Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

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 +1,29 @@
+{
+  "_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.48.2",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.48.2"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd16e8f2a45f46cc98c3bb79773ffdc1b7db4957a72548a2447ff785d09e4d47
+size 4877660776

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a50fc94ea48cfc71500565412206c0615e19406ad51b0f4f92dbfeb13ff9029f
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff1e641ea5e5a4cb88f77ed43f7be8ce187ba61a33211607f006227849e7aa9d
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6dc24147239bc8d160fa1b5610874033d0af1e360aaf43d916a194f8faeedef
+size 1089994880

model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15231233024
+  },
+  "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"
+  }
+}

special_tokens_map.json

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

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 20000,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,3624 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.97265625,
+  "eval_steps": 500,
+  "global_step": 513,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01171875,
+      "grad_norm": 0.9319692850112915,
+      "learning_rate": 1.1764705882352942e-07,
+      "loss": 0.7028,
+      "step": 1
+    },
+    {
+      "epoch": 0.0234375,
+      "grad_norm": 0.9757155179977417,
+      "learning_rate": 2.3529411764705883e-07,
+      "loss": 0.7416,
+      "step": 2
+    },
+    {
+      "epoch": 0.03515625,
+      "grad_norm": 1.0889487266540527,
+      "learning_rate": 3.529411764705883e-07,
+      "loss": 0.8392,
+      "step": 3
+    },
+    {
+      "epoch": 0.046875,
+      "grad_norm": 1.0020272731781006,
+      "learning_rate": 4.7058823529411767e-07,
+      "loss": 0.7549,
+      "step": 4
+    },
+    {
+      "epoch": 0.05859375,
+      "grad_norm": 1.0064201354980469,
+      "learning_rate": 5.882352941176471e-07,
+      "loss": 0.802,
+      "step": 5
+    },
+    {
+      "epoch": 0.0703125,
+      "grad_norm": 0.9806166291236877,
+      "learning_rate": 7.058823529411766e-07,
+      "loss": 0.7754,
+      "step": 6
+    },
+    {
+      "epoch": 0.08203125,
+      "grad_norm": 0.9506519436836243,
+      "learning_rate": 8.235294117647059e-07,
+      "loss": 0.7591,
+      "step": 7
+    },
+    {
+      "epoch": 0.09375,
+      "grad_norm": 0.9138185977935791,
+      "learning_rate": 9.411764705882353e-07,
+      "loss": 0.7681,
+      "step": 8
+    },
+    {
+      "epoch": 0.10546875,
+      "grad_norm": 0.872790515422821,
+      "learning_rate": 1.0588235294117648e-06,
+      "loss": 0.7332,
+      "step": 9
+    },
+    {
+      "epoch": 0.1171875,
+      "grad_norm": 0.8308555483818054,
+      "learning_rate": 1.1764705882352942e-06,
+      "loss": 0.7991,
+      "step": 10
+    },
+    {
+      "epoch": 0.12890625,
+      "grad_norm": 0.7250374555587769,
+      "learning_rate": 1.2941176470588237e-06,
+      "loss": 0.6812,
+      "step": 11
+    },
+    {
+      "epoch": 0.140625,
+      "grad_norm": 0.6780915856361389,
+      "learning_rate": 1.4117647058823531e-06,
+      "loss": 0.7178,
+      "step": 12
+    },
+    {
+      "epoch": 0.15234375,
+      "grad_norm": 0.6684752702713013,
+      "learning_rate": 1.5294117647058826e-06,
+      "loss": 0.7335,
+      "step": 13
+    },
+    {
+      "epoch": 0.1640625,
+      "grad_norm": 0.48017001152038574,
+      "learning_rate": 1.6470588235294118e-06,
+      "loss": 0.7293,
+      "step": 14
+    },
+    {
+      "epoch": 0.17578125,
+      "grad_norm": 0.4176006317138672,
+      "learning_rate": 1.7647058823529414e-06,
+      "loss": 0.6453,
+      "step": 15
+    },
+    {
+      "epoch": 0.1875,
+      "grad_norm": 0.4581579566001892,
+      "learning_rate": 1.8823529411764707e-06,
+      "loss": 0.709,
+      "step": 16
+    },
+    {
+      "epoch": 0.19921875,
+      "grad_norm": 0.41040804982185364,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.6965,
+      "step": 17
+    },
+    {
+      "epoch": 0.2109375,
+      "grad_norm": 0.38801896572113037,
+      "learning_rate": 2.1176470588235296e-06,
+      "loss": 0.7142,
+      "step": 18
+    },
+    {
+      "epoch": 0.22265625,
+      "grad_norm": 0.5587085485458374,
+      "learning_rate": 2.2352941176470592e-06,
+      "loss": 0.6624,
+      "step": 19
+    },
+    {
+      "epoch": 0.234375,
+      "grad_norm": 0.6821652054786682,
+      "learning_rate": 2.3529411764705885e-06,
+      "loss": 0.6708,
+      "step": 20
+    },
+    {
+      "epoch": 0.24609375,
+      "grad_norm": 0.7130780816078186,
+      "learning_rate": 2.470588235294118e-06,
+      "loss": 0.68,
+      "step": 21
+    },
+    {
+      "epoch": 0.2578125,
+      "grad_norm": 0.6799156069755554,
+      "learning_rate": 2.5882352941176473e-06,
+      "loss": 0.6069,
+      "step": 22
+    },
+    {
+      "epoch": 0.26953125,
+      "grad_norm": 0.6155521869659424,
+      "learning_rate": 2.7058823529411766e-06,
+      "loss": 0.6541,
+      "step": 23
+    },
+    {
+      "epoch": 0.28125,
+      "grad_norm": 0.5290796160697937,
+      "learning_rate": 2.8235294117647062e-06,
+      "loss": 0.6276,
+      "step": 24
+    },
+    {
+      "epoch": 0.29296875,
+      "grad_norm": 0.460702121257782,
+      "learning_rate": 2.9411764705882355e-06,
+      "loss": 0.6172,
+      "step": 25
+    },
+    {
+      "epoch": 0.3046875,
+      "grad_norm": 0.4080013334751129,
+      "learning_rate": 3.058823529411765e-06,
+      "loss": 0.609,
+      "step": 26
+    },
+    {
+      "epoch": 0.31640625,
+      "grad_norm": 0.32690924406051636,
+      "learning_rate": 3.1764705882352943e-06,
+      "loss": 0.585,
+      "step": 27
+    },
+    {
+      "epoch": 0.328125,
+      "grad_norm": 0.3010313808917999,
+      "learning_rate": 3.2941176470588236e-06,
+      "loss": 0.5684,
+      "step": 28
+    },
+    {
+      "epoch": 0.33984375,
+      "grad_norm": 0.30681392550468445,
+      "learning_rate": 3.4117647058823532e-06,
+      "loss": 0.6241,
+      "step": 29
+    },
+    {
+      "epoch": 0.3515625,
+      "grad_norm": 0.2821861505508423,
+      "learning_rate": 3.529411764705883e-06,
+      "loss": 0.5612,
+      "step": 30
+    },
+    {
+      "epoch": 0.36328125,
+      "grad_norm": 0.3114885985851288,
+      "learning_rate": 3.6470588235294117e-06,
+      "loss": 0.5777,
+      "step": 31
+    },
+    {
+      "epoch": 0.375,
+      "grad_norm": 0.2961554527282715,
+      "learning_rate": 3.7647058823529414e-06,
+      "loss": 0.6118,
+      "step": 32
+    },
+    {
+      "epoch": 0.38671875,
+      "grad_norm": 0.2794322371482849,
+      "learning_rate": 3.882352941176471e-06,
+      "loss": 0.5221,
+      "step": 33
+    },
+    {
+      "epoch": 0.3984375,
+      "grad_norm": 0.33694833517074585,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.5857,
+      "step": 34
+    },
+    {
+      "epoch": 0.41015625,
+      "grad_norm": 0.2691885530948639,
+      "learning_rate": 4.11764705882353e-06,
+      "loss": 0.6122,
+      "step": 35
+    },
+    {
+      "epoch": 0.421875,
+      "grad_norm": 0.2636789381504059,
+      "learning_rate": 4.235294117647059e-06,
+      "loss": 0.5355,
+      "step": 36
+    },
+    {
+      "epoch": 0.43359375,
+      "grad_norm": 0.25277385115623474,
+      "learning_rate": 4.352941176470588e-06,
+      "loss": 0.5373,
+      "step": 37
+    },
+    {
+      "epoch": 0.4453125,
+      "grad_norm": 0.2117689847946167,
+      "learning_rate": 4.4705882352941184e-06,
+      "loss": 0.5308,
+      "step": 38
+    },
+    {
+      "epoch": 0.45703125,
+      "grad_norm": 0.19506287574768066,
+      "learning_rate": 4.588235294117647e-06,
+      "loss": 0.5372,
+      "step": 39
+    },
+    {
+      "epoch": 0.46875,
+      "grad_norm": 0.21475745737552643,
+      "learning_rate": 4.705882352941177e-06,
+      "loss": 0.565,
+      "step": 40
+    },
+    {
+      "epoch": 0.48046875,
+      "grad_norm": 0.20359668135643005,
+      "learning_rate": 4.823529411764706e-06,
+      "loss": 0.5349,
+      "step": 41
+    },
+    {
+      "epoch": 0.4921875,
+      "grad_norm": 0.2028248906135559,
+      "learning_rate": 4.941176470588236e-06,
+      "loss": 0.5587,
+      "step": 42
+    },
+    {
+      "epoch": 0.50390625,
+      "grad_norm": 0.21315933763980865,
+      "learning_rate": 5.058823529411765e-06,
+      "loss": 0.542,
+      "step": 43
+    },
+    {
+      "epoch": 0.515625,
+      "grad_norm": 0.21315297484397888,
+      "learning_rate": 5.176470588235295e-06,
+      "loss": 0.5878,
+      "step": 44
+    },
+    {
+      "epoch": 0.52734375,
+      "grad_norm": 0.19848909974098206,
+      "learning_rate": 5.294117647058824e-06,
+      "loss": 0.5238,
+      "step": 45
+    },
+    {
+      "epoch": 0.5390625,
+      "grad_norm": 0.20311777293682098,
+      "learning_rate": 5.411764705882353e-06,
+      "loss": 0.548,
+      "step": 46
+    },
+    {
+      "epoch": 0.55078125,
+      "grad_norm": 0.18243664503097534,
+      "learning_rate": 5.529411764705883e-06,
+      "loss": 0.5282,
+      "step": 47
+    },
+    {
+      "epoch": 0.5625,
+      "grad_norm": 0.16974785923957825,
+      "learning_rate": 5.6470588235294125e-06,
+      "loss": 0.4836,
+      "step": 48
+    },
+    {
+      "epoch": 0.57421875,
+      "grad_norm": 0.1762179285287857,
+      "learning_rate": 5.764705882352941e-06,
+      "loss": 0.5304,
+      "step": 49
+    },
+    {
+      "epoch": 0.5859375,
+      "grad_norm": 0.20087581872940063,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 0.5317,
+      "step": 50
+    },
+    {
+      "epoch": 0.59765625,
+      "grad_norm": 0.1862439662218094,
+      "learning_rate": 6e-06,
+      "loss": 0.5536,
+      "step": 51
+    },
+    {
+      "epoch": 0.609375,
+      "grad_norm": 0.1928662210702896,
+      "learning_rate": 6.11764705882353e-06,
+      "loss": 0.4844,
+      "step": 52
+    },
+    {
+      "epoch": 0.62109375,
+      "grad_norm": 0.20256248116493225,
+      "learning_rate": 6.2352941176470595e-06,
+      "loss": 0.5428,
+      "step": 53
+    },
+    {
+      "epoch": 0.6328125,
+      "grad_norm": 0.2007371336221695,
+      "learning_rate": 6.352941176470589e-06,
+      "loss": 0.5225,
+      "step": 54
+    },
+    {
+      "epoch": 0.64453125,
+      "grad_norm": 0.16175827383995056,
+      "learning_rate": 6.470588235294119e-06,
+      "loss": 0.5137,
+      "step": 55
+    },
+    {
+      "epoch": 0.65625,
+      "grad_norm": 0.17586955428123474,
+      "learning_rate": 6.588235294117647e-06,
+      "loss": 0.5034,
+      "step": 56
+    },
+    {
+      "epoch": 0.66796875,
+      "grad_norm": 0.17159290611743927,
+      "learning_rate": 6.705882352941176e-06,
+      "loss": 0.5267,
+      "step": 57
+    },
+    {
+      "epoch": 0.6796875,
+      "grad_norm": 0.17129066586494446,
+      "learning_rate": 6.8235294117647065e-06,
+      "loss": 0.4634,
+      "step": 58
+    },
+    {
+      "epoch": 0.69140625,
+      "grad_norm": 0.14943340420722961,
+      "learning_rate": 6.941176470588236e-06,
+      "loss": 0.4749,
+      "step": 59
+    },
+    {
+      "epoch": 0.703125,
+      "grad_norm": 0.17984403669834137,
+      "learning_rate": 7.058823529411766e-06,
+      "loss": 0.5254,
+      "step": 60
+    },
+    {
+      "epoch": 0.71484375,
+      "grad_norm": 0.15619614720344543,
+      "learning_rate": 7.176470588235295e-06,
+      "loss": 0.4941,
+      "step": 61
+    },
+    {
+      "epoch": 0.7265625,
+      "grad_norm": 0.14417926967144012,
+      "learning_rate": 7.294117647058823e-06,
+      "loss": 0.4234,
+      "step": 62
+    },
+    {
+      "epoch": 0.73828125,
+      "grad_norm": 0.16936203837394714,
+      "learning_rate": 7.4117647058823535e-06,
+      "loss": 0.5016,
+      "step": 63
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.17044682800769806,
+      "learning_rate": 7.529411764705883e-06,
+      "loss": 0.4817,
+      "step": 64
+    },
+    {
+      "epoch": 0.76171875,
+      "grad_norm": 0.1539342850446701,
+      "learning_rate": 7.647058823529411e-06,
+      "loss": 0.4804,
+      "step": 65
+    },
+    {
+      "epoch": 0.7734375,
+      "grad_norm": 0.15344035625457764,
+      "learning_rate": 7.764705882352941e-06,
+      "loss": 0.4875,
+      "step": 66
+    },
+    {
+      "epoch": 0.78515625,
+      "grad_norm": 0.15734320878982544,
+      "learning_rate": 7.882352941176471e-06,
+      "loss": 0.436,
+      "step": 67
+    },
+    {
+      "epoch": 0.796875,
+      "grad_norm": 0.17874149978160858,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.5315,
+      "step": 68
+    },
+    {
+      "epoch": 0.80859375,
+      "grad_norm": 0.18438568711280823,
+      "learning_rate": 8.11764705882353e-06,
+      "loss": 0.5206,
+      "step": 69
+    },
+    {
+      "epoch": 0.8203125,
+      "grad_norm": 0.16773667931556702,
+      "learning_rate": 8.23529411764706e-06,
+      "loss": 0.4604,
+      "step": 70
+    },
+    {
+      "epoch": 0.83203125,
+      "grad_norm": 0.146653413772583,
+      "learning_rate": 8.35294117647059e-06,
+      "loss": 0.4445,
+      "step": 71
+    },
+    {
+      "epoch": 0.84375,
+      "grad_norm": 0.1633101850748062,
+      "learning_rate": 8.470588235294118e-06,
+      "loss": 0.4741,
+      "step": 72
+    },
+    {
+      "epoch": 0.85546875,
+      "grad_norm": 0.1426013708114624,
+      "learning_rate": 8.588235294117647e-06,
+      "loss": 0.4447,
+      "step": 73
+    },
+    {
+      "epoch": 0.8671875,
+      "grad_norm": 0.19708958268165588,
+      "learning_rate": 8.705882352941177e-06,
+      "loss": 0.4915,
+      "step": 74
+    },
+    {
+      "epoch": 0.87890625,
+      "grad_norm": 0.13479125499725342,
+      "learning_rate": 8.823529411764707e-06,
+      "loss": 0.4302,
+      "step": 75
+    },
+    {
+      "epoch": 0.890625,
+      "grad_norm": 0.16112269461154938,
+      "learning_rate": 8.941176470588237e-06,
+      "loss": 0.4457,
+      "step": 76
+    },
+    {
+      "epoch": 0.90234375,
+      "grad_norm": 0.1683078110218048,
+      "learning_rate": 9.058823529411765e-06,
+      "loss": 0.4617,
+      "step": 77
+    },
+    {
+      "epoch": 0.9140625,
+      "grad_norm": 0.17289473116397858,
+      "learning_rate": 9.176470588235294e-06,
+      "loss": 0.4885,
+      "step": 78
+    },
+    {
+      "epoch": 0.92578125,
+      "grad_norm": 0.14186030626296997,
+      "learning_rate": 9.294117647058824e-06,
+      "loss": 0.4063,
+      "step": 79
+    },
+    {
+      "epoch": 0.9375,
+      "grad_norm": 0.17006702721118927,
+      "learning_rate": 9.411764705882354e-06,
+      "loss": 0.4684,
+      "step": 80
+    },
+    {
+      "epoch": 0.94921875,
+      "grad_norm": 0.16714246571063995,
+      "learning_rate": 9.529411764705882e-06,
+      "loss": 0.4416,
+      "step": 81
+    },
+    {
+      "epoch": 0.9609375,
+      "grad_norm": 0.1373886913061142,
+      "learning_rate": 9.647058823529412e-06,
+      "loss": 0.4677,
+      "step": 82
+    },
+    {
+      "epoch": 0.97265625,
+      "grad_norm": 0.15918415784835815,
+      "learning_rate": 9.764705882352942e-06,
+      "loss": 0.4479,
+      "step": 83
+    },
+    {
+      "epoch": 0.984375,
+      "grad_norm": 0.12358763813972473,
+      "learning_rate": 9.882352941176472e-06,
+      "loss": 0.4033,
+      "step": 84
+    },
+    {
+      "epoch": 0.99609375,
+      "grad_norm": 0.19624371826648712,
+      "learning_rate": 1e-05,
+      "loss": 0.4579,
+      "step": 85
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.19624371826648712,
+      "learning_rate": 9.99995783847866e-06,
+      "loss": 0.4138,
+      "step": 86
+    },
+    {
+      "epoch": 1.01171875,
+      "grad_norm": 0.2939195930957794,
+      "learning_rate": 9.999831354625678e-06,
+      "loss": 0.3971,
+      "step": 87
+    },
+    {
+      "epoch": 1.0234375,
+      "grad_norm": 0.14151132106781006,
+      "learning_rate": 9.999620550574155e-06,
+      "loss": 0.4168,
+      "step": 88
+    },
+    {
+      "epoch": 1.03515625,
+      "grad_norm": 0.174483984708786,
+      "learning_rate": 9.999325429879215e-06,
+      "loss": 0.3952,
+      "step": 89
+    },
+    {
+      "epoch": 1.046875,
+      "grad_norm": 0.16392698884010315,
+      "learning_rate": 9.998945997517957e-06,
+      "loss": 0.4031,
+      "step": 90
+    },
+    {
+      "epoch": 1.05859375,
+      "grad_norm": 0.13462132215499878,
+      "learning_rate": 9.99848225988936e-06,
+      "loss": 0.3974,
+      "step": 91
+    },
+    {
+      "epoch": 1.0703125,
+      "grad_norm": 0.20017389953136444,
+      "learning_rate": 9.997934224814173e-06,
+      "loss": 0.4054,
+      "step": 92
+    },
+    {
+      "epoch": 1.08203125,
+      "grad_norm": 0.15260270237922668,
+      "learning_rate": 9.997301901534797e-06,
+      "loss": 0.43,
+      "step": 93
+    },
+    {
+      "epoch": 1.09375,
+      "grad_norm": 0.15193504095077515,
+      "learning_rate": 9.996585300715117e-06,
+      "loss": 0.3885,
+      "step": 94
+    },
+    {
+      "epoch": 1.10546875,
+      "grad_norm": 0.16993655264377594,
+      "learning_rate": 9.99578443444032e-06,
+      "loss": 0.4191,
+      "step": 95
+    },
+    {
+      "epoch": 1.1171875,
+      "grad_norm": 0.15632706880569458,
+      "learning_rate": 9.994899316216709e-06,
+      "loss": 0.3439,
+      "step": 96
+    },
+    {
+      "epoch": 1.12890625,
+      "grad_norm": 0.1436368227005005,
+      "learning_rate": 9.99392996097145e-06,
+      "loss": 0.3609,
+      "step": 97
+    },
+    {
+      "epoch": 1.140625,
+      "grad_norm": 0.14202651381492615,
+      "learning_rate": 9.992876385052346e-06,
+      "loss": 0.4133,
+      "step": 98
+    },
+    {
+      "epoch": 1.15234375,
+      "grad_norm": 0.17068006098270416,
+      "learning_rate": 9.991738606227537e-06,
+      "loss": 0.3775,
+      "step": 99
+    },
+    {
+      "epoch": 1.1640625,
+      "grad_norm": 0.15222977101802826,
+      "learning_rate": 9.990516643685222e-06,
+      "loss": 0.4191,
+      "step": 100
+    },
+    {
+      "epoch": 1.17578125,
+      "grad_norm": 0.1581757664680481,
+      "learning_rate": 9.989210518033316e-06,
+      "loss": 0.4098,
+      "step": 101
+    },
+    {
+      "epoch": 1.1875,
+      "grad_norm": 0.14198894798755646,
+      "learning_rate": 9.987820251299121e-06,
+      "loss": 0.4031,
+      "step": 102
+    },
+    {
+      "epoch": 1.19921875,
+      "grad_norm": 0.1587005853652954,
+      "learning_rate": 9.98634586692894e-06,
+      "loss": 0.3632,
+      "step": 103
+    },
+    {
+      "epoch": 1.2109375,
+      "grad_norm": 0.1380324512720108,
+      "learning_rate": 9.984787389787689e-06,
+      "loss": 0.4229,
+      "step": 104
+    },
+    {
+      "epoch": 1.22265625,
+      "grad_norm": 0.1664944589138031,
+      "learning_rate": 9.983144846158472e-06,
+      "loss": 0.3952,
+      "step": 105
+    },
+    {
+      "epoch": 1.234375,
+      "grad_norm": 0.15593458712100983,
+      "learning_rate": 9.981418263742148e-06,
+      "loss": 0.3657,
+      "step": 106
+    },
+    {
+      "epoch": 1.24609375,
+      "grad_norm": 0.14692318439483643,
+      "learning_rate": 9.979607671656852e-06,
+      "loss": 0.3777,
+      "step": 107
+    },
+    {
+      "epoch": 1.2578125,
+      "grad_norm": 0.19152411818504333,
+      "learning_rate": 9.97771310043751e-06,
+      "loss": 0.378,
+      "step": 108
+    },
+    {
+      "epoch": 1.26953125,
+      "grad_norm": 0.1495964378118515,
+      "learning_rate": 9.975734582035323e-06,
+      "loss": 0.3662,
+      "step": 109
+    },
+    {
+      "epoch": 1.28125,
+      "grad_norm": 0.14938651025295258,
+      "learning_rate": 9.973672149817232e-06,
+      "loss": 0.4197,
+      "step": 110
+    },
+    {
+      "epoch": 1.29296875,
+      "grad_norm": 0.15581446886062622,
+      "learning_rate": 9.971525838565348e-06,
+      "loss": 0.3804,
+      "step": 111
+    },
+    {
+      "epoch": 1.3046875,
+      "grad_norm": 0.14621272683143616,
+      "learning_rate": 9.96929568447637e-06,
+      "loss": 0.3372,
+      "step": 112
+    },
+    {
+      "epoch": 1.31640625,
+      "grad_norm": 0.14326448738574982,
+      "learning_rate": 9.966981725160972e-06,
+      "loss": 0.4092,
+      "step": 113
+    },
+    {
+      "epoch": 1.328125,
+      "grad_norm": 0.1629864126443863,
+      "learning_rate": 9.964583999643174e-06,
+      "loss": 0.3829,
+      "step": 114
+    },
+    {
+      "epoch": 1.33984375,
+      "grad_norm": 0.16448885202407837,
+      "learning_rate": 9.96210254835968e-06,
+      "loss": 0.3952,
+      "step": 115
+    },
+    {
+      "epoch": 1.3515625,
+      "grad_norm": 0.13906875252723694,
+      "learning_rate": 9.95953741315919e-06,
+      "loss": 0.3501,
+      "step": 116
+    },
+    {
+      "epoch": 1.36328125,
+      "grad_norm": 0.1344955563545227,
+      "learning_rate": 9.95688863730171e-06,
+      "loss": 0.4212,
+      "step": 117
+    },
+    {
+      "epoch": 1.375,
+      "grad_norm": 0.2054166942834854,
+      "learning_rate": 9.954156265457801e-06,
+      "loss": 0.4155,
+      "step": 118
+    },
+    {
+      "epoch": 1.38671875,
+      "grad_norm": 0.14791074395179749,
+      "learning_rate": 9.951340343707852e-06,
+      "loss": 0.4104,
+      "step": 119
+    },
+    {
+      "epoch": 1.3984375,
+      "grad_norm": 0.17642416059970856,
+      "learning_rate": 9.948440919541277e-06,
+      "loss": 0.3502,
+      "step": 120
+    },
+    {
+      "epoch": 1.41015625,
+      "grad_norm": 0.14525847136974335,
+      "learning_rate": 9.945458041855732e-06,
+      "loss": 0.383,
+      "step": 121
+    },
+    {
+      "epoch": 1.421875,
+      "grad_norm": 0.15956953167915344,
+      "learning_rate": 9.942391760956277e-06,
+      "loss": 0.3864,
+      "step": 122
+    },
+    {
+      "epoch": 1.43359375,
+      "grad_norm": 0.16518352925777435,
+      "learning_rate": 9.939242128554542e-06,
+      "loss": 0.4374,
+      "step": 123
+    },
+    {
+      "epoch": 1.4453125,
+      "grad_norm": 0.15952655673027039,
+      "learning_rate": 9.936009197767847e-06,
+      "loss": 0.3719,
+      "step": 124
+    },
+    {
+      "epoch": 1.45703125,
+      "grad_norm": 0.13431760668754578,
+      "learning_rate": 9.932693023118299e-06,
+      "loss": 0.3693,
+      "step": 125
+    },
+    {
+      "epoch": 1.46875,
+      "grad_norm": 0.15348534286022186,
+      "learning_rate": 9.929293660531889e-06,
+      "loss": 0.3708,
+      "step": 126
+    },
+    {
+      "epoch": 1.48046875,
+      "grad_norm": 0.1592601090669632,
+      "learning_rate": 9.925811167337533e-06,
+      "loss": 0.3598,
+      "step": 127
+    },
+    {
+      "epoch": 1.4921875,
+      "grad_norm": 0.16625793278217316,
+      "learning_rate": 9.922245602266119e-06,
+      "loss": 0.3548,
+      "step": 128
+    },
+    {
+      "epoch": 1.50390625,
+      "grad_norm": 0.16050715744495392,
+      "learning_rate": 9.918597025449505e-06,
+      "loss": 0.3732,
+      "step": 129
+    },
+    {
+      "epoch": 1.515625,
+      "grad_norm": 0.1567108929157257,
+      "learning_rate": 9.91486549841951e-06,
+      "loss": 0.3875,
+      "step": 130
+    },
+    {
+      "epoch": 1.52734375,
+      "grad_norm": 0.13900204002857208,
+      "learning_rate": 9.911051084106877e-06,
+      "loss": 0.3134,
+      "step": 131
+    },
+    {
+      "epoch": 1.5390625,
+      "grad_norm": 0.17397968471050262,
+      "learning_rate": 9.90715384684021e-06,
+      "loss": 0.3283,
+      "step": 132
+    },
+    {
+      "epoch": 1.55078125,
+      "grad_norm": 0.13944688439369202,
+      "learning_rate": 9.903173852344889e-06,
+      "loss": 0.3801,
+      "step": 133
+    },
+    {
+      "epoch": 1.5625,
+      "grad_norm": 0.1763840615749359,
+      "learning_rate": 9.899111167741966e-06,
+      "loss": 0.3426,
+      "step": 134
+    },
+    {
+      "epoch": 1.57421875,
+      "grad_norm": 0.15428921580314636,
+      "learning_rate": 9.894965861547023e-06,
+      "loss": 0.4046,
+      "step": 135
+    },
+    {
+      "epoch": 1.5859375,
+      "grad_norm": 0.18792587518692017,
+      "learning_rate": 9.890738003669029e-06,
+      "loss": 0.3948,
+      "step": 136
+    },
+    {
+      "epoch": 1.59765625,
+      "grad_norm": 0.14246441423892975,
+      "learning_rate": 9.88642766540915e-06,
+      "loss": 0.343,
+      "step": 137
+    },
+    {
+      "epoch": 1.609375,
+      "grad_norm": 0.14254647493362427,
+      "learning_rate": 9.882034919459556e-06,
+      "loss": 0.3908,
+      "step": 138
+    },
+    {
+      "epoch": 1.62109375,
+      "grad_norm": 0.15610483288764954,
+      "learning_rate": 9.877559839902185e-06,
+      "loss": 0.4127,
+      "step": 139
+    },
+    {
+      "epoch": 1.6328125,
+      "grad_norm": 0.14824527502059937,
+      "learning_rate": 9.873002502207502e-06,
+      "loss": 0.3469,
+      "step": 140
+    },
+    {
+      "epoch": 1.64453125,
+      "grad_norm": 0.13025063276290894,
+      "learning_rate": 9.868362983233226e-06,
+      "loss": 0.3467,
+      "step": 141
+    },
+    {
+      "epoch": 1.65625,
+      "grad_norm": 0.12863893806934357,
+      "learning_rate": 9.863641361223025e-06,
+      "loss": 0.3441,
+      "step": 142
+    },
+    {
+      "epoch": 1.66796875,
+      "grad_norm": 0.1754492223262787,
+      "learning_rate": 9.858837715805207e-06,
+      "loss": 0.3893,
+      "step": 143
+    },
+    {
+      "epoch": 1.6796875,
+      "grad_norm": 0.14538809657096863,
+      "learning_rate": 9.853952127991374e-06,
+      "loss": 0.361,
+      "step": 144
+    },
+    {
+      "epoch": 1.69140625,
+      "grad_norm": 0.13839296996593475,
+      "learning_rate": 9.848984680175049e-06,
+      "loss": 0.4422,
+      "step": 145
+    },
+    {
+      "epoch": 1.703125,
+      "grad_norm": 0.16836000978946686,
+      "learning_rate": 9.843935456130295e-06,
+      "loss": 0.3741,
+      "step": 146
+    },
+    {
+      "epoch": 1.71484375,
+      "grad_norm": 0.1621960699558258,
+      "learning_rate": 9.8388045410103e-06,
+      "loss": 0.4299,
+      "step": 147
+    },
+    {
+      "epoch": 1.7265625,
+      "grad_norm": 0.15016399323940277,
+      "learning_rate": 9.833592021345938e-06,
+      "loss": 0.3613,
+      "step": 148
+    },
+    {
+      "epoch": 1.73828125,
+      "grad_norm": 0.1778838336467743,
+      "learning_rate": 9.828297985044314e-06,
+      "loss": 0.4127,
+      "step": 149
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 0.20492997765541077,
+      "learning_rate": 9.822922521387277e-06,
+      "loss": 0.4566,
+      "step": 150
+    },
+    {
+      "epoch": 1.76171875,
+      "grad_norm": 0.142822265625,
+      "learning_rate": 9.817465721029916e-06,
+      "loss": 0.3561,
+      "step": 151
+    },
+    {
+      "epoch": 1.7734375,
+      "grad_norm": 0.15397801995277405,
+      "learning_rate": 9.811927675999035e-06,
+      "loss": 0.3854,
+      "step": 152
+    },
+    {
+      "epoch": 1.78515625,
+      "grad_norm": 0.13618412613868713,
+      "learning_rate": 9.806308479691595e-06,
+      "loss": 0.3933,
+      "step": 153
+    },
+    {
+      "epoch": 1.796875,
+      "grad_norm": 0.19063451886177063,
+      "learning_rate": 9.800608226873143e-06,
+      "loss": 0.382,
+      "step": 154
+    },
+    {
+      "epoch": 1.80859375,
+      "grad_norm": 0.16610917448997498,
+      "learning_rate": 9.794827013676206e-06,
+      "loss": 0.402,
+      "step": 155
+    },
+    {
+      "epoch": 1.8203125,
+      "grad_norm": 0.14735649526119232,
+      "learning_rate": 9.788964937598688e-06,
+      "loss": 0.3646,
+      "step": 156
+    },
+    {
+      "epoch": 1.83203125,
+      "grad_norm": 0.1583123356103897,
+      "learning_rate": 9.783022097502204e-06,
+      "loss": 0.3519,
+      "step": 157
+    },
+    {
+      "epoch": 1.84375,
+      "grad_norm": 0.14677587151527405,
+      "learning_rate": 9.776998593610428e-06,
+      "loss": 0.3739,
+      "step": 158
+    },
+    {
+      "epoch": 1.85546875,
+      "grad_norm": 0.15498070418834686,
+      "learning_rate": 9.770894527507393e-06,
+      "loss": 0.344,
+      "step": 159
+    },
+    {
+      "epoch": 1.8671875,
+      "grad_norm": 0.164178267121315,
+      "learning_rate": 9.764710002135784e-06,
+      "loss": 0.3701,
+      "step": 160
+    },
+    {
+      "epoch": 1.87890625,
+      "grad_norm": 0.12638989090919495,
+      "learning_rate": 9.7584451217952e-06,
+      "loss": 0.3985,
+      "step": 161
+    },
+    {
+      "epoch": 1.890625,
+      "grad_norm": 0.17551939189434052,
+      "learning_rate": 9.752099992140401e-06,
+      "loss": 0.367,
+      "step": 162
+    },
+    {
+      "epoch": 1.90234375,
+      "grad_norm": 0.13494940102100372,
+      "learning_rate": 9.745674720179507e-06,
+      "loss": 0.345,
+      "step": 163
+    },
+    {
+      "epoch": 1.9140625,
+      "grad_norm": 0.15139806270599365,
+      "learning_rate": 9.739169414272219e-06,
+      "loss": 0.3767,
+      "step": 164
+    },
+    {
+      "epoch": 1.92578125,
+      "grad_norm": 0.16584157943725586,
+      "learning_rate": 9.732584184127973e-06,
+      "loss": 0.4002,
+      "step": 165
+    },
+    {
+      "epoch": 1.9375,
+      "grad_norm": 0.14180031418800354,
+      "learning_rate": 9.7259191408041e-06,
+      "loss": 0.3494,
+      "step": 166
+    },
+    {
+      "epoch": 1.94921875,
+      "grad_norm": 0.15869130194187164,
+      "learning_rate": 9.719174396703941e-06,
+      "loss": 0.3527,
+      "step": 167
+    },
+    {
+      "epoch": 1.9609375,
+      "grad_norm": 0.15733729302883148,
+      "learning_rate": 9.71235006557497e-06,
+      "loss": 0.4277,
+      "step": 168
+    },
+    {
+      "epoch": 1.97265625,
+      "grad_norm": 0.16619561612606049,
+      "learning_rate": 9.705446262506858e-06,
+      "loss": 0.398,
+      "step": 169
+    },
+    {
+      "epoch": 1.984375,
+      "grad_norm": 0.156788632273674,
+      "learning_rate": 9.698463103929542e-06,
+      "loss": 0.4272,
+      "step": 170
+    },
+    {
+      "epoch": 1.99609375,
+      "grad_norm": 0.17090734839439392,
+      "learning_rate": 9.691400707611258e-06,
+      "loss": 0.3683,
+      "step": 171
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.17090734839439392,
+      "learning_rate": 9.684259192656554e-06,
+      "loss": 0.3783,
+      "step": 172
+    },
+    {
+      "epoch": 2.01171875,
+      "grad_norm": 0.269039511680603,
+      "learning_rate": 9.677038679504285e-06,
+      "loss": 0.3033,
+      "step": 173
+    },
+    {
+      "epoch": 2.0234375,
+      "grad_norm": 0.18037453293800354,
+      "learning_rate": 9.669739289925578e-06,
+      "loss": 0.2647,
+      "step": 174
+    },
+    {
+      "epoch": 2.03515625,
+      "grad_norm": 0.16502933204174042,
+      "learning_rate": 9.66236114702178e-06,
+      "loss": 0.2689,
+      "step": 175
+    },
+    {
+      "epoch": 2.046875,
+      "grad_norm": 0.13461141288280487,
+      "learning_rate": 9.654904375222384e-06,
+      "loss": 0.299,
+      "step": 176
+    },
+    {
+      "epoch": 2.05859375,
+      "grad_norm": 0.15441769361495972,
+      "learning_rate": 9.647369100282928e-06,
+      "loss": 0.2987,
+      "step": 177
+    },
+    {
+      "epoch": 2.0703125,
+      "grad_norm": 0.22800779342651367,
+      "learning_rate": 9.639755449282874e-06,
+      "loss": 0.2801,
+      "step": 178
+    },
+    {
+      "epoch": 2.08203125,
+      "grad_norm": 0.1886795461177826,
+      "learning_rate": 9.632063550623465e-06,
+      "loss": 0.264,
+      "step": 179
+    },
+    {
+      "epoch": 2.09375,
+      "grad_norm": 0.16359379887580872,
+      "learning_rate": 9.62429353402556e-06,
+      "loss": 0.278,
+      "step": 180
+    },
+    {
+      "epoch": 2.10546875,
+      "grad_norm": 0.1788463145494461,
+      "learning_rate": 9.616445530527448e-06,
+      "loss": 0.2813,
+      "step": 181
+    },
+    {
+      "epoch": 2.1171875,
+      "grad_norm": 0.12653109431266785,
+      "learning_rate": 9.608519672482635e-06,
+      "loss": 0.2997,
+      "step": 182
+    },
+    {
+      "epoch": 2.12890625,
+      "grad_norm": 0.20340153574943542,
+      "learning_rate": 9.600516093557618e-06,
+      "loss": 0.3049,
+      "step": 183
+    },
+    {
+      "epoch": 2.140625,
+      "grad_norm": 0.1629520207643509,
+      "learning_rate": 9.592434928729617e-06,
+      "loss": 0.2765,
+      "step": 184
+    },
+    {
+      "epoch": 2.15234375,
+      "grad_norm": 0.17957361042499542,
+      "learning_rate": 9.584276314284316e-06,
+      "loss": 0.281,
+      "step": 185
+    },
+    {
+      "epoch": 2.1640625,
+      "grad_norm": 0.13229969143867493,
+      "learning_rate": 9.576040387813553e-06,
+      "loss": 0.2668,
+      "step": 186
+    },
+    {
+      "epoch": 2.17578125,
+      "grad_norm": 0.19403038918972015,
+      "learning_rate": 9.567727288213005e-06,
+      "loss": 0.2529,
+      "step": 187
+    },
+    {
+      "epoch": 2.1875,
+      "grad_norm": 0.1554100513458252,
+      "learning_rate": 9.559337155679843e-06,
+      "loss": 0.2756,
+      "step": 188
+    },
+    {
+      "epoch": 2.19921875,
+      "grad_norm": 0.1644996702671051,
+      "learning_rate": 9.550870131710366e-06,
+      "loss": 0.2765,
+      "step": 189
+    },
+    {
+      "epoch": 2.2109375,
+      "grad_norm": 0.16561852395534515,
+      "learning_rate": 9.542326359097619e-06,
+      "loss": 0.2834,
+      "step": 190
+    },
+    {
+      "epoch": 2.22265625,
+      "grad_norm": 0.15680573880672455,
+      "learning_rate": 9.533705981928984e-06,
+      "loss": 0.2611,
+      "step": 191
+    },
+    {
+      "epoch": 2.234375,
+      "grad_norm": 0.14902271330356598,
+      "learning_rate": 9.525009145583746e-06,
+      "loss": 0.2759,
+      "step": 192
+    },
+    {
+      "epoch": 2.24609375,
+      "grad_norm": 0.1453092247247696,
+      "learning_rate": 9.516235996730645e-06,
+      "loss": 0.2453,
+      "step": 193
+    },
+    {
+      "epoch": 2.2578125,
+      "grad_norm": 0.15511390566825867,
+      "learning_rate": 9.507386683325404e-06,
+      "loss": 0.2878,
+      "step": 194
+    },
+    {
+      "epoch": 2.26953125,
+      "grad_norm": 0.146243616938591,
+      "learning_rate": 9.498461354608228e-06,
+      "loss": 0.2813,
+      "step": 195
+    },
+    {
+      "epoch": 2.28125,
+      "grad_norm": 0.13351713120937347,
+      "learning_rate": 9.489460161101291e-06,
+      "loss": 0.2464,
+      "step": 196
+    },
+    {
+      "epoch": 2.29296875,
+      "grad_norm": 0.12981724739074707,
+      "learning_rate": 9.4803832546062e-06,
+      "loss": 0.2364,
+      "step": 197
+    },
+    {
+      "epoch": 2.3046875,
+      "grad_norm": 0.1395965814590454,
+      "learning_rate": 9.471230788201429e-06,
+      "loss": 0.2699,
+      "step": 198
+    },
+    {
+      "epoch": 2.31640625,
+      "grad_norm": 0.13987243175506592,
+      "learning_rate": 9.46200291623974e-06,
+      "loss": 0.2634,
+      "step": 199
+    },
+    {
+      "epoch": 2.328125,
+      "grad_norm": 0.17181117832660675,
+      "learning_rate": 9.452699794345583e-06,
+      "loss": 0.2501,
+      "step": 200
+    },
+    {
+      "epoch": 2.33984375,
+      "grad_norm": 0.12983955442905426,
+      "learning_rate": 9.443321579412465e-06,
+      "loss": 0.2383,
+      "step": 201
+    },
+    {
+      "epoch": 2.3515625,
+      "grad_norm": 0.17716065049171448,
+      "learning_rate": 9.43386842960031e-06,
+      "loss": 0.3096,
+      "step": 202
+    },
+    {
+      "epoch": 2.36328125,
+      "grad_norm": 0.16192656755447388,
+      "learning_rate": 9.42434050433279e-06,
+      "loss": 0.2825,
+      "step": 203
+    },
+    {
+      "epoch": 2.375,
+      "grad_norm": 0.1350512057542801,
+      "learning_rate": 9.414737964294636e-06,
+      "loss": 0.2301,
+      "step": 204
+    },
+    {
+      "epoch": 2.38671875,
+      "grad_norm": 0.1251179575920105,
+      "learning_rate": 9.405060971428924e-06,
+      "loss": 0.2682,
+      "step": 205
+    },
+    {
+      "epoch": 2.3984375,
+      "grad_norm": 0.13068172335624695,
+      "learning_rate": 9.39530968893435e-06,
+      "loss": 0.3303,
+      "step": 206
+    },
+    {
+      "epoch": 2.41015625,
+      "grad_norm": 0.16019192337989807,
+      "learning_rate": 9.38548428126248e-06,
+      "loss": 0.2651,
+      "step": 207
+    },
+    {
+      "epoch": 2.421875,
+      "grad_norm": 0.14367878437042236,
+      "learning_rate": 9.375584914114963e-06,
+      "loss": 0.2712,
+      "step": 208
+    },
+    {
+      "epoch": 2.43359375,
+      "grad_norm": 0.14105841517448425,
+      "learning_rate": 9.365611754440756e-06,
+      "loss": 0.2541,
+      "step": 209
+    },
+    {
+      "epoch": 2.4453125,
+      "grad_norm": 0.14219817519187927,
+      "learning_rate": 9.355564970433288e-06,
+      "loss": 0.2874,
+      "step": 210
+    },
+    {
+      "epoch": 2.45703125,
+      "grad_norm": 0.14144977927207947,
+      "learning_rate": 9.345444731527642e-06,
+      "loss": 0.255,
+      "step": 211
+    },
+    {
+      "epoch": 2.46875,
+      "grad_norm": 0.16394628584384918,
+      "learning_rate": 9.335251208397684e-06,
+      "loss": 0.2946,
+      "step": 212
+    },
+    {
+      "epoch": 2.48046875,
+      "grad_norm": 0.1422003209590912,
+      "learning_rate": 9.32498457295319e-06,
+      "loss": 0.275,
+      "step": 213
+    },
+    {
+      "epoch": 2.4921875,
+      "grad_norm": 0.153361514210701,
+      "learning_rate": 9.31464499833695e-06,
+      "loss": 0.2695,
+      "step": 214
+    },
+    {
+      "epoch": 2.50390625,
+      "grad_norm": 0.13666994869709015,
+      "learning_rate": 9.30423265892184e-06,
+      "loss": 0.2708,
+      "step": 215
+    },
+    {
+      "epoch": 2.515625,
+      "grad_norm": 0.13980181515216827,
+      "learning_rate": 9.29374773030789e-06,
+      "loss": 0.3132,
+      "step": 216
+    },
+    {
+      "epoch": 2.52734375,
+      "grad_norm": 0.1572592854499817,
+      "learning_rate": 9.283190389319315e-06,
+      "loss": 0.2802,
+      "step": 217
+    },
+    {
+      "epoch": 2.5390625,
+      "grad_norm": 0.12849898636341095,
+      "learning_rate": 9.27256081400154e-06,
+      "loss": 0.2813,
+      "step": 218
+    },
+    {
+      "epoch": 2.55078125,
+      "grad_norm": 0.14952799677848816,
+      "learning_rate": 9.26185918361819e-06,
+      "loss": 0.2662,
+      "step": 219
+    },
+    {
+      "epoch": 2.5625,
+      "grad_norm": 0.1414070427417755,
+      "learning_rate": 9.251085678648072e-06,
+      "loss": 0.2833,
+      "step": 220
+    },
+    {
+      "epoch": 2.57421875,
+      "grad_norm": 0.14631874859333038,
+      "learning_rate": 9.24024048078213e-06,
+      "loss": 0.2645,
+      "step": 221
+    },
+    {
+      "epoch": 2.5859375,
+      "grad_norm": 0.16351047158241272,
+      "learning_rate": 9.229323772920383e-06,
+      "loss": 0.2563,
+      "step": 222
+    },
+    {
+      "epoch": 2.59765625,
+      "grad_norm": 0.1420244574546814,
+      "learning_rate": 9.218335739168833e-06,
+      "loss": 0.3132,
+      "step": 223
+    },
+    {
+      "epoch": 2.609375,
+      "grad_norm": 0.16125503182411194,
+      "learning_rate": 9.207276564836367e-06,
+      "loss": 0.2533,
+      "step": 224
+    },
+    {
+      "epoch": 2.62109375,
+      "grad_norm": 0.16347342729568481,
+      "learning_rate": 9.196146436431635e-06,
+      "loss": 0.2884,
+      "step": 225
+    },
+    {
+      "epoch": 2.6328125,
+      "grad_norm": 0.15693005919456482,
+      "learning_rate": 9.18494554165989e-06,
+      "loss": 0.2644,
+      "step": 226
+    },
+    {
+      "epoch": 2.64453125,
+      "grad_norm": 0.15239854156970978,
+      "learning_rate": 9.173674069419843e-06,
+      "loss": 0.259,
+      "step": 227
+    },
+    {
+      "epoch": 2.65625,
+      "grad_norm": 0.14987824857234955,
+      "learning_rate": 9.162332209800455e-06,
+      "loss": 0.2792,
+      "step": 228
+    },
+    {
+      "epoch": 2.66796875,
+      "grad_norm": 0.14547227323055267,
+      "learning_rate": 9.150920154077753e-06,
+      "loss": 0.2663,
+      "step": 229
+    },
+    {
+      "epoch": 2.6796875,
+      "grad_norm": 0.15145206451416016,
+      "learning_rate": 9.13943809471159e-06,
+      "loss": 0.2885,
+      "step": 230
+    },
+    {
+      "epoch": 2.69140625,
+      "grad_norm": 0.17260988056659698,
+      "learning_rate": 9.1278862253424e-06,
+      "loss": 0.3377,
+      "step": 231
+    },
+    {
+      "epoch": 2.703125,
+      "grad_norm": 0.1483244150876999,
+      "learning_rate": 9.116264740787937e-06,
+      "loss": 0.2742,
+      "step": 232
+    },
+    {
+      "epoch": 2.71484375,
+      "grad_norm": 0.13184918463230133,
+      "learning_rate": 9.104573837039992e-06,
+      "loss": 0.2746,
+      "step": 233
+    },
+    {
+      "epoch": 2.7265625,
+      "grad_norm": 0.14674793183803558,
+      "learning_rate": 9.092813711261075e-06,
+      "loss": 0.2372,
+      "step": 234
+    },
+    {
+      "epoch": 2.73828125,
+      "grad_norm": 0.13162650167942047,
+      "learning_rate": 9.08098456178111e-06,
+      "loss": 0.287,
+      "step": 235
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 0.14429397881031036,
+      "learning_rate": 9.069086588094067e-06,
+      "loss": 0.2561,
+      "step": 236
+    },
+    {
+      "epoch": 2.76171875,
+      "grad_norm": 0.1422855108976364,
+      "learning_rate": 9.057119990854617e-06,
+      "loss": 0.3037,
+      "step": 237
+    },
+    {
+      "epoch": 2.7734375,
+      "grad_norm": 0.15853242576122284,
+      "learning_rate": 9.045084971874738e-06,
+      "loss": 0.2888,
+      "step": 238
+    },
+    {
+      "epoch": 2.78515625,
+      "grad_norm": 0.12172834575176239,
+      "learning_rate": 9.032981734120312e-06,
+      "loss": 0.2477,
+      "step": 239
+    },
+    {
+      "epoch": 2.796875,
+      "grad_norm": 0.1423560231924057,
+      "learning_rate": 9.020810481707709e-06,
+      "loss": 0.2867,
+      "step": 240
+    },
+    {
+      "epoch": 2.80859375,
+      "grad_norm": 0.1702727973461151,
+      "learning_rate": 9.008571419900334e-06,
+      "loss": 0.3015,
+      "step": 241
+    },
+    {
+      "epoch": 2.8203125,
+      "grad_norm": 0.14154493808746338,
+      "learning_rate": 8.996264755105173e-06,
+      "loss": 0.2771,
+      "step": 242
+    },
+    {
+      "epoch": 2.83203125,
+      "grad_norm": 0.14596149325370789,
+      "learning_rate": 8.983890694869312e-06,
+      "loss": 0.3017,
+      "step": 243
+    },
+    {
+      "epoch": 2.84375,
+      "grad_norm": 0.13621939718723297,
+      "learning_rate": 8.97144944787643e-06,
+      "loss": 0.2496,
+      "step": 244
+    },
+    {
+      "epoch": 2.85546875,
+      "grad_norm": 0.18428286910057068,
+      "learning_rate": 8.958941223943292e-06,
+      "loss": 0.2751,
+      "step": 245
+    },
+    {
+      "epoch": 2.8671875,
+      "grad_norm": 0.1273455023765564,
+      "learning_rate": 8.946366234016192e-06,
+      "loss": 0.2678,
+      "step": 246
+    },
+    {
+      "epoch": 2.87890625,
+      "grad_norm": 0.1442483514547348,
+      "learning_rate": 8.933724690167417e-06,
+      "loss": 0.2574,
+      "step": 247
+    },
+    {
+      "epoch": 2.890625,
+      "grad_norm": 0.12419985979795456,
+      "learning_rate": 8.921016805591654e-06,
+      "loss": 0.2512,
+      "step": 248
+    },
+    {
+      "epoch": 2.90234375,
+      "grad_norm": 0.13012221455574036,
+      "learning_rate": 8.908242794602401e-06,
+      "loss": 0.2674,
+      "step": 249
+    },
+    {
+      "epoch": 2.9140625,
+      "grad_norm": 0.13028892874717712,
+      "learning_rate": 8.895402872628352e-06,
+      "loss": 0.2566,
+      "step": 250
+    },
+    {
+      "epoch": 2.92578125,
+      "grad_norm": 0.1266089677810669,
+      "learning_rate": 8.882497256209767e-06,
+      "loss": 0.2671,
+      "step": 251
+    },
+    {
+      "epoch": 2.9375,
+      "grad_norm": 0.15992388129234314,
+      "learning_rate": 8.869526162994814e-06,
+      "loss": 0.276,
+      "step": 252
+    },
+    {
+      "epoch": 2.94921875,
+      "grad_norm": 0.13113290071487427,
+      "learning_rate": 8.856489811735904e-06,
+      "loss": 0.3073,
+      "step": 253
+    },
+    {
+      "epoch": 2.9609375,
+      "grad_norm": 0.15523600578308105,
+      "learning_rate": 8.843388422285995e-06,
+      "loss": 0.2704,
+      "step": 254
+    },
+    {
+      "epoch": 2.97265625,
+      "grad_norm": 0.13755646347999573,
+      "learning_rate": 8.83022221559489e-06,
+      "loss": 0.3174,
+      "step": 255
+    },
+    {
+      "epoch": 2.984375,
+      "grad_norm": 0.15444748103618622,
+      "learning_rate": 8.816991413705515e-06,
+      "loss": 0.2955,
+      "step": 256
+    },
+    {
+      "epoch": 2.99609375,
+      "grad_norm": 0.13223521411418915,
+      "learning_rate": 8.803696239750163e-06,
+      "loss": 0.2563,
+      "step": 257
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.28410595655441284,
+      "learning_rate": 8.790336917946737e-06,
+      "loss": 0.242,
+      "step": 258
+    },
+    {
+      "epoch": 3.01171875,
+      "grad_norm": 0.16228830814361572,
+      "learning_rate": 8.776913673594968e-06,
+      "loss": 0.1626,
+      "step": 259
+    },
+    {
+      "epoch": 3.0234375,
+      "grad_norm": 0.14956367015838623,
+      "learning_rate": 8.763426733072624e-06,
+      "loss": 0.1751,
+      "step": 260
+    },
+    {
+      "epoch": 3.03515625,
+      "grad_norm": 0.15641574561595917,
+      "learning_rate": 8.74987632383167e-06,
+      "loss": 0.1932,
+      "step": 261
+    },
+    {
+      "epoch": 3.046875,
+      "grad_norm": 0.18344224989414215,
+      "learning_rate": 8.736262674394455e-06,
+      "loss": 0.1646,
+      "step": 262
+    },
+    {
+      "epoch": 3.05859375,
+      "grad_norm": 0.23181785643100739,
+      "learning_rate": 8.722586014349851e-06,
+      "loss": 0.1855,
+      "step": 263
+    },
+    {
+      "epoch": 3.0703125,
+      "grad_norm": 0.1877327412366867,
+      "learning_rate": 8.708846574349372e-06,
+      "loss": 0.1779,
+      "step": 264
+    },
+    {
+      "epoch": 3.08203125,
+      "grad_norm": 0.1676245629787445,
+      "learning_rate": 8.695044586103297e-06,
+      "loss": 0.1621,
+      "step": 265
+    },
+    {
+      "epoch": 3.09375,
+      "grad_norm": 0.16094529628753662,
+      "learning_rate": 8.681180282376754e-06,
+      "loss": 0.1732,
+      "step": 266
+    },
+    {
+      "epoch": 3.10546875,
+      "grad_norm": 0.14831165969371796,
+      "learning_rate": 8.667253896985796e-06,
+      "loss": 0.1678,
+      "step": 267
+    },
+    {
+      "epoch": 3.1171875,
+      "grad_norm": 0.1500885933637619,
+      "learning_rate": 8.653265664793466e-06,
+      "loss": 0.1597,
+      "step": 268
+    },
+    {
+      "epoch": 3.12890625,
+      "grad_norm": 0.13528233766555786,
+      "learning_rate": 8.639215821705821e-06,
+      "loss": 0.2005,
+      "step": 269
+    },
+    {
+      "epoch": 3.140625,
+      "grad_norm": 0.1544928252696991,
+      "learning_rate": 8.625104604667965e-06,
+      "loss": 0.1658,
+      "step": 270
+    },
+    {
+      "epoch": 3.15234375,
+      "grad_norm": 0.1660798341035843,
+      "learning_rate": 8.610932251660046e-06,
+      "loss": 0.2022,
+      "step": 271
+    },
+    {
+      "epoch": 3.1640625,
+      "grad_norm": 0.17334631085395813,
+      "learning_rate": 8.596699001693257e-06,
+      "loss": 0.1821,
+      "step": 272
+    },
+    {
+      "epoch": 3.17578125,
+      "grad_norm": 0.23266500234603882,
+      "learning_rate": 8.58240509480578e-06,
+      "loss": 0.1905,
+      "step": 273
+    },
+    {
+      "epoch": 3.1875,
+      "grad_norm": 0.14414437115192413,
+      "learning_rate": 8.568050772058763e-06,
+      "loss": 0.1792,
+      "step": 274
+    },
+    {
+      "epoch": 3.19921875,
+      "grad_norm": 0.1535835564136505,
+      "learning_rate": 8.553636275532236e-06,
+      "loss": 0.1699,
+      "step": 275
+    },
+    {
+      "epoch": 3.2109375,
+      "grad_norm": 0.17544692754745483,
+      "learning_rate": 8.539161848321047e-06,
+      "loss": 0.216,
+      "step": 276
+    },
+    {
+      "epoch": 3.22265625,
+      "grad_norm": 0.14662209153175354,
+      "learning_rate": 8.524627734530738e-06,
+      "loss": 0.1655,
+      "step": 277
+    },
+    {
+      "epoch": 3.234375,
+      "grad_norm": 0.14544612169265747,
+      "learning_rate": 8.51003417927345e-06,
+      "loss": 0.1764,
+      "step": 278
+    },
+    {
+      "epoch": 3.24609375,
+      "grad_norm": 0.16317211091518402,
+      "learning_rate": 8.49538142866378e-06,
+      "loss": 0.1791,
+      "step": 279
+    },
+    {
+      "epoch": 3.2578125,
+      "grad_norm": 0.1408390998840332,
+      "learning_rate": 8.480669729814635e-06,
+      "loss": 0.1571,
+      "step": 280
+    },
+    {
+      "epoch": 3.26953125,
+      "grad_norm": 0.14905256032943726,
+      "learning_rate": 8.465899330833051e-06,
+      "loss": 0.1513,
+      "step": 281
+    },
+    {
+      "epoch": 3.28125,
+      "grad_norm": 0.13584789633750916,
+      "learning_rate": 8.451070480816027e-06,
+      "loss": 0.1677,
+      "step": 282
+    },
+    {
+      "epoch": 3.29296875,
+      "grad_norm": 0.14696229994297028,
+      "learning_rate": 8.436183429846314e-06,
+      "loss": 0.1812,
+      "step": 283
+    },
+    {
+      "epoch": 3.3046875,
+      "grad_norm": 0.21364903450012207,
+      "learning_rate": 8.421238428988199e-06,
+      "loss": 0.1875,
+      "step": 284
+    },
+    {
+      "epoch": 3.31640625,
+      "grad_norm": 0.15450192987918854,
+      "learning_rate": 8.40623573028327e-06,
+      "loss": 0.1763,
+      "step": 285
+    },
+    {
+      "epoch": 3.328125,
+      "grad_norm": 0.14626288414001465,
+      "learning_rate": 8.39117558674617e-06,
+      "loss": 0.1625,
+      "step": 286
+    },
+    {
+      "epoch": 3.33984375,
+      "grad_norm": 0.15964478254318237,
+      "learning_rate": 8.376058252360322e-06,
+      "loss": 0.1896,
+      "step": 287
+    },
+    {
+      "epoch": 3.3515625,
+      "grad_norm": 0.14050042629241943,
+      "learning_rate": 8.360883982073653e-06,
+      "loss": 0.1695,
+      "step": 288
+    },
+    {
+      "epoch": 3.36328125,
+      "grad_norm": 0.16250889003276825,
+      "learning_rate": 8.345653031794292e-06,
+      "loss": 0.1638,
+      "step": 289
+    },
+    {
+      "epoch": 3.375,
+      "grad_norm": 0.14831937849521637,
+      "learning_rate": 8.330365658386252e-06,
+      "loss": 0.1714,
+      "step": 290
+    },
+    {
+      "epoch": 3.38671875,
+      "grad_norm": 0.15942546725273132,
+      "learning_rate": 8.3150221196651e-06,
+      "loss": 0.1744,
+      "step": 291
+    },
+    {
+      "epoch": 3.3984375,
+      "grad_norm": 0.14313259720802307,
+      "learning_rate": 8.299622674393615e-06,
+      "loss": 0.1578,
+      "step": 292
+    },
+    {
+      "epoch": 3.41015625,
+      "grad_norm": 0.1886913925409317,
+      "learning_rate": 8.284167582277406e-06,
+      "loss": 0.2234,
+      "step": 293
+    },
+    {
+      "epoch": 3.421875,
+      "grad_norm": 0.16603605449199677,
+      "learning_rate": 8.268657103960558e-06,
+      "loss": 0.1891,
+      "step": 294
+    },
+    {
+      "epoch": 3.43359375,
+      "grad_norm": 0.18278054893016815,
+      "learning_rate": 8.25309150102121e-06,
+      "loss": 0.1847,
+      "step": 295
+    },
+    {
+      "epoch": 3.4453125,
+      "grad_norm": 0.15222598612308502,
+      "learning_rate": 8.237471035967168e-06,
+      "loss": 0.181,
+      "step": 296
+    },
+    {
+      "epoch": 3.45703125,
+      "grad_norm": 0.16404354572296143,
+      "learning_rate": 8.221795972231459e-06,
+      "loss": 0.1682,
+      "step": 297
+    },
+    {
+      "epoch": 3.46875,
+      "grad_norm": 0.15001647174358368,
+      "learning_rate": 8.206066574167893e-06,
+      "loss": 0.1589,
+      "step": 298
+    },
+    {
+      "epoch": 3.48046875,
+      "grad_norm": 0.1620359718799591,
+      "learning_rate": 8.190283107046613e-06,
+      "loss": 0.1914,
+      "step": 299
+    },
+    {
+      "epoch": 3.4921875,
+      "grad_norm": 0.17575767636299133,
+      "learning_rate": 8.174445837049614e-06,
+      "loss": 0.1713,
+      "step": 300
+    },
+    {
+      "epoch": 3.50390625,
+      "grad_norm": 0.16479137539863586,
+      "learning_rate": 8.158555031266255e-06,
+      "loss": 0.1958,
+      "step": 301
+    },
+    {
+      "epoch": 3.515625,
+      "grad_norm": 0.14474551379680634,
+      "learning_rate": 8.142610957688755e-06,
+      "loss": 0.1741,
+      "step": 302
+    },
+    {
+      "epoch": 3.52734375,
+      "grad_norm": 0.15629877150058746,
+      "learning_rate": 8.12661388520767e-06,
+      "loss": 0.1925,
+      "step": 303
+    },
+    {
+      "epoch": 3.5390625,
+      "grad_norm": 0.21056914329528809,
+      "learning_rate": 8.110564083607371e-06,
+      "loss": 0.1776,
+      "step": 304
+    },
+    {
+      "epoch": 3.55078125,
+      "grad_norm": 0.15697474777698517,
+      "learning_rate": 8.094461823561473e-06,
+      "loss": 0.16,
+      "step": 305
+    },
+    {
+      "epoch": 3.5625,
+      "grad_norm": 0.15944577753543854,
+      "learning_rate": 8.078307376628292e-06,
+      "loss": 0.1784,
+      "step": 306
+    },
+    {
+      "epoch": 3.57421875,
+      "grad_norm": 0.1611972153186798,
+      "learning_rate": 8.06210101524625e-06,
+      "loss": 0.1914,
+      "step": 307
+    },
+    {
+      "epoch": 3.5859375,
+      "grad_norm": 0.18928834795951843,
+      "learning_rate": 8.045843012729288e-06,
+      "loss": 0.2047,
+      "step": 308
+    },
+    {
+      "epoch": 3.59765625,
+      "grad_norm": 0.17320488393306732,
+      "learning_rate": 8.029533643262257e-06,
+      "loss": 0.1882,
+      "step": 309
+    },
+    {
+      "epoch": 3.609375,
+      "grad_norm": 0.1613771617412567,
+      "learning_rate": 8.013173181896283e-06,
+      "loss": 0.1814,
+      "step": 310
+    },
+    {
+      "epoch": 3.62109375,
+      "grad_norm": 0.15916645526885986,
+      "learning_rate": 7.996761904544146e-06,
+      "loss": 0.1905,
+      "step": 311
+    },
+    {
+      "epoch": 3.6328125,
+      "grad_norm": 0.1529555767774582,
+      "learning_rate": 7.980300087975612e-06,
+      "loss": 0.1835,
+      "step": 312
+    },
+    {
+      "epoch": 3.64453125,
+      "grad_norm": 0.16544155776500702,
+      "learning_rate": 7.963788009812775e-06,
+      "loss": 0.1791,
+      "step": 313
+    },
+    {
+      "epoch": 3.65625,
+      "grad_norm": 0.1579754799604416,
+      "learning_rate": 7.94722594852537e-06,
+      "loss": 0.1943,
+      "step": 314
+    },
+    {
+      "epoch": 3.66796875,
+      "grad_norm": 0.16517029702663422,
+      "learning_rate": 7.930614183426074e-06,
+      "loss": 0.1949,
+      "step": 315
+    },
+    {
+      "epoch": 3.6796875,
+      "grad_norm": 0.19699627161026,
+      "learning_rate": 7.913952994665805e-06,
+      "loss": 0.1816,
+      "step": 316
+    },
+    {
+      "epoch": 3.69140625,
+      "grad_norm": 0.15589186549186707,
+      "learning_rate": 7.89724266322899e-06,
+      "loss": 0.1857,
+      "step": 317
+    },
+    {
+      "epoch": 3.703125,
+      "grad_norm": 0.15042664110660553,
+      "learning_rate": 7.880483470928823e-06,
+      "loss": 0.1659,
+      "step": 318
+    },
+    {
+      "epoch": 3.71484375,
+      "grad_norm": 0.1443316787481308,
+      "learning_rate": 7.863675700402527e-06,
+      "loss": 0.1741,
+      "step": 319
+    },
+    {
+      "epoch": 3.7265625,
+      "grad_norm": 0.14981609582901,
+      "learning_rate": 7.846819635106569e-06,
+      "loss": 0.1998,
+      "step": 320
+    },
+    {
+      "epoch": 3.73828125,
+      "grad_norm": 0.13699378073215485,
+      "learning_rate": 7.829915559311892e-06,
+      "loss": 0.1636,
+      "step": 321
+    },
+    {
+      "epoch": 3.75,
+      "grad_norm": 0.14549602568149567,
+      "learning_rate": 7.812963758099118e-06,
+      "loss": 0.1732,
+      "step": 322
+    },
+    {
+      "epoch": 3.76171875,
+      "grad_norm": 0.15110832452774048,
+      "learning_rate": 7.795964517353734e-06,
+      "loss": 0.1816,
+      "step": 323
+    },
+    {
+      "epoch": 3.7734375,
+      "grad_norm": 0.1504000872373581,
+      "learning_rate": 7.778918123761287e-06,
+      "loss": 0.1656,
+      "step": 324
+    },
+    {
+      "epoch": 3.78515625,
+      "grad_norm": 0.15452761948108673,
+      "learning_rate": 7.76182486480253e-06,
+      "loss": 0.1764,
+      "step": 325
+    },
+    {
+      "epoch": 3.796875,
+      "grad_norm": 0.14827081561088562,
+      "learning_rate": 7.744685028748582e-06,
+      "loss": 0.159,
+      "step": 326
+    },
+    {
+      "epoch": 3.80859375,
+      "grad_norm": 0.140132874250412,
+      "learning_rate": 7.72749890465607e-06,
+      "loss": 0.1923,
+      "step": 327
+    },
+    {
+      "epoch": 3.8203125,
+      "grad_norm": 0.1527477502822876,
+      "learning_rate": 7.710266782362248e-06,
+      "loss": 0.1586,
+      "step": 328
+    },
+    {
+      "epoch": 3.83203125,
+      "grad_norm": 0.14466483891010284,
+      "learning_rate": 7.692988952480114e-06,
+      "loss": 0.1669,
+      "step": 329
+    },
+    {
+      "epoch": 3.84375,
+      "grad_norm": 0.15228520333766937,
+      "learning_rate": 7.675665706393502e-06,
+      "loss": 0.167,
+      "step": 330
+    },
+    {
+      "epoch": 3.85546875,
+      "grad_norm": 0.18122923374176025,
+      "learning_rate": 7.658297336252181e-06,
+      "loss": 0.1981,
+      "step": 331
+    },
+    {
+      "epoch": 3.8671875,
+      "grad_norm": 0.1492803692817688,
+      "learning_rate": 7.64088413496691e-06,
+      "loss": 0.1802,
+      "step": 332
+    },
+    {
+      "epoch": 3.87890625,
+      "grad_norm": 0.15203335881233215,
+      "learning_rate": 7.623426396204516e-06,
+      "loss": 0.1935,
+      "step": 333
+    },
+    {
+      "epoch": 3.890625,
+      "grad_norm": 0.15901945531368256,
+      "learning_rate": 7.605924414382926e-06,
+      "loss": 0.1711,
+      "step": 334
+    },
+    {
+      "epoch": 3.90234375,
+      "grad_norm": 0.14871086180210114,
+      "learning_rate": 7.588378484666214e-06,
+      "loss": 0.1658,
+      "step": 335
+    },
+    {
+      "epoch": 3.9140625,
+      "grad_norm": 0.16054388880729675,
+      "learning_rate": 7.570788902959612e-06,
+      "loss": 0.1915,
+      "step": 336
+    },
+    {
+      "epoch": 3.92578125,
+      "grad_norm": 0.14552587270736694,
+      "learning_rate": 7.553155965904535e-06,
+      "loss": 0.1901,
+      "step": 337
+    },
+    {
+      "epoch": 3.9375,
+      "grad_norm": 0.15138490498065948,
+      "learning_rate": 7.535479970873563e-06,
+      "loss": 0.2054,
+      "step": 338
+    },
+    {
+      "epoch": 3.94921875,
+      "grad_norm": 0.15584439039230347,
+      "learning_rate": 7.517761215965429e-06,
+      "loss": 0.1863,
+      "step": 339
+    },
+    {
+      "epoch": 3.9609375,
+      "grad_norm": 0.15878607332706451,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.2298,
+      "step": 340
+    },
+    {
+      "epoch": 3.97265625,
+      "grad_norm": 0.14984527230262756,
+      "learning_rate": 7.482196622513233e-06,
+      "loss": 0.1862,
+      "step": 341
+    },
+    {
+      "epoch": 3.984375,
+      "grad_norm": 0.1370866894721985,
+      "learning_rate": 7.464351383752117e-06,
+      "loss": 0.1711,
+      "step": 342
+    },
+    {
+      "epoch": 3.99609375,
+      "grad_norm": 0.13851135969161987,
+      "learning_rate": 7.4464645846696186e-06,
+      "loss": 0.1564,
+      "step": 343
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.13851135969161987,
+      "learning_rate": 7.428536526919603e-06,
+      "loss": 0.1124,
+      "step": 344
+    },
+    {
+      "epoch": 4.01171875,
+      "grad_norm": 0.21845901012420654,
+      "learning_rate": 7.4105675128517456e-06,
+      "loss": 0.1096,
+      "step": 345
+    },
+    {
+      "epoch": 4.0234375,
+      "grad_norm": 0.18661820888519287,
+      "learning_rate": 7.392557845506433e-06,
+      "loss": 0.1378,
+      "step": 346
+    },
+    {
+      "epoch": 4.03515625,
+      "grad_norm": 0.16629564762115479,
+      "learning_rate": 7.374507828609657e-06,
+      "loss": 0.1173,
+      "step": 347
+    },
+    {
+      "epoch": 4.046875,
+      "grad_norm": 0.15008185803890228,
+      "learning_rate": 7.356417766567887e-06,
+      "loss": 0.1028,
+      "step": 348
+    },
+    {
+      "epoch": 4.05859375,
+      "grad_norm": 0.1788836270570755,
+      "learning_rate": 7.3382879644629345e-06,
+      "loss": 0.1065,
+      "step": 349
+    },
+    {
+      "epoch": 4.0703125,
+      "grad_norm": 0.19552384316921234,
+      "learning_rate": 7.320118728046818e-06,
+      "loss": 0.0964,
+      "step": 350
+    },
+    {
+      "epoch": 4.08203125,
+      "grad_norm": 0.17942851781845093,
+      "learning_rate": 7.301910363736596e-06,
+      "loss": 0.093,
+      "step": 351
+    },
+    {
+      "epoch": 4.09375,
+      "grad_norm": 0.1425376981496811,
+      "learning_rate": 7.283663178609204e-06,
+      "loss": 0.0923,
+      "step": 352
+    },
+    {
+      "epoch": 4.10546875,
+      "grad_norm": 0.14930015802383423,
+      "learning_rate": 7.265377480396277e-06,
+      "loss": 0.1254,
+      "step": 353
+    },
+    {
+      "epoch": 4.1171875,
+      "grad_norm": 0.1838270127773285,
+      "learning_rate": 7.247053577478955e-06,
+      "loss": 0.0997,
+      "step": 354
+    },
+    {
+      "epoch": 4.12890625,
+      "grad_norm": 0.16155952215194702,
+      "learning_rate": 7.2286917788826926e-06,
+      "loss": 0.0889,
+      "step": 355
+    },
+    {
+      "epoch": 4.140625,
+      "grad_norm": 0.15314266085624695,
+      "learning_rate": 7.210292394272029e-06,
+      "loss": 0.1011,
+      "step": 356
+    },
+    {
+      "epoch": 4.15234375,
+      "grad_norm": 0.14763030409812927,
+      "learning_rate": 7.191855733945388e-06,
+      "loss": 0.1017,
+      "step": 357
+    },
+    {
+      "epoch": 4.1640625,
+      "grad_norm": 0.1658443957567215,
+      "learning_rate": 7.173382108829826e-06,
+      "loss": 0.0825,
+      "step": 358
+    },
+    {
+      "epoch": 4.17578125,
+      "grad_norm": 0.14803840219974518,
+      "learning_rate": 7.154871830475798e-06,
+      "loss": 0.0986,
+      "step": 359
+    },
+    {
+      "epoch": 4.1875,
+      "grad_norm": 0.13587775826454163,
+      "learning_rate": 7.136325211051905e-06,
+      "loss": 0.1074,
+      "step": 360
+    },
+    {
+      "epoch": 4.19921875,
+      "grad_norm": 0.1521381437778473,
+      "learning_rate": 7.117742563339622e-06,
+      "loss": 0.095,
+      "step": 361
+    },
+    {
+      "epoch": 4.2109375,
+      "grad_norm": 0.1659468561410904,
+      "learning_rate": 7.099124200728028e-06,
+      "loss": 0.1028,
+      "step": 362
+    },
+    {
+      "epoch": 4.22265625,
+      "grad_norm": 0.17924466729164124,
+      "learning_rate": 7.0804704372085205e-06,
+      "loss": 0.0952,
+      "step": 363
+    },
+    {
+      "epoch": 4.234375,
+      "grad_norm": 0.17670685052871704,
+      "learning_rate": 7.061781587369518e-06,
+      "loss": 0.1299,
+      "step": 364
+    },
+    {
+      "epoch": 4.24609375,
+      "grad_norm": 0.15658561885356903,
+      "learning_rate": 7.043057966391158e-06,
+      "loss": 0.1098,
+      "step": 365
+    },
+    {
+      "epoch": 4.2578125,
+      "grad_norm": 0.1510300636291504,
+      "learning_rate": 7.024299890039978e-06,
+      "loss": 0.0922,
+      "step": 366
+    },
+    {
+      "epoch": 4.26953125,
+      "grad_norm": 0.15231317281723022,
+      "learning_rate": 7.005507674663594e-06,
+      "loss": 0.1098,
+      "step": 367
+    },
+    {
+      "epoch": 4.28125,
+      "grad_norm": 0.16875913739204407,
+      "learning_rate": 6.986681637185361e-06,
+      "loss": 0.0994,
+      "step": 368
+    },
+    {
+      "epoch": 4.29296875,
+      "grad_norm": 0.19646456837654114,
+      "learning_rate": 6.967822095099031e-06,
+      "loss": 0.1197,
+      "step": 369
+    },
+    {
+      "epoch": 4.3046875,
+      "grad_norm": 0.14720042049884796,
+      "learning_rate": 6.948929366463397e-06,
+      "loss": 0.1208,
+      "step": 370
+    },
+    {
+      "epoch": 4.31640625,
+      "grad_norm": 0.1687857210636139,
+      "learning_rate": 6.930003769896936e-06,
+      "loss": 0.1019,
+      "step": 371
+    },
+    {
+      "epoch": 4.328125,
+      "grad_norm": 0.17267028987407684,
+      "learning_rate": 6.91104562457242e-06,
+      "loss": 0.1051,
+      "step": 372
+    },
+    {
+      "epoch": 4.33984375,
+      "grad_norm": 0.15575772523880005,
+      "learning_rate": 6.892055250211552e-06,
+      "loss": 0.1319,
+      "step": 373
+    },
+    {
+      "epoch": 4.3515625,
+      "grad_norm": 0.16643179953098297,
+      "learning_rate": 6.873032967079562e-06,
+      "loss": 0.1241,
+      "step": 374
+    },
+    {
+      "epoch": 4.36328125,
+      "grad_norm": 0.18457186222076416,
+      "learning_rate": 6.8539790959798045e-06,
+      "loss": 0.1018,
+      "step": 375
+    },
+    {
+      "epoch": 4.375,
+      "grad_norm": 0.15639866888523102,
+      "learning_rate": 6.834893958248361e-06,
+      "loss": 0.1073,
+      "step": 376
+    },
+    {
+      "epoch": 4.38671875,
+      "grad_norm": 0.14814910292625427,
+      "learning_rate": 6.815777875748607e-06,
+      "loss": 0.1018,
+      "step": 377
+    },
+    {
+      "epoch": 4.3984375,
+      "grad_norm": 0.14821702241897583,
+      "learning_rate": 6.7966311708657884e-06,
+      "loss": 0.096,
+      "step": 378
+    },
+    {
+      "epoch": 4.41015625,
+      "grad_norm": 0.16026665270328522,
+      "learning_rate": 6.77745416650159e-06,
+      "loss": 0.0961,
+      "step": 379
+    },
+    {
+      "epoch": 4.421875,
+      "grad_norm": 0.135820209980011,
+      "learning_rate": 6.758247186068684e-06,
+      "loss": 0.0898,
+      "step": 380
+    },
+    {
+      "epoch": 4.43359375,
+      "grad_norm": 0.1573071926832199,
+      "learning_rate": 6.739010553485276e-06,
+      "loss": 0.1238,
+      "step": 381
+    },
+    {
+      "epoch": 4.4453125,
+      "grad_norm": 0.15640759468078613,
+      "learning_rate": 6.719744593169642e-06,
+      "loss": 0.1109,
+      "step": 382
+    },
+    {
+      "epoch": 4.45703125,
+      "grad_norm": 0.1494980901479721,
+      "learning_rate": 6.700449630034662e-06,
+      "loss": 0.0996,
+      "step": 383
+    },
+    {
+      "epoch": 4.46875,
+      "grad_norm": 0.13157877326011658,
+      "learning_rate": 6.681125989482337e-06,
+      "loss": 0.1078,
+      "step": 384
+    },
+    {
+      "epoch": 4.48046875,
+      "grad_norm": 0.17184874415397644,
+      "learning_rate": 6.6617739973982985e-06,
+      "loss": 0.1042,
+      "step": 385
+    },
+    {
+      "epoch": 4.4921875,
+      "grad_norm": 0.14946678280830383,
+      "learning_rate": 6.64239398014632e-06,
+      "loss": 0.1083,
+      "step": 386
+    },
+    {
+      "epoch": 4.50390625,
+      "grad_norm": 0.19056496024131775,
+      "learning_rate": 6.622986264562804e-06,
+      "loss": 0.127,
+      "step": 387
+    },
+    {
+      "epoch": 4.515625,
+      "grad_norm": 0.14448995888233185,
+      "learning_rate": 6.6035511779512764e-06,
+      "loss": 0.1038,
+      "step": 388
+    },
+    {
+      "epoch": 4.52734375,
+      "grad_norm": 0.1500634104013443,
+      "learning_rate": 6.584089048076866e-06,
+      "loss": 0.1069,
+      "step": 389
+    },
+    {
+      "epoch": 4.5390625,
+      "grad_norm": 0.14609850943088531,
+      "learning_rate": 6.5646002031607726e-06,
+      "loss": 0.103,
+      "step": 390
+    },
+    {
+      "epoch": 4.55078125,
+      "grad_norm": 0.15358710289001465,
+      "learning_rate": 6.545084971874738e-06,
+      "loss": 0.1128,
+      "step": 391
+    },
+    {
+      "epoch": 4.5625,
+      "grad_norm": 0.158953458070755,
+      "learning_rate": 6.525543683335497e-06,
+      "loss": 0.1028,
+      "step": 392
+    },
+    {
+      "epoch": 4.57421875,
+      "grad_norm": 0.1573840230703354,
+      "learning_rate": 6.505976667099233e-06,
+      "loss": 0.1167,
+      "step": 393
+    },
+    {
+      "epoch": 4.5859375,
+      "grad_norm": 0.16527414321899414,
+      "learning_rate": 6.486384253156014e-06,
+      "loss": 0.1178,
+      "step": 394
+    },
+    {
+      "epoch": 4.59765625,
+      "grad_norm": 0.14780762791633606,
+      "learning_rate": 6.466766771924231e-06,
+      "loss": 0.0999,
+      "step": 395
+    },
+    {
+      "epoch": 4.609375,
+      "grad_norm": 0.175028994679451,
+      "learning_rate": 6.447124554245026e-06,
+      "loss": 0.1273,
+      "step": 396
+    },
+    {
+      "epoch": 4.62109375,
+      "grad_norm": 0.16705100238323212,
+      "learning_rate": 6.427457931376712e-06,
+      "loss": 0.0969,
+      "step": 397
+    },
+    {
+      "epoch": 4.6328125,
+      "grad_norm": 0.1503540724515915,
+      "learning_rate": 6.407767234989181e-06,
+      "loss": 0.0962,
+      "step": 398
+    },
+    {
+      "epoch": 4.64453125,
+      "grad_norm": 0.17881572246551514,
+      "learning_rate": 6.388052797158324e-06,
+      "loss": 0.1143,
+      "step": 399
+    },
+    {
+      "epoch": 4.65625,
+      "grad_norm": 0.13578246533870697,
+      "learning_rate": 6.368314950360416e-06,
+      "loss": 0.122,
+      "step": 400
+    },
+    {
+      "epoch": 4.66796875,
+      "grad_norm": 0.19132906198501587,
+      "learning_rate": 6.3485540274665134e-06,
+      "loss": 0.0979,
+      "step": 401
+    },
+    {
+      "epoch": 4.6796875,
+      "grad_norm": 0.12635491788387299,
+      "learning_rate": 6.32877036173685e-06,
+      "loss": 0.1021,
+      "step": 402
+    },
+    {
+      "epoch": 4.69140625,
+      "grad_norm": 0.2278641015291214,
+      "learning_rate": 6.308964286815203e-06,
+      "loss": 0.1107,
+      "step": 403
+    },
+    {
+      "epoch": 4.703125,
+      "grad_norm": 0.2541256844997406,
+      "learning_rate": 6.289136136723268e-06,
+      "loss": 0.1026,
+      "step": 404
+    },
+    {
+      "epoch": 4.71484375,
+      "grad_norm": 0.16736242175102234,
+      "learning_rate": 6.269286245855039e-06,
+      "loss": 0.1211,
+      "step": 405
+    },
+    {
+      "epoch": 4.7265625,
+      "grad_norm": 0.22079572081565857,
+      "learning_rate": 6.249414948971154e-06,
+      "loss": 0.1321,
+      "step": 406
+    },
+    {
+      "epoch": 4.73828125,
+      "grad_norm": 0.16761334240436554,
+      "learning_rate": 6.229522581193257e-06,
+      "loss": 0.1219,
+      "step": 407
+    },
+    {
+      "epoch": 4.75,
+      "grad_norm": 0.19656169414520264,
+      "learning_rate": 6.209609477998339e-06,
+      "loss": 0.1115,
+      "step": 408
+    },
+    {
+      "epoch": 4.76171875,
+      "grad_norm": 0.14829809963703156,
+      "learning_rate": 6.189675975213094e-06,
+      "loss": 0.0969,
+      "step": 409
+    },
+    {
+      "epoch": 4.7734375,
+      "grad_norm": 0.14715968072414398,
+      "learning_rate": 6.169722409008244e-06,
+      "loss": 0.1179,
+      "step": 410
+    },
+    {
+      "epoch": 4.78515625,
+      "grad_norm": 0.17647211253643036,
+      "learning_rate": 6.1497491158928694e-06,
+      "loss": 0.0951,
+      "step": 411
+    },
+    {
+      "epoch": 4.796875,
+      "grad_norm": 0.1545528769493103,
+      "learning_rate": 6.129756432708739e-06,
+      "loss": 0.1028,
+      "step": 412
+    },
+    {
+      "epoch": 4.80859375,
+      "grad_norm": 0.15555883944034576,
+      "learning_rate": 6.109744696624631e-06,
+      "loss": 0.0929,
+      "step": 413
+    },
+    {
+      "epoch": 4.8203125,
+      "grad_norm": 0.1452905237674713,
+      "learning_rate": 6.089714245130639e-06,
+      "loss": 0.1001,
+      "step": 414
+    },
+    {
+      "epoch": 4.83203125,
+      "grad_norm": 0.1887577474117279,
+      "learning_rate": 6.0696654160324875e-06,
+      "loss": 0.1297,
+      "step": 415
+    },
+    {
+      "epoch": 4.84375,
+      "grad_norm": 0.15246422588825226,
+      "learning_rate": 6.049598547445829e-06,
+      "loss": 0.1085,
+      "step": 416
+    },
+    {
+      "epoch": 4.85546875,
+      "grad_norm": 0.1306416392326355,
+      "learning_rate": 6.02951397779055e-06,
+      "loss": 0.0879,
+      "step": 417
+    },
+    {
+      "epoch": 4.8671875,
+      "grad_norm": 0.16870683431625366,
+      "learning_rate": 6.009412045785051e-06,
+      "loss": 0.114,
+      "step": 418
+    },
+    {
+      "epoch": 4.87890625,
+      "grad_norm": 0.16079163551330566,
+      "learning_rate": 5.98929309044055e-06,
+      "loss": 0.1084,
+      "step": 419
+    },
+    {
+      "epoch": 4.890625,
+      "grad_norm": 0.1713971495628357,
+      "learning_rate": 5.9691574510553505e-06,
+      "loss": 0.1185,
+      "step": 420
+    },
+    {
+      "epoch": 4.90234375,
+      "grad_norm": 0.14745403826236725,
+      "learning_rate": 5.9490054672091305e-06,
+      "loss": 0.1064,
+      "step": 421
+    },
+    {
+      "epoch": 4.9140625,
+      "grad_norm": 0.1969185173511505,
+      "learning_rate": 5.928837478757206e-06,
+      "loss": 0.0908,
+      "step": 422
+    },
+    {
+      "epoch": 4.92578125,
+      "grad_norm": 0.2246403843164444,
+      "learning_rate": 5.908653825824808e-06,
+      "loss": 0.1175,
+      "step": 423
+    },
+    {
+      "epoch": 4.9375,
+      "grad_norm": 0.149653822183609,
+      "learning_rate": 5.888454848801345e-06,
+      "loss": 0.0957,
+      "step": 424
+    },
+    {
+      "epoch": 4.94921875,
+      "grad_norm": 0.13323569297790527,
+      "learning_rate": 5.8682408883346535e-06,
+      "loss": 0.0923,
+      "step": 425
+    },
+    {
+      "epoch": 4.9609375,
+      "grad_norm": 0.16845755279064178,
+      "learning_rate": 5.848012285325264e-06,
+      "loss": 0.1047,
+      "step": 426
+    },
+    {
+      "epoch": 4.97265625,
+      "grad_norm": 0.12530267238616943,
+      "learning_rate": 5.82776938092065e-06,
+      "loss": 0.1052,
+      "step": 427
+    },
+    {
+      "epoch": 4.984375,
+      "grad_norm": 0.16033221781253815,
+      "learning_rate": 5.807512516509468e-06,
+      "loss": 0.1021,
+      "step": 428
+    },
+    {
+      "epoch": 4.99609375,
+      "grad_norm": 0.15662384033203125,
+      "learning_rate": 5.787242033715809e-06,
+      "loss": 0.1029,
+      "step": 429
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.15662384033203125,
+      "learning_rate": 5.766958274393428e-06,
+      "loss": 0.0769,
+      "step": 430
+    },
+    {
+      "epoch": 5.01171875,
+      "grad_norm": 0.23790201544761658,
+      "learning_rate": 5.746661580619985e-06,
+      "loss": 0.0638,
+      "step": 431
+    },
+    {
+      "epoch": 5.0234375,
+      "grad_norm": 0.13754722476005554,
+      "learning_rate": 5.726352294691275e-06,
+      "loss": 0.0464,
+      "step": 432
+    },
+    {
+      "epoch": 5.03515625,
+      "grad_norm": 0.10853724926710129,
+      "learning_rate": 5.706030759115458e-06,
+      "loss": 0.0506,
+      "step": 433
+    },
+    {
+      "epoch": 5.046875,
+      "grad_norm": 0.13977329432964325,
+      "learning_rate": 5.685697316607274e-06,
+      "loss": 0.0551,
+      "step": 434
+    },
+    {
+      "epoch": 5.05859375,
+      "grad_norm": 0.15899504721164703,
+      "learning_rate": 5.66535231008227e-06,
+      "loss": 0.0577,
+      "step": 435
+    },
+    {
+      "epoch": 5.0703125,
+      "grad_norm": 0.15688322484493256,
+      "learning_rate": 5.644996082651018e-06,
+      "loss": 0.0613,
+      "step": 436
+    },
+    {
+      "epoch": 5.08203125,
+      "grad_norm": 0.1396789848804474,
+      "learning_rate": 5.6246289776133246e-06,
+      "loss": 0.0594,
+      "step": 437
+    },
+    {
+      "epoch": 5.09375,
+      "grad_norm": 0.18655186891555786,
+      "learning_rate": 5.604251338452444e-06,
+      "loss": 0.0511,
+      "step": 438
+    },
+    {
+      "epoch": 5.10546875,
+      "grad_norm": 0.1351338028907776,
+      "learning_rate": 5.583863508829281e-06,
+      "loss": 0.0652,
+      "step": 439
+    },
+    {
+      "epoch": 5.1171875,
+      "grad_norm": 0.14356586337089539,
+      "learning_rate": 5.5634658325766066e-06,
+      "loss": 0.0672,
+      "step": 440
+    },
+    {
+      "epoch": 5.12890625,
+      "grad_norm": 0.17796984314918518,
+      "learning_rate": 5.543058653693241e-06,
+      "loss": 0.0604,
+      "step": 441
+    },
+    {
+      "epoch": 5.140625,
+      "grad_norm": 0.14155617356300354,
+      "learning_rate": 5.522642316338268e-06,
+      "loss": 0.0638,
+      "step": 442
+    },
+    {
+      "epoch": 5.15234375,
+      "grad_norm": 0.15060634911060333,
+      "learning_rate": 5.5022171648252265e-06,
+      "loss": 0.0489,
+      "step": 443
+    },
+    {
+      "epoch": 5.1640625,
+      "grad_norm": 0.1309528797864914,
+      "learning_rate": 5.4817835436163e-06,
+      "loss": 0.0561,
+      "step": 444
+    },
+    {
+      "epoch": 5.17578125,
+      "grad_norm": 0.14138580858707428,
+      "learning_rate": 5.46134179731651e-06,
+      "loss": 0.0556,
+      "step": 445
+    },
+    {
+      "epoch": 5.1875,
+      "grad_norm": 0.13814423978328705,
+      "learning_rate": 5.440892270667909e-06,
+      "loss": 0.0614,
+      "step": 446
+    },
+    {
+      "epoch": 5.19921875,
+      "grad_norm": 0.15363521873950958,
+      "learning_rate": 5.420435308543757e-06,
+      "loss": 0.0622,
+      "step": 447
+    },
+    {
+      "epoch": 5.2109375,
+      "grad_norm": 0.12922365963459015,
+      "learning_rate": 5.399971255942708e-06,
+      "loss": 0.0611,
+      "step": 448
+    },
+    {
+      "epoch": 5.22265625,
+      "grad_norm": 0.15916524827480316,
+      "learning_rate": 5.379500457983006e-06,
+      "loss": 0.0611,
+      "step": 449
+    },
+    {
+      "epoch": 5.234375,
+      "grad_norm": 0.14161787927150726,
+      "learning_rate": 5.359023259896638e-06,
+      "loss": 0.0463,
+      "step": 450
+    },
+    {
+      "epoch": 5.24609375,
+      "grad_norm": 0.14080356061458588,
+      "learning_rate": 5.338540007023538e-06,
+      "loss": 0.0574,
+      "step": 451
+    },
+    {
+      "epoch": 5.2578125,
+      "grad_norm": 0.14684629440307617,
+      "learning_rate": 5.318051044805745e-06,
+      "loss": 0.0508,
+      "step": 452
+    },
+    {
+      "epoch": 5.26953125,
+      "grad_norm": 0.14448480308055878,
+      "learning_rate": 5.297556718781588e-06,
+      "loss": 0.0649,
+      "step": 453
+    },
+    {
+      "epoch": 5.28125,
+      "grad_norm": 0.15748435258865356,
+      "learning_rate": 5.27705737457985e-06,
+      "loss": 0.0672,
+      "step": 454
+    },
+    {
+      "epoch": 5.29296875,
+      "grad_norm": 0.13277721405029297,
+      "learning_rate": 5.2565533579139484e-06,
+      "loss": 0.0474,
+      "step": 455
+    },
+    {
+      "epoch": 5.3046875,
+      "grad_norm": 0.14888741075992584,
+      "learning_rate": 5.236045014576098e-06,
+      "loss": 0.0699,
+      "step": 456
+    },
+    {
+      "epoch": 5.31640625,
+      "grad_norm": 0.1279619187116623,
+      "learning_rate": 5.2155326904314795e-06,
+      "loss": 0.0746,
+      "step": 457
+    },
+    {
+      "epoch": 5.328125,
+      "grad_norm": 0.17560450732707977,
+      "learning_rate": 5.1950167314124085e-06,
+      "loss": 0.058,
+      "step": 458
+    },
+    {
+      "epoch": 5.33984375,
+      "grad_norm": 0.15874771773815155,
+      "learning_rate": 5.174497483512506e-06,
+      "loss": 0.0575,
+      "step": 459
+    },
+    {
+      "epoch": 5.3515625,
+      "grad_norm": 0.13492193818092346,
+      "learning_rate": 5.153975292780852e-06,
+      "loss": 0.0469,
+      "step": 460
+    },
+    {
+      "epoch": 5.36328125,
+      "grad_norm": 0.14720965921878815,
+      "learning_rate": 5.133450505316162e-06,
+      "loss": 0.0605,
+      "step": 461
+    },
+    {
+      "epoch": 5.375,
+      "grad_norm": 0.1355433613061905,
+      "learning_rate": 5.112923467260941e-06,
+      "loss": 0.0509,
+      "step": 462
+    },
+    {
+      "epoch": 5.38671875,
+      "grad_norm": 0.15568669140338898,
+      "learning_rate": 5.09239452479565e-06,
+      "loss": 0.0687,
+      "step": 463
+    },
+    {
+      "epoch": 5.3984375,
+      "grad_norm": 0.15161432325839996,
+      "learning_rate": 5.071864024132868e-06,
+      "loss": 0.064,
+      "step": 464
+    },
+    {
+      "epoch": 5.41015625,
+      "grad_norm": 0.17019574344158173,
+      "learning_rate": 5.05133231151145e-06,
+      "loss": 0.0451,
+      "step": 465
+    },
+    {
+      "epoch": 5.421875,
+      "grad_norm": 0.11086393147706985,
+      "learning_rate": 5.030799733190694e-06,
+      "loss": 0.0559,
+      "step": 466
+    },
+    {
+      "epoch": 5.43359375,
+      "grad_norm": 0.15558047592639923,
+      "learning_rate": 5.010266635444495e-06,
+      "loss": 0.0585,
+      "step": 467
+    },
+    {
+      "epoch": 5.4453125,
+      "grad_norm": 0.148808091878891,
+      "learning_rate": 4.989733364555507e-06,
+      "loss": 0.0593,
+      "step": 468
+    },
+    {
+      "epoch": 5.45703125,
+      "grad_norm": 0.158802792429924,
+      "learning_rate": 4.9692002668093075e-06,
+      "loss": 0.0525,
+      "step": 469
+    },
+    {
+      "epoch": 5.46875,
+      "grad_norm": 0.12689101696014404,
+      "learning_rate": 4.948667688488552e-06,
+      "loss": 0.0543,
+      "step": 470
+    },
+    {
+      "epoch": 5.48046875,
+      "grad_norm": 0.13123546540737152,
+      "learning_rate": 4.928135975867134e-06,
+      "loss": 0.0617,
+      "step": 471
+    },
+    {
+      "epoch": 5.4921875,
+      "grad_norm": 0.16200099885463715,
+      "learning_rate": 4.907605475204352e-06,
+      "loss": 0.0588,
+      "step": 472
+    },
+    {
+      "epoch": 5.50390625,
+      "grad_norm": 0.16697384417057037,
+      "learning_rate": 4.887076532739061e-06,
+      "loss": 0.0578,
+      "step": 473
+    },
+    {
+      "epoch": 5.515625,
+      "grad_norm": 0.14646796882152557,
+      "learning_rate": 4.866549494683839e-06,
+      "loss": 0.0781,
+      "step": 474
+    },
+    {
+      "epoch": 5.52734375,
+      "grad_norm": 0.14703866839408875,
+      "learning_rate": 4.846024707219149e-06,
+      "loss": 0.0597,
+      "step": 475
+    },
+    {
+      "epoch": 5.5390625,
+      "grad_norm": 0.1428564488887787,
+      "learning_rate": 4.825502516487497e-06,
+      "loss": 0.0497,
+      "step": 476
+    },
+    {
+      "epoch": 5.55078125,
+      "grad_norm": 0.13470599055290222,
+      "learning_rate": 4.804983268587593e-06,
+      "loss": 0.0574,
+      "step": 477
+    },
+    {
+      "epoch": 5.5625,
+      "grad_norm": 0.13724403083324432,
+      "learning_rate": 4.784467309568524e-06,
+      "loss": 0.0626,
+      "step": 478
+    },
+    {
+      "epoch": 5.57421875,
+      "grad_norm": 0.13949063420295715,
+      "learning_rate": 4.7639549854239045e-06,
+      "loss": 0.0577,
+      "step": 479
+    },
+    {
+      "epoch": 5.5859375,
+      "grad_norm": 0.14538227021694183,
+      "learning_rate": 4.7434466420860515e-06,
+      "loss": 0.054,
+      "step": 480
+    },
+    {
+      "epoch": 5.59765625,
+      "grad_norm": 0.13742738962173462,
+      "learning_rate": 4.7229426254201504e-06,
+      "loss": 0.0541,
+      "step": 481
+    },
+    {
+      "epoch": 5.609375,
+      "grad_norm": 0.1454109251499176,
+      "learning_rate": 4.702443281218413e-06,
+      "loss": 0.0492,
+      "step": 482
+    },
+    {
+      "epoch": 5.62109375,
+      "grad_norm": 0.1340155303478241,
+      "learning_rate": 4.681948955194256e-06,
+      "loss": 0.0519,
+      "step": 483
+    },
+    {
+      "epoch": 5.6328125,
+      "grad_norm": 0.11645597219467163,
+      "learning_rate": 4.661459992976463e-06,
+      "loss": 0.0772,
+      "step": 484
+    },
+    {
+      "epoch": 5.64453125,
+      "grad_norm": 0.1487962156534195,
+      "learning_rate": 4.640976740103363e-06,
+      "loss": 0.0461,
+      "step": 485
+    },
+    {
+      "epoch": 5.65625,
+      "grad_norm": 0.1314113289117813,
+      "learning_rate": 4.620499542016996e-06,
+      "loss": 0.0498,
+      "step": 486
+    },
+    {
+      "epoch": 5.66796875,
+      "grad_norm": 0.11778878420591354,
+      "learning_rate": 4.6000287440572925e-06,
+      "loss": 0.0553,
+      "step": 487
+    },
+    {
+      "epoch": 5.6796875,
+      "grad_norm": 0.14744260907173157,
+      "learning_rate": 4.579564691456245e-06,
+      "loss": 0.0582,
+      "step": 488
+    },
+    {
+      "epoch": 5.69140625,
+      "grad_norm": 0.1958065778017044,
+      "learning_rate": 4.5591077293320925e-06,
+      "loss": 0.0481,
+      "step": 489
+    },
+    {
+      "epoch": 5.703125,
+      "grad_norm": 0.11204442381858826,
+      "learning_rate": 4.53865820268349e-06,
+      "loss": 0.0417,
+      "step": 490
+    },
+    {
+      "epoch": 5.71484375,
+      "grad_norm": 0.14066173136234283,
+      "learning_rate": 4.5182164563837015e-06,
+      "loss": 0.042,
+      "step": 491
+    },
+    {
+      "epoch": 5.7265625,
+      "grad_norm": 0.11793823540210724,
+      "learning_rate": 4.497782835174775e-06,
+      "loss": 0.0536,
+      "step": 492
+    },
+    {
+      "epoch": 5.73828125,
+      "grad_norm": 0.12355687469244003,
+      "learning_rate": 4.477357683661734e-06,
+      "loss": 0.0621,
+      "step": 493
+    },
+    {
+      "epoch": 5.75,
+      "grad_norm": 0.16047550737857819,
+      "learning_rate": 4.456941346306761e-06,
+      "loss": 0.0603,
+      "step": 494
+    },
+    {
+      "epoch": 5.76171875,
+      "grad_norm": 0.1558404564857483,
+      "learning_rate": 4.436534167423395e-06,
+      "loss": 0.0578,
+      "step": 495
+    },
+    {
+      "epoch": 5.7734375,
+      "grad_norm": 0.1236644983291626,
+      "learning_rate": 4.41613649117072e-06,
+      "loss": 0.0455,
+      "step": 496
+    },
+    {
+      "epoch": 5.78515625,
+      "grad_norm": 0.10916972905397415,
+      "learning_rate": 4.395748661547558e-06,
+      "loss": 0.0505,
+      "step": 497
+    },
+    {
+      "epoch": 5.796875,
+      "grad_norm": 0.13890743255615234,
+      "learning_rate": 4.375371022386677e-06,
+      "loss": 0.0476,
+      "step": 498
+    },
+    {
+      "epoch": 5.80859375,
+      "grad_norm": 0.14354218542575836,
+      "learning_rate": 4.355003917348985e-06,
+      "loss": 0.055,
+      "step": 499
+    },
+    {
+      "epoch": 5.8203125,
+      "grad_norm": 0.11266707628965378,
+      "learning_rate": 4.334647689917734e-06,
+      "loss": 0.0429,
+      "step": 500
+    },
+    {
+      "epoch": 5.83203125,
+      "grad_norm": 0.12821117043495178,
+      "learning_rate": 4.314302683392729e-06,
+      "loss": 0.0497,
+      "step": 501
+    },
+    {
+      "epoch": 5.84375,
+      "grad_norm": 0.1462571769952774,
+      "learning_rate": 4.293969240884545e-06,
+      "loss": 0.0563,
+      "step": 502
+    },
+    {
+      "epoch": 5.85546875,
+      "grad_norm": 0.12532669305801392,
+      "learning_rate": 4.273647705308726e-06,
+      "loss": 0.0745,
+      "step": 503
+    },
+    {
+      "epoch": 5.8671875,
+      "grad_norm": 0.4210283160209656,
+      "learning_rate": 4.253338419380016e-06,
+      "loss": 0.0481,
+      "step": 504
+    },
+    {
+      "epoch": 5.87890625,
+      "grad_norm": 0.13547341525554657,
+      "learning_rate": 4.233041725606573e-06,
+      "loss": 0.0609,
+      "step": 505
+    },
+    {
+      "epoch": 5.890625,
+      "grad_norm": 0.11836741119623184,
+      "learning_rate": 4.212757966284191e-06,
+      "loss": 0.051,
+      "step": 506
+    },
+    {
+      "epoch": 5.90234375,
+      "grad_norm": 0.13804543018341064,
+      "learning_rate": 4.192487483490532e-06,
+      "loss": 0.0619,
+      "step": 507
+    },
+    {
+      "epoch": 5.9140625,
+      "grad_norm": 0.1685003638267517,
+      "learning_rate": 4.17223061907935e-06,
+      "loss": 0.0515,
+      "step": 508
+    },
+    {
+      "epoch": 5.92578125,
+      "grad_norm": 0.14643150568008423,
+      "learning_rate": 4.151987714674737e-06,
+      "loss": 0.0538,
+      "step": 509
+    },
+    {
+      "epoch": 5.9375,
+      "grad_norm": 0.11812538653612137,
+      "learning_rate": 4.131759111665349e-06,
+      "loss": 0.0417,
+      "step": 510
+    },
+    {
+      "epoch": 5.94921875,
+      "grad_norm": 0.12235013395547867,
+      "learning_rate": 4.111545151198657e-06,
+      "loss": 0.0488,
+      "step": 511
+    },
+    {
+      "epoch": 5.9609375,
+      "grad_norm": 0.14277981221675873,
+      "learning_rate": 4.091346174175193e-06,
+      "loss": 0.052,
+      "step": 512
+    },
+    {
+      "epoch": 5.97265625,
+      "grad_norm": 0.13800203800201416,
+      "learning_rate": 4.071162521242796e-06,
+      "loss": 0.0462,
+      "step": 513
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 850,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 171,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 109706402504704.0,
+  "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:cabd7eea3a1004f1d2ad68be4ea28e77619033566d7b279da3d607abfda7062d
+size 7672