543,
+      "grad_norm": 0.6464673181225589,
+      "kl": 0.2041015625,
+      "learning_rate": 4.511587123607232e-06,
+      "loss": 0.2242,
+      "step": 114,
+      "step_loss": 0.271484375
+    },
+    {
+      "epoch": 1.116504854368932,
+      "grad_norm": 0.5620248695246799,
+      "kl": 0.23046875,
+      "learning_rate": 4.502931292682759e-06,
+      "loss": 0.225,
+      "step": 115,
+      "step_loss": 0.255859375
+    },
+    {
+      "epoch": 1.1262135922330097,
+      "grad_norm": 0.6212910752562149,
+      "kl": 0.3203125,
+      "learning_rate": 4.494208946250042e-06,
+      "loss": 0.2322,
+      "step": 116,
+      "step_loss": 0.166015625
+    },
+    {
+      "epoch": 1.1359223300970873,
+      "grad_norm": 0.6102956517360447,
+      "kl": 0.134765625,
+      "learning_rate": 4.485420415281157e-06,
+      "loss": 0.246,
+      "step": 117,
+      "step_loss": 0.2099609375
+    },
+    {
+      "epoch": 1.145631067961165,
+      "grad_norm": 0.8959975572620882,
+      "kl": 0.341796875,
+      "learning_rate": 4.4765660332595686e-06,
+      "loss": 0.2519,
+      "step": 118,
+      "step_loss": 0.154296875
+    },
+    {
+      "epoch": 1.1553398058252426,
+      "grad_norm": 0.5883698395469736,
+      "kl": 0.30078125,
+      "learning_rate": 4.467646136167482e-06,
+      "loss": 0.2548,
+      "step": 119,
+      "step_loss": 0.22265625
+    },
+    {
+      "epoch": 1.1650485436893203,
+      "grad_norm": 0.5899290313710992,
+      "kl": 0.1669921875,
+      "learning_rate": 4.458661062473091e-06,
+      "loss": 0.2327,
+      "step": 120,
+      "step_loss": 0.1533203125
+    },
+    {
+      "epoch": 1.174757281553398,
+      "grad_norm": 0.7434899396388887,
+      "kl": 0.294921875,
+      "learning_rate": 4.449611153117736e-06,
+      "loss": 0.2415,
+      "step": 121,
+      "step_loss": 0.16796875
+    },
+    {
+      "epoch": 1.1844660194174756,
+      "grad_norm": 0.5645672375917218,
+      "kl": 0.21875,
+      "learning_rate": 4.4404967515029655e-06,
+      "loss": 0.198,
+      "step": 122,
+      "step_loss": 0.27734375
+    },
+    {
+      "epoch": 1.1941747572815533,
+      "grad_norm": 0.61620433580154,
+      "kl": 0.28125,
+      "learning_rate": 4.431318203477505e-06,
+      "loss": 0.214,
+      "step": 123,
+      "step_loss": 0.18359375
+    },
+    {
+      "epoch": 1.203883495145631,
+      "grad_norm": 0.5927124243417978,
+      "kl": 0.29296875,
+      "learning_rate": 4.422075857324138e-06,
+      "loss": 0.2182,
+      "step": 124,
+      "step_loss": 0.25
+    },
+    {
+      "epoch": 1.2135922330097086,
+      "grad_norm": 0.6201091855952103,
+      "kl": 0.18359375,
+      "learning_rate": 4.412770063746483e-06,
+      "loss": 0.2207,
+      "step": 125,
+      "step_loss": 0.1865234375
+    },
+    {
+      "epoch": 1.2233009708737863,
+      "grad_norm": 0.911604159360823,
+      "kl": 0.162109375,
+      "learning_rate": 4.403401175855695e-06,
+      "loss": 0.2471,
+      "step": 126,
+      "step_loss": 0.1474609375
+    },
+    {
+      "epoch": 1.233009708737864,
+      "grad_norm": 0.6792470228209272,
+      "kl": 0.0927734375,
+      "learning_rate": 4.3939695491570596e-06,
+      "loss": 0.2351,
+      "step": 127,
+      "step_loss": 0.146484375
+    },
+    {
+      "epoch": 1.2427184466019416,
+      "grad_norm": 0.6012469615978469,
+      "kl": 0.212890625,
+      "learning_rate": 4.384475541536505e-06,
+      "loss": 0.2434,
+      "step": 128,
+      "step_loss": 0.2041015625
+    },
+    {
+      "epoch": 1.2524271844660193,
+      "grad_norm": 0.589209563012478,
+      "kl": 0.2041015625,
+      "learning_rate": 4.374919513247021e-06,
+      "loss": 0.2129,
+      "step": 129,
+      "step_loss": 0.109375
+    },
+    {
+      "epoch": 1.262135922330097,
+      "grad_norm": 0.5987454630415056,
+      "kl": 0.22265625,
+      "learning_rate": 4.3653018268949945e-06,
+      "loss": 0.2179,
+      "step": 130,
+      "step_loss": 0.166015625
+    },
+    {
+      "epoch": 1.2718446601941746,
+      "grad_norm": 0.5813161566217064,
+      "kl": 0.2041015625,
+      "learning_rate": 4.355622847426444e-06,
+      "loss": 0.2183,
+      "step": 131,
+      "step_loss": 0.431640625
+    },
+    {
+      "epoch": 1.2815533980582523,
+      "grad_norm": 0.644992270460315,
+      "kl": 0.25,
+      "learning_rate": 4.345882942113171e-06,
+      "loss": 0.2412,
+      "step": 132,
+      "step_loss": 0.1474609375
+    },
+    {
+      "epoch": 1.29126213592233,
+      "grad_norm": 0.5838255650379847,
+      "kl": 0.4140625,
+      "learning_rate": 4.336082480538832e-06,
+      "loss": 0.2223,
+      "step": 133,
+      "step_loss": 0.171875
+    },
+    {
+      "epoch": 1.3009708737864076,
+      "grad_norm": 0.7975539886352869,
+      "kl": 0.3125,
+      "learning_rate": 4.326221834584905e-06,
+      "loss": 0.2218,
+      "step": 134,
+      "step_loss": 0.47265625
+    },
+    {
+      "epoch": 1.3106796116504853,
+      "grad_norm": 0.5800964403953691,
+      "kl": 0.181640625,
+      "learning_rate": 4.316301378416585e-06,
+      "loss": 0.239,
+      "step": 135,
+      "step_loss": 0.1552734375
+    },
+    {
+      "epoch": 1.3203883495145632,
+      "grad_norm": 0.5247261922814125,
+      "kl": 0.1923828125,
+      "learning_rate": 4.306321488468583e-06,
+      "loss": 0.2337,
+      "step": 136,
+      "step_loss": 0.119140625
+    },
+    {
+      "epoch": 1.3300970873786409,
+      "grad_norm": 0.6297520228191021,
+      "kl": 0.1435546875,
+      "learning_rate": 4.296282543430841e-06,
+      "loss": 0.24,
+      "step": 137,
+      "step_loss": 0.29296875
+    },
+    {
+      "epoch": 1.3398058252427185,
+      "grad_norm": 0.5734205082343037,
+      "kl": 0.216796875,
+      "learning_rate": 4.286184924234168e-06,
+      "loss": 0.2218,
+      "step": 138,
+      "step_loss": 0.15234375
+    },
+    {
+      "epoch": 1.3495145631067962,
+      "grad_norm": 0.5442629148889834,
+      "kl": 0.2255859375,
+      "learning_rate": 4.276029014035777e-06,
+      "loss": 0.2189,
+      "step": 139,
+      "step_loss": 0.2265625
+    },
+    {
+      "epoch": 1.3592233009708738,
+      "grad_norm": 0.5288732036595232,
+      "kl": 0.12060546875,
+      "learning_rate": 4.265815198204754e-06,
+      "loss": 0.2428,
+      "step": 140,
+      "step_loss": 0.169921875
+    },
+    {
+      "epoch": 1.3689320388349515,
+      "grad_norm": 0.6546290678837918,
+      "kl": 0.2451171875,
+      "learning_rate": 4.2555438643074315e-06,
+      "loss": 0.2422,
+      "step": 141,
+      "step_loss": 0.21875
+    },
+    {
+      "epoch": 1.3786407766990292,
+      "grad_norm": 0.585078334332048,
+      "kl": 0.2041015625,
+      "learning_rate": 4.245215402092681e-06,
+      "loss": 0.2455,
+      "step": 142,
+      "step_loss": 0.205078125
+    },
+    {
+      "epoch": 1.3883495145631068,
+      "grad_norm": 0.556943790797957,
+      "kl": 0.1748046875,
+      "learning_rate": 4.234830203477126e-06,
+      "loss": 0.2064,
+      "step": 143,
+      "step_loss": 0.201171875
+    },
+    {
+      "epoch": 1.3980582524271845,
+      "grad_norm": 0.5946911580488923,
+      "kl": 0.14453125,
+      "learning_rate": 4.224388662530271e-06,
+      "loss": 0.2218,
+      "step": 144,
+      "step_loss": 0.2431640625
+    },
+    {
+      "epoch": 1.4077669902912622,
+      "grad_norm": 0.5810717668347283,
+      "kl": 0.326171875,
+      "learning_rate": 4.213891175459545e-06,
+      "loss": 0.2442,
+      "step": 145,
+      "step_loss": 0.1796875
+    },
+    {
+      "epoch": 1.4174757281553398,
+      "grad_norm": 0.6115382263091327,
+      "kl": 0.1865234375,
+      "learning_rate": 4.203338140595272e-06,
+      "loss": 0.2202,
+      "step": 146,
+      "step_loss": 0.26171875
+    },
+    {
+      "epoch": 1.4271844660194175,
+      "grad_norm": 0.5465729015270102,
+      "kl": 0.248046875,
+      "learning_rate": 4.192729958375552e-06,
+      "loss": 0.2129,
+      "step": 147,
+      "step_loss": 0.28125
+    },
+    {
+      "epoch": 1.4368932038834952,
+      "grad_norm": 0.5834391929303593,
+      "kl": 0.162109375,
+      "learning_rate": 4.1820670313310684e-06,
+      "loss": 0.2328,
+      "step": 148,
+      "step_loss": 0.1103515625
+    },
+    {
+      "epoch": 1.4466019417475728,
+      "grad_norm": 0.6239303256729286,
+      "kl": 0.19921875,
+      "learning_rate": 4.171349764069815e-06,
+      "loss": 0.2468,
+      "step": 149,
+      "step_loss": 0.2275390625
+    },
+    {
+      "epoch": 1.4563106796116505,
+      "grad_norm": 0.5464914339167269,
+      "kl": 0.154296875,
+      "learning_rate": 4.16057856326174e-06,
+      "loss": 0.2145,
+      "step": 150,
+      "step_loss": 0.21484375
+    },
+    {
+      "epoch": 1.4660194174757282,
+      "grad_norm": 0.5714192820756144,
+      "kl": 0.34765625,
+      "learning_rate": 4.149753837623317e-06,
+      "loss": 0.2293,
+      "step": 151,
+      "step_loss": 0.1142578125
+    },
+    {
+      "epoch": 1.4757281553398058,
+      "grad_norm": 0.6222270965621777,
+      "kl": 0.20703125,
+      "learning_rate": 4.1388759979020346e-06,
+      "loss": 0.2378,
+      "step": 152,
+      "step_loss": 0.2158203125
+    },
+    {
+      "epoch": 1.4854368932038835,
+      "grad_norm": 0.6459648372732245,
+      "kl": 0.1728515625,
+      "learning_rate": 4.127945456860813e-06,
+      "loss": 0.2301,
+      "step": 153,
+      "step_loss": 0.154296875
+    },
+    {
+      "epoch": 1.4951456310679612,
+      "grad_norm": 0.613782544176392,
+      "kl": 0.21875,
+      "learning_rate": 4.116962629262339e-06,
+      "loss": 0.2155,
+      "step": 154,
+      "step_loss": 0.2099609375
+    },
+    {
+      "epoch": 1.5048543689320388,
+      "grad_norm": 0.569655314094469,
+      "kl": 0.220703125,
+      "learning_rate": 4.105927931853327e-06,
+      "loss": 0.231,
+      "step": 155,
+      "step_loss": 0.26953125
+    },
+    {
+      "epoch": 1.5145631067961165,
+      "grad_norm": 0.6309412616695066,
+      "kl": 0.369140625,
+      "learning_rate": 4.094841783348711e-06,
+      "loss": 0.2435,
+      "step": 156,
+      "step_loss": 0.1689453125
+    },
+    {
+      "epoch": 1.5242718446601942,
+      "grad_norm": 0.6105234133068124,
+      "kl": 0.2138671875,
+      "learning_rate": 4.083704604415749e-06,
+      "loss": 0.2052,
+      "step": 157,
+      "step_loss": 0.2412109375
+    },
+    {
+      "epoch": 1.5339805825242718,
+      "grad_norm": 0.5804176749635781,
+      "kl": 0.2197265625,
+      "learning_rate": 4.072516817658065e-06,
+      "loss": 0.2508,
+      "step": 158,
+      "step_loss": 0.171875
+    },
+    {
+      "epoch": 1.5436893203883495,
+      "grad_norm": 0.55765374326049,
+      "kl": 0.306640625,
+      "learning_rate": 4.0612788475996125e-06,
+      "loss": 0.2164,
+      "step": 159,
+      "step_loss": 0.2001953125
+    },
+    {
+      "epoch": 1.5533980582524272,
+      "grad_norm": 0.6074565129354751,
+      "kl": 0.7265625,
+      "learning_rate": 4.049991120668566e-06,
+      "loss": 0.229,
+      "step": 160,
+      "step_loss": 0.40625
+    },
+    {
+      "epoch": 1.5631067961165048,
+      "grad_norm": 0.7748587294998153,
+      "kl": 1.328125,
+      "learning_rate": 4.038654065181137e-06,
+      "loss": 0.281,
+      "step": 161,
+      "step_loss": 0.8984375
+    },
+    {
+      "epoch": 1.5728155339805825,
+      "grad_norm": 0.5644659021112801,
+      "kl": 0.205078125,
+      "learning_rate": 4.027268111325328e-06,
+      "loss": 0.2397,
+      "step": 162,
+      "step_loss": 0.26953125
+    },
+    {
+      "epoch": 1.5825242718446602,
+      "grad_norm": 0.5648454404940179,
+      "kl": 0.1533203125,
+      "learning_rate": 4.015833691144603e-06,
+      "loss": 0.2157,
+      "step": 163,
+      "step_loss": 0.1416015625
+    },
+    {
+      "epoch": 1.5922330097087378,
+      "grad_norm": 0.5903127460469126,
+      "kl": 0.224609375,
+      "learning_rate": 4.0043512385214924e-06,
+      "loss": 0.2308,
+      "step": 164,
+      "step_loss": 0.1669921875
+    },
+    {
+      "epoch": 1.6019417475728155,
+      "grad_norm": 0.5951640957860757,
+      "kl": 0.26953125,
+      "learning_rate": 3.9928211891611385e-06,
+      "loss": 0.2136,
+      "step": 165,
+      "step_loss": 0.1875
+    },
+    {
+      "epoch": 1.6116504854368932,
+      "grad_norm": 0.5261435237006341,
+      "kl": 0.17578125,
+      "learning_rate": 3.981243980574751e-06,
+      "loss": 0.2382,
+      "step": 166,
+      "step_loss": 0.234375
+    },
+    {
+      "epoch": 1.6213592233009708,
+      "grad_norm": 0.5887020505829587,
+      "kl": 0.2255859375,
+      "learning_rate": 3.969620052063012e-06,
+      "loss": 0.2352,
+      "step": 167,
+      "step_loss": 0.2236328125
+    },
+    {
+      "epoch": 1.6310679611650487,
+      "grad_norm": 0.6254868280303059,
+      "kl": 0.2001953125,
+      "learning_rate": 3.957949844699405e-06,
+      "loss": 0.2235,
+      "step": 168,
+      "step_loss": 0.162109375
+    },
+    {
+      "epoch": 1.6407766990291264,
+      "grad_norm": 0.5478297482531744,
+      "kl": 0.23046875,
+      "learning_rate": 3.946233801313482e-06,
+      "loss": 0.2201,
+      "step": 169,
+      "step_loss": 0.30859375
+    },
+    {
+      "epoch": 1.650485436893204,
+      "grad_norm": 0.5691412826879887,
+      "kl": 0.2177734375,
+      "learning_rate": 3.934472366474051e-06,
+      "loss": 0.2014,
+      "step": 170,
+      "step_loss": 0.2001953125
+    },
+    {
+      "epoch": 1.6601941747572817,
+      "grad_norm": 0.5604014457233358,
+      "kl": 0.16796875,
+      "learning_rate": 3.922665986472316e-06,
+      "loss": 0.2348,
+      "step": 171,
+      "step_loss": 0.27734375
+    },
+    {
+      "epoch": 1.6699029126213594,
+      "grad_norm": 0.5166321657054415,
+      "kl": 0.1708984375,
+      "learning_rate": 3.91081510930494e-06,
+      "loss": 0.2125,
+      "step": 172,
+      "step_loss": 0.263671875
+    },
+    {
+      "epoch": 1.679611650485437,
+      "grad_norm": 0.567633227241871,
+      "kl": 0.2197265625,
+      "learning_rate": 3.8989201846570405e-06,
+      "loss": 0.2257,
+      "step": 173,
+      "step_loss": 0.16796875
+    },
+    {
+      "epoch": 1.6893203883495147,
+      "grad_norm": 0.6483568118346154,
+      "kl": 0.185546875,
+      "learning_rate": 3.886981663885134e-06,
+      "loss": 0.2426,
+      "step": 174,
+      "step_loss": 0.25390625
+    },
+    {
+      "epoch": 1.6990291262135924,
+      "grad_norm": 0.6784974565198278,
+      "kl": 0.28125,
+      "learning_rate": 3.875e-06,
+      "loss": 0.2508,
+      "step": 175,
+      "step_loss": 0.453125
+    },
+    {
+      "epoch": 1.70873786407767,
+      "grad_norm": 0.5532390595417498,
+      "kl": 0.20703125,
+      "learning_rate": 3.862975647649503e-06,
+      "loss": 0.2183,
+      "step": 176,
+      "step_loss": 0.12890625
+    },
+    {
+      "epoch": 1.7184466019417477,
+      "grad_norm": 0.5636800996705107,
+      "kl": 0.228515625,
+      "learning_rate": 3.850909063101328e-06,
+      "loss": 0.2107,
+      "step": 177,
+      "step_loss": 0.1845703125
+    },
+    {
+      "epoch": 1.7281553398058254,
+      "grad_norm": 0.5703475366568808,
+      "kl": 0.1337890625,
+      "learning_rate": 3.838800704225679e-06,
+      "loss": 0.2135,
+      "step": 178,
+      "step_loss": 0.1708984375
+    },
+    {
+      "epoch": 1.737864077669903,
+      "grad_norm": 0.6101234925448211,
+      "kl": 0.1904296875,
+      "learning_rate": 3.826651030477896e-06,
+      "loss": 0.2385,
+      "step": 179,
+      "step_loss": 0.177734375
+    },
+    {
+      "epoch": 1.7475728155339807,
+      "grad_norm": 0.5393184109703284,
+      "kl": 0.1708984375,
+      "learning_rate": 3.8144605028810233e-06,
+      "loss": 0.2215,
+      "step": 180,
+      "step_loss": 0.18359375
+    },
+    {
+      "epoch": 1.7572815533980584,
+      "grad_norm": 0.560735690651583,
+      "kl": 0.322265625,
+      "learning_rate": 3.802229584008321e-06,
+      "loss": 0.2235,
+      "step": 181,
+      "step_loss": 0.2353515625
+    },
+    {
+      "epoch": 1.766990291262136,
+      "grad_norm": 0.5687197299908796,
+      "kl": 0.318359375,
+      "learning_rate": 3.789958737965705e-06,
+      "loss": 0.2377,
+      "step": 182,
+      "step_loss": 0.42578125
+    },
+    {
+      "epoch": 1.7766990291262137,
+      "grad_norm": 0.517105808317797,
+      "kl": 0.2265625,
+      "learning_rate": 3.777648430374142e-06,
+      "loss": 0.2247,
+      "step": 183,
+      "step_loss": 0.423828125
+    },
+    {
+      "epoch": 1.7864077669902914,
+      "grad_norm": 0.5846543567424347,
+      "kl": 0.150390625,
+      "learning_rate": 3.765299128351977e-06,
+      "loss": 0.2233,
+      "step": 184,
+      "step_loss": 0.14453125
+    },
+    {
+      "epoch": 1.796116504854369,
+      "grad_norm": 0.550349343500017,
+      "kl": 0.5078125,
+      "learning_rate": 3.7529113004972117e-06,
+      "loss": 0.2296,
+      "step": 185,
+      "step_loss": 0.32421875
+    },
+    {
+      "epoch": 1.8058252427184467,
+      "grad_norm": 0.6054819413255546,
+      "kl": 0.2890625,
+      "learning_rate": 3.740485416869722e-06,
+      "loss": 0.2266,
+      "step": 186,
+      "step_loss": 0.2373046875
+    },
+    {
+      "epoch": 1.8155339805825244,
+      "grad_norm": 0.5882865611694021,
+      "kl": 0.2001953125,
+      "learning_rate": 3.7280219489734214e-06,
+      "loss": 0.2265,
+      "step": 187,
+      "step_loss": 0.1259765625
+    },
+    {
+      "epoch": 1.825242718446602,
+      "grad_norm": 0.6392508865666573,
+      "kl": 0.328125,
+      "learning_rate": 3.7155213697383702e-06,
+      "loss": 0.2463,
+      "step": 188,
+      "step_loss": 0.173828125
+    },
+    {
+      "epoch": 1.8349514563106797,
+      "grad_norm": 0.5196239704198046,
+      "kl": 0.1923828125,
+      "learning_rate": 3.70298415350283e-06,
+      "loss": 0.2359,
+      "step": 189,
+      "step_loss": 0.123046875
+    },
+    {
+      "epoch": 1.8446601941747574,
+      "grad_norm": 0.6494083949250458,
+      "kl": 1.0859375,
+      "learning_rate": 3.690410775995263e-06,
+      "loss": 0.2307,
+      "step": 190,
+      "step_loss": 0.44921875
+    },
+    {
+      "epoch": 1.854368932038835,
+      "grad_norm": 0.5435542784787667,
+      "kl": 0.3359375,
+      "learning_rate": 3.677801714316283e-06,
+      "loss": 0.2252,
+      "step": 191,
+      "step_loss": 0.2265625
+    },
+    {
+      "epoch": 1.8640776699029127,
+      "grad_norm": 0.6511666190719619,
+      "kl": 0.248046875,
+      "learning_rate": 3.665157446920551e-06,
+      "loss": 0.2525,
+      "step": 192,
+      "step_loss": 0.189453125
+    },
+    {
+      "epoch": 1.8737864077669903,
+      "grad_norm": 0.6128430106715034,
+      "kl": 0.306640625,
+      "learning_rate": 3.6524784535986175e-06,
+      "loss": 0.2723,
+      "step": 193,
+      "step_loss": 0.298828125
+    },
+    {
+      "epoch": 1.883495145631068,
+      "grad_norm": 0.5849812397434826,
+      "kl": 0.1875,
+      "learning_rate": 3.639765215458721e-06,
+      "loss": 0.2487,
+      "step": 194,
+      "step_loss": 0.33984375
+    },
+    {
+      "epoch": 1.8932038834951457,
+      "grad_norm": 0.5998143147199343,
+      "kl": 0.2060546875,
+      "learning_rate": 3.6270182149085288e-06,
+      "loss": 0.2265,
+      "step": 195,
+      "step_loss": 0.1162109375
+    },
+    {
+      "epoch": 1.9029126213592233,
+      "grad_norm": 0.5559548956050891,
+      "kl": 0.2041015625,
+      "learning_rate": 3.6142379356368334e-06,
+      "loss": 0.2266,
+      "step": 196,
+      "step_loss": 0.12109375
+    },
+    {
+      "epoch": 1.912621359223301,
+      "grad_norm": 0.6722163508708774,
+      "kl": 0.2265625,
+      "learning_rate": 3.6014248625951987e-06,
+      "loss": 0.2349,
+      "step": 197,
+      "step_loss": 0.296875
+    },
+    {
+      "epoch": 1.9223300970873787,
+      "grad_norm": 0.5488054807070938,
+      "kl": 0.0947265625,
+      "learning_rate": 3.5885794819795566e-06,
+      "loss": 0.216,
+      "step": 198,
+      "step_loss": 0.09326171875
+    },
+    {
+      "epoch": 1.9320388349514563,
+      "grad_norm": 0.5550608922872593,
+      "kl": 0.423828125,
+      "learning_rate": 3.5757022812117625e-06,
+      "loss": 0.2304,
+      "step": 199,
+      "step_loss": 0.49609375
+    },
+    {
+      "epoch": 1.941747572815534,
+      "grad_norm": 0.5964742703384457,
+      "kl": 0.341796875,
+      "learning_rate": 3.562793748921095e-06,
+      "loss": 0.2283,
+      "step": 200,
+      "step_loss": 0.32421875
+    },
+    {
+      "epoch": 1.9514563106796117,
+      "grad_norm": 0.5752160371372506,
+      "kl": 0.263671875,
+      "learning_rate": 3.5498543749257164e-06,
+      "loss": 0.2166,
+      "step": 201,
+      "step_loss": 0.2041015625
+    },
+    {
+      "epoch": 1.9611650485436893,
+      "grad_norm": 0.547718813991668,
+      "kl": 0.1904296875,
+      "learning_rate": 3.536884650214088e-06,
+      "loss": 0.2374,
+      "step": 202,
+      "step_loss": 0.193359375
+    },
+    {
+      "epoch": 1.970873786407767,
+      "grad_norm": 0.5834419512668307,
+      "kl": 0.20703125,
+      "learning_rate": 3.5238850669263386e-06,
+      "loss": 0.2419,
+      "step": 203,
+      "step_loss": 0.404296875
+    },
+    {
+      "epoch": 1.9805825242718447,
+      "grad_norm": 0.5492481511414377,
+      "kl": 0.171875,
+      "learning_rate": 3.510856118335589e-06,
+      "loss": 0.2304,
+      "step": 204,
+      "step_loss": 0.26953125
+    },
+    {
+      "epoch": 1.9902912621359223,
+      "grad_norm": 0.6611396642693328,
+      "kl": 0.34765625,
+      "learning_rate": 3.497798298829234e-06,
+      "loss": 0.2235,
+      "step": 205,
+      "step_loss": 0.158203125
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.6014585529041143,
+      "kl": 0.19140625,
+      "learning_rate": 3.484712103890188e-06,
+      "loss": 0.2067,
+      "step": 206,
+      "step_loss": 0.23828125
+    },
+    {
+      "epoch": 2.0,
+      "eval_test_transformed.json_loss": NaN,
+      "eval_test_transformed.json_runtime": 8.5733,
+      "eval_test_transformed.json_samples_per_second": 58.321,
+      "eval_test_transformed.json_steps_per_second": 1.866,
+      "step": 206
+    },
+    {
+      "epoch": 2.0097087378640777,
+      "grad_norm": 0.5791202947407729,
+      "kl": NaN,
+      "learning_rate": 3.471598030078074e-06,
+      "loss": 0.1895,
+      "step": 207,
+      "step_loss": NaN
+    },
+    {
+      "epoch": 2.0194174757281553,
+      "grad_norm": 0.5095592680348658,
+      "kl": 0.2373046875,
+      "learning_rate": 3.4584565750103932e-06,
+      "loss": 0.1777,
+      "step": 208,
+      "step_loss": 0.25390625
+    },
+    {
+      "epoch": 2.029126213592233,
+      "grad_norm": 0.52090270216247,
+      "kl": 0.2041015625,
+      "learning_rate": 3.445288237343632e-06,
+      "loss": 0.1885,
+      "step": 209,
+      "step_loss": 0.240234375
+    },
+    {
+      "epoch": 2.0388349514563107,
+      "grad_norm": 0.4976045327134996,
+      "kl": 0.2431640625,
+      "learning_rate": 3.432093516754348e-06,
+      "loss": 0.1863,
+      "step": 210,
+      "step_loss": 0.1064453125
+    },
+    {
+      "epoch": 2.0485436893203883,
+      "grad_norm": 0.47846076806594573,
+      "kl": 0.25390625,
+      "learning_rate": 3.4188729139202063e-06,
+      "loss": 0.1806,
+      "step": 211,
+      "step_loss": 0.138671875
+    },
+    {
+      "epoch": 2.058252427184466,
+      "grad_norm": 0.4911716210346588,
+      "kl": 0.1787109375,
+      "learning_rate": 3.4056269305009807e-06,
+      "loss": 0.1963,
+      "step": 212,
+      "step_loss": 0.1650390625
+    },
+    {
+      "epoch": 2.0679611650485437,
+      "grad_norm": 0.5097950309369077,
+      "kl": 0.1708984375,
+      "learning_rate": 3.3923560691195194e-06,
+      "loss": 0.1745,
+      "step": 213,
+      "step_loss": 0.232421875
+    },
+    {
+      "epoch": 2.0776699029126213,
+      "grad_norm": 0.49796056560568963,
+      "kl": 0.390625,
+      "learning_rate": 3.379060833342673e-06,
+      "loss": 0.1956,
+      "step": 214,
+      "step_loss": 0.134765625
+    },
+    {
+      "epoch": 2.087378640776699,
+      "grad_norm": 0.5533404313296716,
+      "kl": 0.1796875,
+      "learning_rate": 3.3657417276621867e-06,
+      "loss": 0.1977,
+      "step": 215,
+      "step_loss": 0.2890625
+    },
+    {
+      "epoch": 2.0970873786407767,
+      "grad_norm": 0.5296091938932913,
+      "kl": 0.251953125,
+      "learning_rate": 3.352399257475553e-06,
+      "loss": 0.1799,
+      "step": 216,
+      "step_loss": 0.296875
+    },
+    {
+      "epoch": 2.1067961165048543,
+      "grad_norm": 0.6482221636505494,
+      "kl": 0.435546875,
+      "learning_rate": 3.339033929066841e-06,
+      "loss": 0.1938,
+      "step": 217,
+      "step_loss": 0.134765625
+    },
+    {
+      "epoch": 2.116504854368932,
+      "grad_norm": 0.5671573525815887,
+      "kl": 0.28515625,
+      "learning_rate": 3.3256462495874804e-06,
+      "loss": 0.2086,
+      "step": 218,
+      "step_loss": 0.259765625
+    },
+    {
+      "epoch": 2.1262135922330097,
+      "grad_norm": 0.5195068124172081,
+      "kl": 0.1962890625,
+      "learning_rate": 3.3122367270370193e-06,
+      "loss": 0.1836,
+      "step": 219,
+      "step_loss": 0.33984375
+    },
+    {
+      "epoch": 2.1359223300970873,
+      "grad_norm": 0.5586205055924973,
+      "kl": 0.2373046875,
+      "learning_rate": 3.2988058702438493e-06,
+      "loss": 0.2151,
+      "step": 220,
+      "step_loss": 0.1455078125
+    },
+    {
+      "epoch": 2.145631067961165,
+      "grad_norm": 0.4963165143500007,
+      "kl": 0.3046875,
+      "learning_rate": 3.285354188845892e-06,
+      "loss": 0.1908,
+      "step": 221,
+      "step_loss": 0.140625
+    },
+    {
+      "epoch": 2.1553398058252426,
+      "grad_norm": 0.5444372065813423,
+      "kl": 0.296875,
+      "learning_rate": 3.271882193271271e-06,
+      "loss": 0.2084,
+      "step": 222,
+      "step_loss": 0.18359375
+    },
+    {
+      "epoch": 2.1650485436893203,
+      "grad_norm": 0.5399029855207185,
+      "kl": 1.0546875,
+      "learning_rate": 3.258390394718933e-06,
+      "loss": 0.2045,
+      "step": 223,
+      "step_loss": 0.3125
+    },
+    {
+      "epoch": 2.174757281553398,
+      "grad_norm": 0.5319895925471867,
+      "kl": 0.345703125,
+      "learning_rate": 3.2448793051392546e-06,
+      "loss": 0.1971,
+      "step": 224,
+      "step_loss": 0.109375
+    },
+    {
+      "epoch": 2.1844660194174756,
+      "grad_norm": 0.6543149803275666,
+      "kl": 0.162109375,
+      "learning_rate": 3.231349437214619e-06,
+      "loss": 0.1733,
+      "step": 225,
+      "step_loss": 0.267578125
+    },
+    {
+      "epoch": 2.1941747572815533,
+      "grad_norm": 0.5510963728874484,
+      "kl": 0.2021484375,
+      "learning_rate": 3.2178013043399588e-06,
+      "loss": 0.172,
+      "step": 226,
+      "step_loss": 0.2109375
+    },
+    {
+      "epoch": 2.203883495145631,
+      "grad_norm": 0.5155354624371081,
+      "kl": 0.435546875,
+      "learning_rate": 3.2042354206032733e-06,
+      "loss": 0.1785,
+      "step": 227,
+      "step_loss": 0.1484375
+    },
+    {
+      "epoch": 2.2135922330097086,
+      "grad_norm": 0.5053157167916185,
+      "kl": 0.341796875,
+      "learning_rate": 3.190652300766126e-06,
+      "loss": 0.1685,
+      "step": 228,
+      "step_loss": 0.0595703125
+    },
+    {
+      "epoch": 2.2233009708737863,
+      "grad_norm": 0.5412285451182728,
+      "kl": 0.294921875,
+      "learning_rate": 3.1770524602441085e-06,
+      "loss": 0.1945,
+      "step": 229,
+      "step_loss": 0.171875
+    },
+    {
+      "epoch": 2.233009708737864,
+      "grad_norm": 0.7645124019667211,
+      "kl": 0.337890625,
+      "learning_rate": 3.163436415087283e-06,
+      "loss": 0.1786,
+      "step": 230,
+      "step_loss": 0.0947265625
+    },
+    {
+      "epoch": 2.2427184466019416,
+      "grad_norm": 0.5902838675490253,
+      "kl": 0.177734375,
+      "learning_rate": 3.1498046819606046e-06,
+      "loss": 0.1995,
+      "step": 231,
+      "step_loss": 0.1884765625
+    },
+    {
+      "epoch": 2.2524271844660193,
+      "grad_norm": 0.6016854494954839,
+      "kl": 0.416015625,
+      "learning_rate": 3.13615777812431e-06,
+      "loss": 0.2059,
+      "step": 232,
+      "step_loss": 0.189453125
+    },
+    {
+      "epoch": 2.262135922330097,
+      "grad_norm": 0.49283185288305054,
+      "kl": 0.169921875,
+      "learning_rate": 3.122496221414293e-06,
+      "loss": 0.1833,
+      "step": 233,
+      "step_loss": 0.1328125
+    },
+    {
+      "epoch": 2.2718446601941746,
+      "grad_norm": 0.5176674531829142,
+      "kl": 0.1953125,
+      "learning_rate": 3.108820530222458e-06,
+      "loss": 0.2065,
+      "step": 234,
+      "step_loss": 0.2734375
+    },
+    {
+      "epoch": 2.2815533980582523,
+      "grad_norm": 0.571466923601719,
+      "kl": 0.234375,
+      "learning_rate": 3.0951312234770427e-06,
+      "loss": 0.1995,
+      "step": 235,
+      "step_loss": 0.1318359375
+    },
+    {
+      "epoch": 2.29126213592233,
+      "grad_norm": 0.46009947680771285,
+      "kl": 0.283203125,
+      "learning_rate": 3.081428820622935e-06,
+      "loss": 0.1821,
+      "step": 236,
+      "step_loss": 0.1689453125
+    },
+    {
+      "epoch": 2.3009708737864076,
+      "grad_norm": 0.4656876349084823,
+      "kl": 0.2158203125,
+      "learning_rate": 3.067713841601956e-06,
+      "loss": 0.1744,
+      "step": 237,
+      "step_loss": 0.1328125
+    },
+    {
+      "epoch": 2.3106796116504853,
+      "grad_norm": 0.5238872133477445,
+      "kl": 0.1767578125,
+      "learning_rate": 3.0539868068331345e-06,
+      "loss": 0.21,
+      "step": 238,
+      "step_loss": 0.353515625
+    },
+    {
+      "epoch": 2.320388349514563,
+      "grad_norm": 0.5445354007250789,
+      "kl": 0.240234375,
+      "learning_rate": 3.040248237192958e-06,
+      "loss": 0.1729,
+      "step": 239,
+      "step_loss": 0.12109375
+    },
+    {
+      "epoch": 2.3300970873786406,
+      "grad_norm": 0.5420652323121445,
+      "kl": 0.193359375,
+      "learning_rate": 3.026498653995607e-06,
+      "loss": 0.2176,
+      "step": 240,
+      "step_loss": 0.330078125
+    },
+    {
+      "epoch": 2.3398058252427183,
+      "grad_norm": 0.5088097390569825,
+      "kl": 0.361328125,
+      "learning_rate": 3.0127385789731773e-06,
+      "loss": 0.199,
+      "step": 241,
+      "step_loss": 0.107421875
+    },
+    {
+      "epoch": 2.349514563106796,
+      "grad_norm": 0.5272400930748196,
+      "kl": 0.1708984375,
+      "learning_rate": 2.9989685342558776e-06,
+      "loss": 0.1906,
+      "step": 242,
+      "step_loss": 0.21484375
+    },
+    {
+      "epoch": 2.3592233009708736,
+      "grad_norm": 0.5369149150820804,
+      "kl": 0.16015625,
+      "learning_rate": 2.9851890423522214e-06,
+      "loss": 0.1698,
+      "step": 243,
+      "step_loss": 0.2080078125
+    },
+    {
+      "epoch": 2.3689320388349513,
+      "grad_norm": 0.48739367878418577,
+      "kl": 0.18359375,
+      "learning_rate": 2.9714006261291967e-06,
+      "loss": 0.1903,
+      "step": 244,
+      "step_loss": 0.123046875
+    },
+    {
+      "epoch": 2.378640776699029,
+      "grad_norm": 0.5086942793116336,
+      "kl": 0.28515625,
+      "learning_rate": 2.9576038087924304e-06,
+      "loss": 0.1919,
+      "step": 245,
+      "step_loss": 0.1181640625
+    },
+    {
+      "epoch": 2.3883495145631066,
+      "grad_norm": 0.511491215329325,
+      "kl": 0.287109375,
+      "learning_rate": 2.943799113866329e-06,
+      "loss": 0.17,
+      "step": 246,
+      "step_loss": 0.1376953125
+    },
+    {
+      "epoch": 2.3980582524271843,
+      "grad_norm": 0.4862420492331248,
+      "kl": 0.2109375,
+      "learning_rate": 2.929987065174219e-06,
+      "loss": 0.1858,
+      "step": 247,
+      "step_loss": 0.0888671875
+    },
+    {
+      "epoch": 2.407766990291262,
+      "grad_norm": 0.5342355311910822,
+      "kl": 0.408203125,
+      "learning_rate": 2.9161681868184673e-06,
+      "loss": 0.1747,
+      "step": 248,
+      "step_loss": 0.103515625
+    },
+    {
+      "epoch": 2.4174757281553396,
+      "grad_norm": 0.48746202027489255,
+      "kl": 0.359375,
+      "learning_rate": 2.9023430031605928e-06,
+      "loss": 0.1628,
+      "step": 249,
+      "step_loss": 0.12890625
+    },
+    {
+      "epoch": 2.4271844660194173,
+      "grad_norm": 0.5273774877988189,
+      "kl": 0.30078125,
+      "learning_rate": 2.888512038801372e-06,
+      "loss": 0.1826,
+      "step": 250,
+      "step_loss": 0.1435546875
+    },
+    {
+      "epoch": 2.436893203883495,
+      "grad_norm": 0.5621617942900121,
+      "kl": 0.2373046875,
+      "learning_rate": 2.874675818560933e-06,
+      "loss": 0.2022,
+      "step": 251,
+      "step_loss": 0.1572265625
+    },
+    {
+      "epoch": 2.4466019417475726,
+      "grad_norm": 0.5789885643882243,
+      "kl": 0.208984375,
+      "learning_rate": 2.8608348674588383e-06,
+      "loss": 0.1643,
+      "step": 252,
+      "step_loss": 0.1279296875
+    },
+    {
+      "epoch": 2.4563106796116507,
+      "grad_norm": 0.5470908147857678,
+      "kl": 0.12060546875,
+      "learning_rate": 2.8469897106941657e-06,
+      "loss": 0.1879,
+      "step": 253,
+      "step_loss": 0.1455078125
+    },
+    {
+      "epoch": 2.466019417475728,
+      "grad_norm": 0.5350773885496009,
+      "kl": 0.24609375,
+      "learning_rate": 2.8331408736255766e-06,
+      "loss": 0.1936,
+      "step": 254,
+      "step_loss": 0.2197265625
+    },
+    {
+      "epoch": 2.475728155339806,
+      "grad_norm": 0.5529985369495677,
+      "kl": 0.251953125,
+      "learning_rate": 2.8192888817513844e-06,
+      "loss": 0.1665,
+      "step": 255,
+      "step_loss": 0.1728515625
+    },
+    {
+      "epoch": 2.4854368932038833,
+      "grad_norm": 0.5432455331659981,
+      "kl": 0.10888671875,
+      "learning_rate": 2.8054342606896102e-06,
+      "loss": 0.1795,
+      "step": 256,
+      "step_loss": 0.09912109375
+    },
+    {
+      "epoch": 2.4951456310679614,
+      "grad_norm": 0.6472916647922633,
+      "kl": 0.306640625,
+      "learning_rate": 2.7915775361580427e-06,
+      "loss": 0.178,
+      "step": 257,
+      "step_loss": 0.1982421875
+    },
+    {
+      "epoch": 2.5048543689320386,
+      "grad_norm": 0.5736017610369268,
+      "kl": 0.349609375,
+      "learning_rate": 2.7777192339542867e-06,
+      "loss": 0.22,
+      "step": 258,
+      "step_loss": 0.248046875
+    },
+    {
+      "epoch": 2.5145631067961167,
+      "grad_norm": 0.5113226835633047,
+      "kl": 0.19140625,
+      "learning_rate": 2.7638598799358123e-06,
+      "loss": 0.1816,
+      "step": 259,
+      "step_loss": 0.12109375
+    },
+    {
+      "epoch": 2.524271844660194,
+      "grad_norm": 0.4926253445141819,
+      "kl": 0.333984375,
+      "learning_rate": 2.7500000000000004e-06,
+      "loss": 0.187,
+      "step": 260,
+      "step_loss": 0.1181640625
+    },
+    {
+      "epoch": 2.533980582524272,
+      "grad_norm": 0.5196079398321654,
+      "kl": 0.15625,
+      "learning_rate": 2.7361401200641884e-06,
+      "loss": 0.2047,
+      "step": 261,
+      "step_loss": 0.16796875
+    },
+    {
+      "epoch": 2.5436893203883493,
+      "grad_norm": 0.5367574475148829,
+      "kl": 0.134765625,
+      "learning_rate": 2.722280766045714e-06,
+      "loss": 0.1701,
+      "step": 262,
+      "step_loss": 0.2275390625
+    },
+    {
+      "epoch": 2.5533980582524274,
+      "grad_norm": 0.6249660005858519,
+      "kl": 0.298828125,
+      "learning_rate": 2.708422463841958e-06,
+      "loss": 0.2044,
+      "step": 263,
+      "step_loss": 0.3203125
+    },
+    {
+      "epoch": 2.5631067961165046,
+      "grad_norm": 0.5255108947368616,
+      "kl": 0.142578125,
+      "learning_rate": 2.6945657393103913e-06,
+      "loss": 0.2174,
+      "step": 264,
+      "step_loss": 0.267578125
+    },
+    {
+      "epoch": 2.5728155339805827,
+      "grad_norm": 0.5000574555764202,
+      "kl": 0.1171875,
+      "learning_rate": 2.680711118248617e-06,
+      "loss": 0.1692,
+      "step": 265,
+      "step_loss": 0.10302734375
+    },
+    {
+      "epoch": 2.58252427184466,
+      "grad_norm": 0.5928812196080804,
+      "kl": 0.1923828125,
+      "learning_rate": 2.666859126374425e-06,
+      "loss": 0.2014,
+      "step": 266,
+      "step_loss": 0.3203125
+    },
+    {
+      "epoch": 2.592233009708738,
+      "grad_norm": 0.520932366174803,
+      "kl": 0.2490234375,
+      "learning_rate": 2.653010289305835e-06,
+      "loss": 0.1857,
+      "step": 267,
+      "step_loss": 0.134765625
+    },
+    {
+      "epoch": 2.6019417475728153,
+      "grad_norm": 0.5199064176512976,
+      "kl": 0.255859375,
+      "learning_rate": 2.639165132541162e-06,
+      "loss": 0.1805,
+      "step": 268,
+      "step_loss": 0.234375
+    },
+    {
+      "epoch": 2.6116504854368934,
+      "grad_norm": 0.5189304175947423,
+      "kl": 0.259765625,
+      "learning_rate": 2.625324181439068e-06,
+      "loss": 0.1857,
+      "step": 269,
+      "step_loss": 0.333984375
+    },
+    {
+      "epoch": 2.6213592233009706,
+      "grad_norm": 0.5477820911661071,
+      "kl": 0.1962890625,
+      "learning_rate": 2.611487961198629e-06,
+      "loss": 0.1705,
+      "step": 270,
+      "step_loss": 0.267578125
+    },
+    {
+      "epoch": 2.6310679611650487,
+      "grad_norm": 0.5144960586834338,
+      "kl": 0.3515625,
+      "learning_rate": 2.5976569968394084e-06,
+      "loss": 0.1824,
+      "step": 271,
+      "step_loss": 0.12890625
+    },
+    {
+      "epoch": 2.6407766990291264,
+      "grad_norm": 0.5129225163884132,
+      "kl": 0.1611328125,
+      "learning_rate": 2.583831813181534e-06,
+      "loss": 0.1936,
+      "step": 272,
+      "step_loss": 0.240234375
+    },
+    {
+      "epoch": 2.650485436893204,
+      "grad_norm": 0.5747902549826789,
+      "kl": 0.376953125,
+      "learning_rate": 2.570012934825782e-06,
+      "loss": 0.1803,
+      "step": 273,
+      "step_loss": 0.171875
+    },
+    {
+      "epoch": 2.6601941747572817,
+      "grad_norm": 0.5303667804765286,
+      "kl": 0.2353515625,
+      "learning_rate": 2.556200886133672e-06,
+      "loss": 0.1638,
+      "step": 274,
+      "step_loss": 0.353515625
+    },
+    {
+      "epoch": 2.6699029126213594,
+      "grad_norm": 0.552737610069486,
+      "kl": 0.166015625,
+      "learning_rate": 2.542396191207571e-06,
+      "loss": 0.1896,
+      "step": 275,
+      "step_loss": 0.212890625
+    },
+    {
+      "epoch": 2.679611650485437,
+      "grad_norm": 0.5138863525528505,
+      "kl": 0.267578125,
+      "learning_rate": 2.528599373870804e-06,
+      "loss": 0.1647,
+      "step": 276,
+      "step_loss": 0.1650390625
+    },
+    {
+      "epoch": 2.6893203883495147,
+      "grad_norm": 0.5816837784109217,
+      "kl": 0.3203125,
+      "learning_rate": 2.5148109576477798e-06,
+      "loss": 0.2022,
+      "step": 277,
+      "step_loss": 0.2734375
+    },
+    {
+      "epoch": 2.6990291262135924,
+      "grad_norm": 0.5946206686019718,
+      "kl": 0.296875,
+      "learning_rate": 2.501031465744123e-06,
+      "loss": 0.1793,
+      "step": 278,
+      "step_loss": 0.1650390625
+    },
+    {
+      "epoch": 2.70873786407767,
+      "grad_norm": 0.5505786008910999,
+      "kl": 0.26171875,
+      "learning_rate": 2.487261421026823e-06,
+      "loss": 0.1869,
+      "step": 279,
+      "step_loss": 0.189453125
+    },
+    {
+      "epoch": 2.7184466019417477,
+      "grad_norm": 0.5607618391578882,
+      "kl": 0.396484375,
+      "learning_rate": 2.4735013460043933e-06,
+      "loss": 0.1782,
+      "step": 280,
+      "step_loss": 0.140625
+    },
+    {
+      "epoch": 2.7281553398058254,
+      "grad_norm": 0.532008359083315,
+      "kl": 0.443359375,
+      "learning_rate": 2.4597517628070427e-06,
+      "loss": 0.2004,
+      "step": 281,
+      "step_loss": 0.1953125
+    },
+    {
+      "epoch": 2.737864077669903,
+      "grad_norm": 0.5409775868618988,
+      "kl": 0.412109375,
+      "learning_rate": 2.446013193166866e-06,
+      "loss": 0.1968,
+      "step": 282,
+      "step_loss": 0.10302734375
+    },
+    {
+      "epoch": 2.7475728155339807,
+      "grad_norm": 0.5494599421845879,
+      "kl": 0.291015625,
+      "learning_rate": 2.4322861583980445e-06,
+      "loss": 0.1882,
+      "step": 283,
+      "step_loss": 0.138671875
+    },
+    {
+      "epoch": 2.7572815533980584,
+      "grad_norm": 0.5270774451468452,
+      "kl": 0.275390625,
+      "learning_rate": 2.4185711793770662e-06,
+      "loss": 0.1773,
+      "step": 284,
+      "step_loss": 0.1728515625
+    },
+    {
+      "epoch": 2.766990291262136,
+      "grad_norm": 0.556579747133581,
+      "kl": 0.177734375,
+      "learning_rate": 2.4048687765229584e-06,
+      "loss": 0.1875,
+      "step": 285,
+      "step_loss": 0.12109375
+    },
+    {
+      "epoch": 2.7766990291262137,
+      "grad_norm": 0.503176078809862,
+      "kl": 0.189453125,
+      "learning_rate": 2.3911794697775437e-06,
+      "loss": 0.1847,
+      "step": 286,
+      "step_loss": 0.328125
+    },
+    {
+      "epoch": 2.7864077669902914,
+      "grad_norm": 0.5914318608290587,
+      "kl": 0.251953125,
+      "learning_rate": 2.377503778585708e-06,
+      "loss": 0.2007,
+      "step": 287,
+      "step_loss": 0.197265625
+    },
+    {
+      "epoch": 2.796116504854369,
+      "grad_norm": 0.47771283753849275,
+      "kl": 0.365234375,
+      "learning_rate": 2.3638422218756905e-06,
+      "loss": 0.1968,
+      "step": 288,
+      "step_loss": 0.13671875
+    },
+    {
+      "epoch": 2.8058252427184467,
+      "grad_norm": 0.544279916739223,
+      "kl": 0.1630859375,
+      "learning_rate": 2.350195318039396e-06,
+      "loss": 0.1844,
+      "step": 289,
+      "step_loss": 0.1279296875
+    },
+    {
+      "epoch": 2.8155339805825244,
+      "grad_norm": 0.7140740591904084,
+      "kl": 0.2080078125,
+      "learning_rate": 2.336563584912717e-06,
+      "loss": 0.2153,
+      "step": 290,
+      "step_loss": 0.294921875
+    },
+    {
+      "epoch": 2.825242718446602,
+      "grad_norm": 0.5978474529315786,
+      "kl": 0.1943359375,
+      "learning_rate": 2.322947539755892e-06,
+      "loss": 0.2018,
+      "step": 291,
+      "step_loss": 0.25390625
+    },
+    {
+      "epoch": 2.8349514563106797,
+      "grad_norm": 0.5338842771148811,
+      "kl": 0.4296875,
+      "learning_rate": 2.3093476992338747e-06,
+      "loss": 0.2051,
+      "step": 292,
+      "step_loss": 0.142578125
+    },
+    {
+      "epoch": 2.8446601941747574,
+      "grad_norm": 0.5434580731788923,
+      "kl": 0.2021484375,
+      "learning_rate": 2.295764579396727e-06,
+      "loss": 0.1833,
+      "step": 293,
+      "step_loss": 0.1259765625
+    },
+    {
+      "epoch": 2.854368932038835,
+      "grad_norm": 0.48934430178338256,
+      "kl": 0.1484375,
+      "learning_rate": 2.2821986956600415e-06,
+      "loss": 0.1602,
+      "step": 294,
+      "step_loss": 0.10595703125
+    },
+    {
+      "epoch": 2.8640776699029127,
+      "grad_norm": 0.5408973977811805,
+      "kl": 0.169921875,
+      "learning_rate": 2.2686505627853812e-06,
+      "loss": 0.1594,
+      "step": 295,
+      "step_loss": 0.1103515625
+    },
+    {
+      "epoch": 2.8737864077669903,
+      "grad_norm": 0.5283594028868295,
+      "kl": 0.36328125,
+      "learning_rate": 2.255120694860746e-06,
+      "loss": 0.1916,
+      "step": 296,
+      "step_loss": 0.126953125
+    },
+    {
+      "epoch": 2.883495145631068,
+      "grad_norm": 0.5035651183385199,
+      "kl": 0.2177734375,
+      "learning_rate": 2.2416096052810683e-06,
+      "loss": 0.1824,
+      "step": 297,
+      "step_loss": 0.220703125
+    },
+    {
+      "epoch": 2.8932038834951457,
+      "grad_norm": 0.5459862981254268,
+      "kl": 0.2470703125,
+      "learning_rate": 2.2281178067287294e-06,
+      "loss": 0.2043,
+      "step": 298,
+      "step_loss": 0.205078125
+    },
+    {
+      "epoch": 2.9029126213592233,
+      "grad_norm": 0.5238184894583012,
+      "kl": 0.30078125,
+      "learning_rate": 2.214645811154108e-06,
+      "loss": 0.1867,
+      "step": 299,
+      "step_loss": 0.2119140625
+    },
+    {
+      "epoch": 2.912621359223301,
+      "grad_norm": 0.5030938038128431,
+      "kl": 0.251953125,
+      "learning_rate": 2.201194129756152e-06,
+      "loss": 0.1743,
+      "step": 300,
+      "step_loss": 0.1611328125
+    },
+    {
+      "epoch": 2.9223300970873787,
+      "grad_norm": 0.6344887467056015,
+      "kl": 0.1796875,
+      "learning_rate": 2.187763272962981e-06,
+      "loss": 0.1986,
+      "step": 301,
+      "step_loss": 0.1328125
+    },
+    {
+      "epoch": 2.9320388349514563,
+      "grad_norm": 0.5310139131483351,
+      "kl": 0.189453125,
+      "learning_rate": 2.1743537504125208e-06,
+      "loss": 0.1998,
+      "step": 302,
+      "step_loss": 0.287109375
+    },
+    {
+      "epoch": 2.941747572815534,
+      "grad_norm": 0.5172075780561182,
+      "kl": 0.4375,
+      "learning_rate": 2.1609660709331598e-06,
+      "loss": 0.2207,
+      "step": 303,
+      "step_loss": 0.31640625
+    },
+    {
+      "epoch": 2.9514563106796117,
+      "grad_norm": 0.5624076449552908,
+      "kl": 0.1748046875,
+      "learning_rate": 2.1476007425244476e-06,
+      "loss": 0.1955,
+      "step": 304,
+      "step_loss": 0.1015625
+    },
+    {
+      "epoch": 2.9611650485436893,
+      "grad_norm": 0.5232258269061456,
+      "kl": 0.306640625,
+      "learning_rate": 2.134258272337814e-06,
+      "loss": 0.1884,
+      "step": 305,
+      "step_loss": 0.21484375
+    },
+    {
+      "epoch": 2.970873786407767,
+      "grad_norm": 0.5856686801966252,
+      "kl": 0.291015625,
+      "learning_rate": 2.120939166657327e-06,
+      "loss": 0.202,
+      "step": 306,
+      "step_loss": 0.2021484375
+    },
+    {
+      "epoch": 2.9805825242718447,
+      "grad_norm": 0.5360709399599322,
+      "kl": 0.29296875,
+      "learning_rate": 2.1076439308804813e-06,
+      "loss": 0.1946,
+      "step": 307,
+      "step_loss": 0.150390625
+    },
+    {
+      "epoch": 2.9902912621359223,
+      "grad_norm": 0.5186377564757945,
+      "kl": 0.1923828125,
+      "learning_rate": 2.0943730694990204e-06,
+      "loss": 0.1776,
+      "step": 308,
+      "step_loss": 0.1015625
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.5769108575358282,
+      "kl": 0.48828125,
+      "learning_rate": 2.081127086079795e-06,
+      "loss": 0.1656,
+      "step": 309,
+      "step_loss": 0.240234375
+    },
+    {
+      "epoch": 3.0,
+      "eval_test_transformed.json_loss": NaN,
+      "eval_test_transformed.json_runtime": 8.3956,
+      "eval_test_transformed.json_samples_per_second": 59.555,
+      "eval_test_transformed.json_steps_per_second": 1.906,
+      "step": 309
+    },
+    {
+      "epoch": 3.0097087378640777,
+      "grad_norm": 0.5711934327513365,
+      "kl": NaN,
+      "learning_rate": 2.0679064832456523e-06,
+      "loss": 0.1695,
+      "step": 310,
+      "step_loss": NaN
+    },
+    {
+      "epoch": 3.0194174757281553,
+      "grad_norm": 0.5080776402736003,
+      "kl": 0.2255859375,
+      "learning_rate": 2.054711762656369e-06,
+      "loss": 0.1599,
+      "step": 311,
+      "step_loss": 0.2373046875
+    },
+    {
+      "epoch": 3.029126213592233,
+      "grad_norm": 0.460829560524234,
+      "kl": 0.154296875,
+      "learning_rate": 2.0415434249896075e-06,
+      "loss": 0.1554,
+      "step": 312,
+      "step_loss": 0.08984375
+    },
+    {
+      "epoch": 3.0388349514563107,
+      "grad_norm": 0.532111983258558,
+      "kl": 0.259765625,
+      "learning_rate": 2.0284019699219265e-06,
+      "loss": 0.1561,
+      "step": 313,
+      "step_loss": 0.1689453125
+    },
+    {
+      "epoch": 3.0485436893203883,
+      "grad_norm": 0.5807336256828239,
+      "kl": 0.369140625,
+      "learning_rate": 2.0152878961098133e-06,
+      "loss": 0.1513,
+      "step": 314,
+      "step_loss": 0.06201171875
+    },
+    {
+      "epoch": 3.058252427184466,
+      "grad_norm": 0.5177785233543187,
+      "kl": 0.1982421875,
+      "learning_rate": 2.0022017011707663e-06,
+      "loss": 0.1816,
+      "step": 315,
+      "step_loss": 0.2236328125
+    },
+    {
+      "epoch": 3.0679611650485437,
+      "grad_norm": 0.5969479421413822,
+      "kl": 0.333984375,
+      "learning_rate": 1.989143881664412e-06,
+      "loss": 0.1782,
+      "step": 316,
+      "step_loss": 0.240234375
+    },
+    {
+      "epoch": 3.0776699029126213,
+      "grad_norm": 0.489029791228728,
+      "kl": 0.59765625,
+      "learning_rate": 1.9761149330736625e-06,
+      "loss": 0.1308,
+      "step": 317,
+      "step_loss": 0.064453125
+    },
+    {
+      "epoch": 3.087378640776699,
+      "grad_norm": 0.7131615528898937,
+      "kl": 0.177734375,
+      "learning_rate": 1.9631153497859127e-06,
+      "loss": 0.179,
+      "step": 318,
+      "step_loss": 0.07861328125
+    },
+    {
+      "epoch": 3.0970873786407767,
+      "grad_norm": 0.4640669864219324,
+      "kl": 0.390625,
+      "learning_rate": 1.950145625074285e-06,
+      "loss": 0.1331,
+      "step": 319,
+      "step_loss": 0.06494140625
+    },
+    {
+      "epoch": 3.1067961165048543,
+      "grad_norm": 0.4889269717181464,
+      "kl": 0.2138671875,
+      "learning_rate": 1.9372062510789064e-06,
+      "loss": 0.1696,
+      "step": 320,
+      "step_loss": 0.07861328125
+    },
+    {
+      "epoch": 3.116504854368932,
+      "grad_norm": 0.4894140759210115,
+      "kl": 0.294921875,
+      "learning_rate": 1.924297718788238e-06,
+      "loss": 0.1532,
+      "step": 321,
+      "step_loss": 0.1728515625
+    },
+    {
+      "epoch": 3.1262135922330097,
+      "grad_norm": 0.4851533726900198,
+      "kl": 0.271484375,
+      "learning_rate": 1.9114205180204437e-06,
+      "loss": 0.1563,
+      "step": 322,
+      "step_loss": 0.08984375
+    },
+    {
+      "epoch": 3.1359223300970873,
+      "grad_norm": 0.4812749750914589,
+      "kl": 0.3359375,
+      "learning_rate": 1.8985751374048022e-06,
+      "loss": 0.1597,
+      "step": 323,
+      "step_loss": 0.083984375
+    },
+    {
+      "epoch": 3.145631067961165,
+      "grad_norm": 0.4735366094358067,
+      "kl": 0.57421875,
+      "learning_rate": 1.8857620643631675e-06,
+      "loss": 0.164,
+      "step": 324,
+      "step_loss": 0.19140625
+    },
+    {
+      "epoch": 3.1553398058252426,
+      "grad_norm": 0.4616311559837366,
+      "kl": 0.3203125,
+      "learning_rate": 1.8729817850914717e-06,
+      "loss": 0.169,
+      "step": 325,
+      "step_loss": 0.2294921875
+    },
+    {
+      "epoch": 3.1650485436893203,
+      "grad_norm": 0.5563217588279268,
+      "kl": 0.224609375,
+      "learning_rate": 1.8602347845412799e-06,
+      "loss": 0.1502,
+      "step": 326,
+      "step_loss": 0.28515625
+    },
+    {
+      "epoch": 3.174757281553398,
+      "grad_norm": 0.45499622442963633,
+      "kl": 0.330078125,
+      "learning_rate": 1.847521546401383e-06,
+      "loss": 0.1547,
+      "step": 327,
+      "step_loss": 0.062255859375
+    },
+    {
+      "epoch": 3.1844660194174756,
+      "grad_norm": 0.5222912664403042,
+      "kl": 0.33984375,
+      "learning_rate": 1.83484255307945e-06,
+      "loss": 0.1694,
+      "step": 328,
+      "step_loss": 0.1083984375
+    },
+    {
+      "epoch": 3.1941747572815533,
+      "grad_norm": 0.4698877989732488,
+      "kl": 0.1611328125,
+      "learning_rate": 1.8221982856837177e-06,
+      "loss": 0.1616,
+      "step": 329,
+      "step_loss": 0.203125
+    },
+    {
+      "epoch": 3.203883495145631,
+      "grad_norm": 0.5161358585670528,
+      "kl": 0.2578125,
+      "learning_rate": 1.8095892240047375e-06,
+      "loss": 0.1483,
+      "step": 330,
+      "step_loss": 0.193359375
+    },
+    {
+      "epoch": 3.2135922330097086,
+      "grad_norm": 0.5707034508549506,
+      "kl": 0.384765625,
+      "learning_rate": 1.7970158464971704e-06,
+      "loss": 0.1764,
+      "step": 331,
+      "step_loss": 0.1435546875
+    },
+    {
+      "epoch": 3.2233009708737863,
+      "grad_norm": 0.6845693547436684,
+      "kl": 0.30078125,
+      "learning_rate": 1.78447863026163e-06,
+      "loss": 0.1686,
+      "step": 332,
+      "step_loss": 0.146484375
+    },
+    {
+      "epoch": 3.233009708737864,
+      "grad_norm": 0.539242866306224,
+      "kl": 0.279296875,
+      "learning_rate": 1.77197805102658e-06,
+      "loss": 0.1602,
+      "step": 333,
+      "step_loss": 0.30078125
+    },
+    {
+      "epoch": 3.2427184466019416,
+      "grad_norm": 0.5376628174473539,
+      "kl": 0.318359375,
+      "learning_rate": 1.759514583130279e-06,
+      "loss": 0.1596,
+      "step": 334,
+      "step_loss": 0.0859375
+    },
+    {
+      "epoch": 3.2524271844660193,
+      "grad_norm": 0.488395082556356,
+      "kl": 0.421875,
+      "learning_rate": 1.7470886995027902e-06,
+      "loss": 0.1437,
+      "step": 335,
+      "step_loss": 0.07080078125
+    },
+    {
+      "epoch": 3.262135922330097,
+      "grad_norm": 0.5055787180594609,
+      "kl": 0.2099609375,
+      "learning_rate": 1.734700871648024e-06,
+      "loss": 0.1565,
+      "step": 336,
+      "step_loss": 0.1953125
+    },
+    {
+      "epoch": 3.2718446601941746,
+      "grad_norm": 0.4884350496305089,
+      "kl": 0.2431640625,
+      "learning_rate": 1.722351569625859e-06,
+      "loss": 0.1445,
+      "step": 337,
+      "step_loss": 0.07666015625
+    },
+    {
+      "epoch": 3.2815533980582523,
+      "grad_norm": 0.5055878842267785,
+      "kl": 0.1669921875,
+      "learning_rate": 1.710041262034296e-06,
+      "loss": 0.1662,
+      "step": 338,
+      "step_loss": 0.1298828125
+    },
+    {
+      "epoch": 3.29126213592233,
+      "grad_norm": 0.6568055753017068,
+      "kl": 0.27734375,
+      "learning_rate": 1.6977704159916801e-06,
+      "loss": 0.1607,
+      "step": 339,
+      "step_loss": 0.13671875
+    },
+    {
+      "epoch": 3.3009708737864076,
+      "grad_norm": 0.46693217339820037,
+      "kl": 0.326171875,
+      "learning_rate": 1.6855394971189779e-06,
+      "loss": 0.1519,
+      "step": 340,
+      "step_loss": 0.0751953125
+    },
+    {
+      "epoch": 3.3106796116504853,
+      "grad_norm": 0.46750716114679314,
+      "kl": 0.224609375,
+      "learning_rate": 1.6733489695221056e-06,
+      "loss": 0.165,
+      "step": 341,
+      "step_loss": 0.25
+    },
+    {
+      "epoch": 3.320388349514563,
+      "grad_norm": 0.46663445875307946,
+      "kl": 0.359375,
+      "learning_rate": 1.6611992957743217e-06,
+      "loss": 0.151,
+      "step": 342,
+      "step_loss": 0.1318359375
+    },
+    {
+      "epoch": 3.3300970873786406,
+      "grad_norm": 0.48545644382954584,
+      "kl": 0.1982421875,
+      "learning_rate": 1.6490909368986725e-06,
+      "loss": 0.1624,
+      "step": 343,
+      "step_loss": 0.1748046875
+    },
+    {
+      "epoch": 3.3398058252427183,
+      "grad_norm": 0.4692569851993047,
+      "kl": 0.232421875,
+      "learning_rate": 1.637024352350498e-06,
+      "loss": 0.1414,
+      "step": 344,
+      "step_loss": 0.271484375
+    },
+    {
+      "epoch": 3.349514563106796,
+      "grad_norm": 0.4758629432786183,
+      "kl": 0.2490234375,
+      "learning_rate": 1.6250000000000007e-06,
+      "loss": 0.1427,
+      "step": 345,
+      "step_loss": 0.150390625
+    },
+    {
+      "epoch": 3.3592233009708736,
+      "grad_norm": 0.4871466615503954,
+      "kl": 0.29296875,
+      "learning_rate": 1.6130183361148675e-06,
+      "loss": 0.1593,
+      "step": 346,
+      "step_loss": 0.1748046875
+    },
+    {
+      "epoch": 3.3689320388349513,
+      "grad_norm": 0.5474079532412777,
+      "kl": 0.32421875,
+      "learning_rate": 1.60107981534296e-06,
+      "loss": 0.178,
+      "step": 347,
+      "step_loss": 0.267578125
+    },
+    {
+      "epoch": 3.378640776699029,
+      "grad_norm": 2.4760425876288865,
+      "kl": 0.25390625,
+      "learning_rate": 1.5891848906950602e-06,
+      "loss": 0.1768,
+      "step": 348,
+      "step_loss": 0.3515625
+    },
+    {
+      "epoch": 3.3883495145631066,
+      "grad_norm": 0.499334607836154,
+      "kl": 0.3359375,
+      "learning_rate": 1.5773340135276843e-06,
+      "loss": 0.1309,
+      "step": 349,
+      "step_loss": 0.07470703125
+    },
+    {
+      "epoch": 3.3980582524271843,
+      "grad_norm": 0.5121888774785807,
+      "kl": 0.3046875,
+      "learning_rate": 1.5655276335259495e-06,
+      "loss": 0.146,
+      "step": 350,
+      "step_loss": 0.126953125
+    },
+    {
+      "epoch": 3.407766990291262,
+      "grad_norm": 0.5311085568794023,
+      "kl": 0.427734375,
+      "learning_rate": 1.5537661986865196e-06,
+      "loss": 0.1675,
+      "step": 351,
+      "step_loss": 0.1103515625
+    },
+    {
+      "epoch": 3.4174757281553396,
+      "grad_norm": 0.5128192279567273,
+      "kl": 0.271484375,
+      "learning_rate": 1.542050155300595e-06,
+      "loss": 0.1672,
+      "step": 352,
+      "step_loss": 0.2197265625
+    },
+    {
+      "epoch": 3.4271844660194173,
+      "grad_norm": 0.49745836081284944,
+      "kl": 0.26171875,
+      "learning_rate": 1.53037994793699e-06,
+      "loss": 0.1568,
+      "step": 353,
+      "step_loss": 0.18359375
+    },
+    {
+      "epoch": 3.436893203883495,
+      "grad_norm": 0.48051427609991737,
+      "kl": 0.263671875,
+      "learning_rate": 1.51875601942525e-06,
+      "loss": 0.1625,
+      "step": 354,
+      "step_loss": 0.2265625
+    },
+    {
+      "epoch": 3.4466019417475726,
+      "grad_norm": 0.48409810919704643,
+      "kl": 0.29296875,
+      "learning_rate": 1.507178810838863e-06,
+      "loss": 0.1706,
+      "step": 355,
+      "step_loss": 0.1845703125
+    },
+    {
+      "epoch": 3.4563106796116507,
+      "grad_norm": 0.537826558649176,
+      "kl": 0.349609375,
+      "learning_rate": 1.4956487614785076e-06,
+      "loss": 0.16,
+      "step": 356,
+      "step_loss": 0.15625
+    },
+    {
+      "epoch": 3.466019417475728,
+      "grad_norm": 0.507752429637045,
+      "kl": 0.345703125,
+      "learning_rate": 1.4841663088553992e-06,
+      "loss": 0.1625,
+      "step": 357,
+      "step_loss": 0.12890625
+    },
+    {
+      "epoch": 3.475728155339806,
+      "grad_norm": 0.5130052193009929,
+      "kl": 0.341796875,
+      "learning_rate": 1.4727318886746725e-06,
+      "loss": 0.1741,
+      "step": 358,
+      "step_loss": 0.11181640625
+    },
+    {
+      "epoch": 3.4854368932038833,
+      "grad_norm": 0.5556488303190134,
+      "kl": 0.46484375,
+      "learning_rate": 1.4613459348188635e-06,
+      "loss": 0.1242,
+      "step": 359,
+      "step_loss": 0.0869140625
+    },
+    {
+      "epoch": 3.4951456310679614,
+      "grad_norm": 0.477620232983351,
+      "kl": 0.39453125,
+      "learning_rate": 1.4500088793314351e-06,
+      "loss": 0.1439,
+      "step": 360,
+      "step_loss": 0.0703125
+    },
+    {
+      "epoch": 3.5048543689320386,
+      "grad_norm": 0.5136494794829761,
+      "kl": 0.2734375,
+      "learning_rate": 1.438721152400388e-06,
+      "loss": 0.1511,
+      "step": 361,
+      "step_loss": 0.326171875
+    },
+    {
+      "epoch": 3.5145631067961167,
+      "grad_norm": 0.4992556694121463,
+      "kl": 0.26171875,
+      "learning_rate": 1.427483182341936e-06,
+      "loss": 0.1641,
+      "step": 362,
+      "step_loss": 0.0810546875
+    },
+    {
+      "epoch": 3.524271844660194,
+      "grad_norm": 0.4965898781365568,
+      "kl": 0.291015625,
+      "learning_rate": 1.4162953955842518e-06,
+      "loss": 0.1717,
+      "step": 363,
+      "step_loss": 0.06640625
+    },
+    {
+      "epoch": 3.533980582524272,
+      "grad_norm": 0.4885949309810252,
+      "kl": 0.2490234375,
+      "learning_rate": 1.4051582166512895e-06,
+      "loss": 0.1377,
+      "step": 364,
+      "step_loss": 0.0830078125
+    },
+    {
+      "epoch": 3.5436893203883493,
+      "grad_norm": 0.7644545502958816,
+      "kl": 0.3203125,
+      "learning_rate": 1.3940720681466735e-06,
+      "loss": 0.1581,
+      "step": 365,
+      "step_loss": 0.17578125
+    },
+    {
+      "epoch": 3.5533980582524274,
+      "grad_norm": 0.549140542595152,
+      "kl": 0.26953125,
+      "learning_rate": 1.3830373707376623e-06,
+      "loss": 0.1455,
+      "step": 366,
+      "step_loss": 0.1533203125
+    },
+    {
+      "epoch": 3.5631067961165046,
+      "grad_norm": 0.7512378808438823,
+      "kl": 0.333984375,
+      "learning_rate": 1.372054543139188e-06,
+      "loss": 0.15,
+      "step": 367,
+      "step_loss": 0.1103515625
+    },
+    {
+      "epoch": 3.5728155339805827,
+      "grad_norm": 0.5020029362406481,
+      "kl": 0.19921875,
+      "learning_rate": 1.3611240020979655e-06,
+      "loss": 0.1427,
+      "step": 368,
+      "step_loss": 0.06494140625
+    },
+    {
+      "epoch": 3.58252427184466,
+      "grad_norm": 0.46369749478886507,
+      "kl": 0.1650390625,
+      "learning_rate": 1.3502461623766842e-06,
+      "loss": 0.1668,
+      "step": 369,
+      "step_loss": 0.1728515625
+    },
+    {
+      "epoch": 3.592233009708738,
+      "grad_norm": 0.527596307252248,
+      "kl": 0.404296875,
+      "learning_rate": 1.3394214367382602e-06,
+      "loss": 0.1543,
+      "step": 370,
+      "step_loss": 0.19921875
+    },
+    {
+      "epoch": 3.6019417475728153,
+      "grad_norm": 0.5072354485956617,
+      "kl": 0.2578125,
+      "learning_rate": 1.3286502359301863e-06,
+      "loss": 0.1599,
+      "step": 371,
+      "step_loss": 0.1904296875
+    },
+    {
+      "epoch": 3.6116504854368934,
+      "grad_norm": 0.5160691631303242,
+      "kl": 0.439453125,
+      "learning_rate": 1.3179329686689318e-06,
+      "loss": 0.1498,
+      "step": 372,
+      "step_loss": 0.169921875
+    },
+    {
+      "epoch": 3.6213592233009706,
+      "grad_norm": 0.6554485149692931,
+      "kl": 0.2412109375,
+      "learning_rate": 1.3072700416244494e-06,
+      "loss": 0.1542,
+      "step": 373,
+      "step_loss": 0.228515625
+    },
+    {
+      "epoch": 3.6310679611650487,
+      "grad_norm": 0.4910381460677276,
+      "kl": 0.240234375,
+      "learning_rate": 1.2966618594047285e-06,
+      "loss": 0.1407,
+      "step": 374,
+      "step_loss": 0.322265625
+    },
+    {
+      "epoch": 3.6407766990291264,
+      "grad_norm": 0.5460180925997682,
+      "kl": 0.25,
+      "learning_rate": 1.286108824540456e-06,
+      "loss": 0.1558,
+      "step": 375,
+      "step_loss": 0.2236328125
+    },
+    {
+      "epoch": 3.650485436893204,
+      "grad_norm": 0.584914369592297,
+      "kl": 0.53125,
+      "learning_rate": 1.2756113374697294e-06,
+      "loss": 0.1594,
+      "step": 376,
+      "step_loss": 0.1318359375
+    },
+    {
+      "epoch": 3.6601941747572817,
+      "grad_norm": 0.5256624617590849,
+      "kl": 0.189453125,
+      "learning_rate": 1.2651697965228748e-06,
+      "loss": 0.1696,
+      "step": 377,
+      "step_loss": 0.1123046875
+    },
+    {
+      "epoch": 3.6699029126213594,
+      "grad_norm": 0.5033922457112555,
+      "kl": 0.216796875,
+      "learning_rate": 1.2547845979073194e-06,
+      "loss": 0.1567,
+      "step": 378,
+      "step_loss": 0.2412109375
+    },
+    {
+      "epoch": 3.679611650485437,
+      "grad_norm": 0.5191896792873844,
+      "kl": 0.1494140625,
+      "learning_rate": 1.2444561356925692e-06,
+      "loss": 0.1413,
+      "step": 379,
+      "step_loss": 0.10205078125
+    },
+    {
+      "epoch": 3.6893203883495147,
+      "grad_norm": 0.5166370978149198,
+      "kl": 0.46484375,
+      "learning_rate": 1.2341848017952464e-06,
+      "loss": 0.1658,
+      "step": 380,
+      "step_loss": 0.1005859375
+    },
+    {
+      "epoch": 3.6990291262135924,
+      "grad_norm": 0.5574979370660633,
+      "kl": 0.251953125,
+      "learning_rate": 1.2239709859642237e-06,
+      "loss": 0.1484,
+      "step": 381,
+      "step_loss": 0.2119140625
+    },
+    {
+      "epoch": 3.70873786407767,
+      "grad_norm": 0.5625584889838852,
+      "kl": 0.5390625,
+      "learning_rate": 1.2138150757658328e-06,
+      "loss": 0.1802,
+      "step": 382,
+      "step_loss": 0.17578125
+    },
+    {
+      "epoch": 3.7184466019417477,
+      "grad_norm": 0.47927966720420184,
+      "kl": 0.283203125,
+      "learning_rate": 1.2037174565691591e-06,
+      "loss": 0.1506,
+      "step": 383,
+      "step_loss": 0.146484375
+    },
+    {
+      "epoch": 3.7281553398058254,
+      "grad_norm": 0.5194167964748884,
+      "kl": 0.4453125,
+      "learning_rate": 1.1936785115314176e-06,
+      "loss": 0.1395,
+      "step": 384,
+      "step_loss": 0.10498046875
+    },
+    {
+      "epoch": 3.737864077669903,
+      "grad_norm": 0.5468254611874791,
+      "kl": 0.306640625,
+      "learning_rate": 1.1836986215834153e-06,
+      "loss": 0.1423,
+      "step": 385,
+      "step_loss": 0.12158203125
+    },
+    {
+      "epoch": 3.7475728155339807,
+      "grad_norm": 0.5071633518038373,
+      "kl": 0.11865234375,
+      "learning_rate": 1.1737781654150955e-06,
+      "loss": 0.162,
+      "step": 386,
+      "step_loss": 0.12451171875
+    },
+    {
+      "epoch": 3.7572815533980584,
+      "grad_norm": 0.48157726901389747,
+      "kl": 0.2353515625,
+      "learning_rate": 1.1639175194611693e-06,
+      "loss": 0.1547,
+      "step": 387,
+      "step_loss": 0.1328125
+    },
+    {
+      "epoch": 3.766990291262136,
+      "grad_norm": 0.541213622511581,
+      "kl": 0.251953125,
+      "learning_rate": 1.15411705788683e-06,
+      "loss": 0.1655,
+      "step": 388,
+      "step_loss": 0.107421875
+    },
+    {
+      "epoch": 3.7766990291262137,
+      "grad_norm": 0.5615405005474776,
+      "kl": 0.212890625,
+      "learning_rate": 1.1443771525735577e-06,
+      "loss": 0.1576,
+      "step": 389,
+      "step_loss": 0.1767578125
+    },
+    {
+      "epoch": 3.7864077669902914,
+      "grad_norm": 0.4950987080693749,
+      "kl": 0.169921875,
+      "learning_rate": 1.1346981731050051e-06,
+      "loss": 0.1565,
+      "step": 390,
+      "step_loss": 0.2578125
+    },
+    {
+      "epoch": 3.796116504854369,
+      "grad_norm": 0.49370094007294774,
+      "kl": 0.375,
+      "learning_rate": 1.1250804867529794e-06,
+      "loss": 0.17,
+      "step": 391,
+      "step_loss": 0.0888671875
+    },
+    {
+      "epoch": 3.8058252427184467,
+      "grad_norm": 0.5639925863204843,
+      "kl": 0.251953125,
+      "learning_rate": 1.1155244584634953e-06,
+      "loss": 0.1395,
+      "step": 392,
+      "step_loss": 0.08984375
+    },
+    {
+      "epoch": 3.8155339805825244,
+      "grad_norm": 0.5187019670608374,
+      "kl": 0.173828125,
+      "learning_rate": 1.1060304508429407e-06,
+      "loss": 0.1556,
+      "step": 393,
+      "step_loss": 0.154296875
+    },
+    {
+      "epoch": 3.825242718446602,
+      "grad_norm": 0.4683311180263721,
+      "kl": 0.23046875,
+      "learning_rate": 1.0965988241443043e-06,
+      "loss": 0.1584,
+      "step": 394,
+      "step_loss": 0.1357421875
+    },
+    {
+      "epoch": 3.8349514563106797,
+      "grad_norm": 0.46054186247992845,
+      "kl": 0.2412109375,
+      "learning_rate": 1.0872299362535175e-06,
+      "loss": 0.1232,
+      "step": 395,
+      "step_loss": 0.0966796875
+    },
+    {
+      "epoch": 3.8446601941747574,
+      "grad_norm": 0.514740300530401,
+      "kl": 0.404296875,
+      "learning_rate": 1.0779241426758628e-06,
+      "loss": 0.1634,
+      "step": 396,
+      "step_loss": 0.09912109375
+    },
+    {
+      "epoch": 3.854368932038835,
+      "grad_norm": 0.5249887226800867,
+      "kl": 0.2421875,
+      "learning_rate": 1.068681796522496e-06,
+      "loss": 0.1515,
+      "step": 397,
+      "step_loss": 0.078125
+    },
+    {
+      "epoch": 3.8640776699029127,
+      "grad_norm": 0.4882053926671313,
+      "kl": 0.2734375,
+      "learning_rate": 1.0595032484970354e-06,
+      "loss": 0.1623,
+      "step": 398,
+      "step_loss": 0.06787109375
+    },
+    {
+      "epoch": 3.8737864077669903,
+      "grad_norm": 0.5046889857220114,
+      "kl": 0.1982421875,
+      "learning_rate": 1.0503888468822648e-06,
+      "loss": 0.1485,
+      "step": 399,
+      "step_loss": 0.2275390625
+    },
+    {
+      "epoch": 3.883495145631068,
+      "grad_norm": 0.5742122623829672,
+      "kl": 0.291015625,
+      "learning_rate": 1.0413389375269098e-06,
+      "loss": 0.1645,
+      "step": 400,
+      "step_loss": 0.109375
+    },
+    {
+      "epoch": 3.8932038834951457,
+      "grad_norm": 0.4869305965525205,
+      "kl": 0.373046875,
+      "learning_rate": 1.0323538638325185e-06,
+      "loss": 0.1506,
+      "step": 401,
+      "step_loss": 0.06298828125
+    },
+    {
+      "epoch": 3.9029126213592233,
+      "grad_norm": 0.4895770160456859,
+      "kl": 0.1748046875,
+      "learning_rate": 1.0234339667404326e-06,
+      "loss": 0.1535,
+      "step": 402,
+      "step_loss": 0.12255859375
+    },
+    {
+      "epoch": 3.912621359223301,
+      "grad_norm": 0.4718475674872005,
+      "kl": 0.2138671875,
+      "learning_rate": 1.0145795847188435e-06,
+      "loss": 0.1563,
+      "step": 403,
+      "step_loss": 0.244140625
+    },
+    {
+      "epoch": 3.9223300970873787,
+      "grad_norm": 0.46310640287447863,
+      "kl": 0.28125,
+      "learning_rate": 1.0057910537499585e-06,
+      "loss": 0.1502,
+      "step": 404,
+      "step_loss": 0.08642578125
+    },
+    {
+      "epoch": 3.9320388349514563,
+      "grad_norm": 0.5464878892001633,
+      "kl": 0.2275390625,
+      "learning_rate": 9.970687073172416e-07,
+      "loss": 0.1648,
+      "step": 405,
+      "step_loss": 0.1259765625
+    },
+    {
+      "epoch": 3.941747572815534,
+      "grad_norm": 0.5101772415025155,
+      "kl": 0.1318359375,
+      "learning_rate": 9.884128763927692e-07,
+      "loss": 0.1584,
+      "step": 406,
+      "step_loss": 0.0927734375
+    },
+    {
+      "epoch": 3.9514563106796117,
+      "grad_norm": 0.5262085741097766,
+      "kl": 0.3359375,
+      "learning_rate": 9.798238894246628e-07,
+      "loss": 0.1498,
+      "step": 407,
+      "step_loss": 0.134765625
+    },
+    {
+      "epoch": 3.9611650485436893,
+      "grad_norm": 0.5133055914879928,
+      "kl": 0.16015625,
+      "learning_rate": 9.713020723246332e-07,
+      "loss": 0.1707,
+      "step": 408,
+      "step_loss": 0.1513671875
+    },
+    {
+      "epoch": 3.970873786407767,
+      "grad_norm": 0.5499931811863684,
+      "kl": 0.306640625,
+      "learning_rate": 9.628477484556066e-07,
+      "loss": 0.1547,
+      "step": 409,
+      "step_loss": 0.11474609375
+    },
+    {
+      "epoch": 3.9805825242718447,
+      "grad_norm": 0.5051010516067944,
+      "kl": 0.162109375,
+      "learning_rate": 9.54461238619462e-07,
+      "loss": 0.1586,
+      "step": 410,
+      "step_loss": 0.109375
+    },
+    {
+      "epoch": 3.9902912621359223,
+      "grad_norm": 0.49187747334266807,
+      "kl": 0.2333984375,
+      "learning_rate": 9.461428610448503e-07,
+      "loss": 0.1461,
+      "step": 411,
+      "step_loss": 0.1416015625
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.44028016997121006,
+      "kl": 0.2734375,
+      "learning_rate": 9.378929313751267e-07,
+      "loss": 0.1499,
+      "step": 412,
+      "step_loss": 0.08056640625
+    },
+    {
+      "epoch": 4.0,
+      "eval_test_transformed.json_loss": NaN,
+      "eval_test_transformed.json_runtime": 8.4799,
+      "eval_test_transformed.json_samples_per_second": 58.963,
+      "eval_test_transformed.json_steps_per_second": 1.887,
+      "step": 412
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 515,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 50.0,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 43413797257216.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79e5b4b10998829a8a29dbe2f5bc6a735273c77a3141e21261948a43a478b72a
+size 9016

zero_to_fp32.py

@@ -0,0 +1,587 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel()
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file, tag=args.tag)