First model init

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": "/data3/bytedance/users/booommer/megatron/mp/",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 131072,
+  "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": "4.46.0",
+  "use_cache": false,
+  "use_mrope": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

configuration.json

@@ -0,0 +1 @@
+{}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.46.0"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:232e230035b978ae474a3f2c8e70503f02032d59233740ac2156b2aac53007b9
+size 4877660776

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa423cb45fb48595c4785d7a000a0c5f8d6998c05634b2c5539909c2e6ab4ab6
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c60ba04ede1dede2a40a69dd199b5730882cad98d525001b3388a5b599656ef
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:feaa4d44fc742162ddaf92b6444025dc192f995639b25e37ffad691388ca491f
+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": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "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 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 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": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null,
+  "use_fast": true
+}

trainer_state.json

@@ -0,0 +1,3750 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9133519673188562,
+  "eval_steps": 500,
+  "global_step": 531,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0017200602021070737,
+      "grad_norm": 9.863033294677734,
+      "learning_rate": 1.0000000000000001e-07,
+      "loss": 46.5684,
+      "step": 1
+    },
+    {
+      "epoch": 0.0034401204042141475,
+      "grad_norm": 9.501362800598145,
+      "learning_rate": 2.0000000000000002e-07,
+      "loss": 46.7594,
+      "step": 2
+    },
+    {
+      "epoch": 0.005160180606321221,
+      "grad_norm": 10.306024551391602,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 47.7285,
+      "step": 3
+    },
+    {
+      "epoch": 0.006880240808428295,
+      "grad_norm": 10.35928726196289,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 46.2416,
+      "step": 4
+    },
+    {
+      "epoch": 0.008600301010535369,
+      "grad_norm": 10.284590721130371,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 46.4638,
+      "step": 5
+    },
+    {
+      "epoch": 0.010320361212642442,
+      "grad_norm": 9.508557319641113,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 47.0967,
+      "step": 6
+    },
+    {
+      "epoch": 0.012040421414749517,
+      "grad_norm": 9.53107738494873,
+      "learning_rate": 7.000000000000001e-07,
+      "loss": 46.8945,
+      "step": 7
+    },
+    {
+      "epoch": 0.01376048161685659,
+      "grad_norm": 8.917387008666992,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 46.0318,
+      "step": 8
+    },
+    {
+      "epoch": 0.015480541818963663,
+      "grad_norm": 9.046576499938965,
+      "learning_rate": 9.000000000000001e-07,
+      "loss": 46.7626,
+      "step": 9
+    },
+    {
+      "epoch": 0.017200602021070738,
+      "grad_norm": 8.346121788024902,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 46.4023,
+      "step": 10
+    },
+    {
+      "epoch": 0.01892066222317781,
+      "grad_norm": 7.368544578552246,
+      "learning_rate": 1.1e-06,
+      "loss": 46.9599,
+      "step": 11
+    },
+    {
+      "epoch": 0.020640722425284884,
+      "grad_norm": 7.151632308959961,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 46.8859,
+      "step": 12
+    },
+    {
+      "epoch": 0.022360782627391957,
+      "grad_norm": 6.858343124389648,
+      "learning_rate": 1.3e-06,
+      "loss": 47.0505,
+      "step": 13
+    },
+    {
+      "epoch": 0.024080842829499034,
+      "grad_norm": 6.4683918952941895,
+      "learning_rate": 1.4000000000000001e-06,
+      "loss": 46.6439,
+      "step": 14
+    },
+    {
+      "epoch": 0.025800903031606107,
+      "grad_norm": 8.405121803283691,
+      "learning_rate": 1.5e-06,
+      "loss": 46.8736,
+      "step": 15
+    },
+    {
+      "epoch": 0.02752096323371318,
+      "grad_norm": 10.191279411315918,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 46.2945,
+      "step": 16
+    },
+    {
+      "epoch": 0.029241023435820253,
+      "grad_norm": 10.24738597869873,
+      "learning_rate": 1.7000000000000002e-06,
+      "loss": 46.0618,
+      "step": 17
+    },
+    {
+      "epoch": 0.030961083637927326,
+      "grad_norm": 9.236139297485352,
+      "learning_rate": 1.8000000000000001e-06,
+      "loss": 47.0773,
+      "step": 18
+    },
+    {
+      "epoch": 0.0326811438400344,
+      "grad_norm": 8.437068939208984,
+      "learning_rate": 1.9000000000000002e-06,
+      "loss": 46.6797,
+      "step": 19
+    },
+    {
+      "epoch": 0.034401204042141476,
+      "grad_norm": 8.35542106628418,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 46.8348,
+      "step": 20
+    },
+    {
+      "epoch": 0.03612126424424855,
+      "grad_norm": 9.210644721984863,
+      "learning_rate": 2.1000000000000002e-06,
+      "loss": 46.7521,
+      "step": 21
+    },
+    {
+      "epoch": 0.03784132444635562,
+      "grad_norm": 9.216105461120605,
+      "learning_rate": 2.2e-06,
+      "loss": 46.4541,
+      "step": 22
+    },
+    {
+      "epoch": 0.0395613846484627,
+      "grad_norm": 8.33311939239502,
+      "learning_rate": 2.3000000000000004e-06,
+      "loss": 46.7192,
+      "step": 23
+    },
+    {
+      "epoch": 0.04128144485056977,
+      "grad_norm": 7.9267449378967285,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 46.7559,
+      "step": 24
+    },
+    {
+      "epoch": 0.043001505052676844,
+      "grad_norm": 7.363073348999023,
+      "learning_rate": 2.5e-06,
+      "loss": 45.9578,
+      "step": 25
+    },
+    {
+      "epoch": 0.044721565254783914,
+      "grad_norm": 7.073836326599121,
+      "learning_rate": 2.6e-06,
+      "loss": 45.9155,
+      "step": 26
+    },
+    {
+      "epoch": 0.04644162545689099,
+      "grad_norm": 7.0409088134765625,
+      "learning_rate": 2.7000000000000004e-06,
+      "loss": 46.6551,
+      "step": 27
+    },
+    {
+      "epoch": 0.04816168565899807,
+      "grad_norm": 6.685385227203369,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 46.3889,
+      "step": 28
+    },
+    {
+      "epoch": 0.04988174586110514,
+      "grad_norm": 6.540524959564209,
+      "learning_rate": 2.9e-06,
+      "loss": 47.0662,
+      "step": 29
+    },
+    {
+      "epoch": 0.05160180606321221,
+      "grad_norm": 6.627730846405029,
+      "learning_rate": 3e-06,
+      "loss": 46.101,
+      "step": 30
+    },
+    {
+      "epoch": 0.05332186626531928,
+      "grad_norm": 7.054595470428467,
+      "learning_rate": 3.1000000000000004e-06,
+      "loss": 46.7379,
+      "step": 31
+    },
+    {
+      "epoch": 0.05504192646742636,
+      "grad_norm": 6.73618221282959,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 46.3172,
+      "step": 32
+    },
+    {
+      "epoch": 0.056761986669533436,
+      "grad_norm": 5.943539619445801,
+      "learning_rate": 3.3000000000000006e-06,
+      "loss": 47.0245,
+      "step": 33
+    },
+    {
+      "epoch": 0.058482046871640506,
+      "grad_norm": 6.056912899017334,
+      "learning_rate": 3.4000000000000005e-06,
+      "loss": 46.0208,
+      "step": 34
+    },
+    {
+      "epoch": 0.06020210707374758,
+      "grad_norm": 5.7798309326171875,
+      "learning_rate": 3.5e-06,
+      "loss": 46.0746,
+      "step": 35
+    },
+    {
+      "epoch": 0.06192216727585465,
+      "grad_norm": 5.896692276000977,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 46.4911,
+      "step": 36
+    },
+    {
+      "epoch": 0.06364222747796174,
+      "grad_norm": 5.496098518371582,
+      "learning_rate": 3.7e-06,
+      "loss": 45.9957,
+      "step": 37
+    },
+    {
+      "epoch": 0.0653622876800688,
+      "grad_norm": 5.253308296203613,
+      "learning_rate": 3.8000000000000005e-06,
+      "loss": 45.7145,
+      "step": 38
+    },
+    {
+      "epoch": 0.06708234788217587,
+      "grad_norm": 5.340756416320801,
+      "learning_rate": 3.900000000000001e-06,
+      "loss": 46.7068,
+      "step": 39
+    },
+    {
+      "epoch": 0.06880240808428295,
+      "grad_norm": 5.312371730804443,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 46.4172,
+      "step": 40
+    },
+    {
+      "epoch": 0.07052246828639003,
+      "grad_norm": 5.484511375427246,
+      "learning_rate": 4.1e-06,
+      "loss": 45.6433,
+      "step": 41
+    },
+    {
+      "epoch": 0.0722425284884971,
+      "grad_norm": 5.260024547576904,
+      "learning_rate": 4.2000000000000004e-06,
+      "loss": 46.1259,
+      "step": 42
+    },
+    {
+      "epoch": 0.07396258869060417,
+      "grad_norm": 5.4440999031066895,
+      "learning_rate": 4.3e-06,
+      "loss": 46.2947,
+      "step": 43
+    },
+    {
+      "epoch": 0.07568264889271124,
+      "grad_norm": 5.153471946716309,
+      "learning_rate": 4.4e-06,
+      "loss": 46.2882,
+      "step": 44
+    },
+    {
+      "epoch": 0.07740270909481832,
+      "grad_norm": 5.6796488761901855,
+      "learning_rate": 4.5e-06,
+      "loss": 46.3314,
+      "step": 45
+    },
+    {
+      "epoch": 0.0791227692969254,
+      "grad_norm": 5.253461837768555,
+      "learning_rate": 4.600000000000001e-06,
+      "loss": 46.4205,
+      "step": 46
+    },
+    {
+      "epoch": 0.08084282949903247,
+      "grad_norm": 5.475714683532715,
+      "learning_rate": 4.7e-06,
+      "loss": 45.9359,
+      "step": 47
+    },
+    {
+      "epoch": 0.08256288970113954,
+      "grad_norm": 5.880488872528076,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 46.0665,
+      "step": 48
+    },
+    {
+      "epoch": 0.08428294990324661,
+      "grad_norm": 5.69149923324585,
+      "learning_rate": 4.9000000000000005e-06,
+      "loss": 45.8811,
+      "step": 49
+    },
+    {
+      "epoch": 0.08600301010535369,
+      "grad_norm": 5.086974143981934,
+      "learning_rate": 5e-06,
+      "loss": 46.1759,
+      "step": 50
+    },
+    {
+      "epoch": 0.08772307030746077,
+      "grad_norm": 5.902041435241699,
+      "learning_rate": 5.1e-06,
+      "loss": 46.4326,
+      "step": 51
+    },
+    {
+      "epoch": 0.08944313050956783,
+      "grad_norm": 5.740163803100586,
+      "learning_rate": 5.2e-06,
+      "loss": 45.607,
+      "step": 52
+    },
+    {
+      "epoch": 0.0911631907116749,
+      "grad_norm": 5.547687530517578,
+      "learning_rate": 5.300000000000001e-06,
+      "loss": 45.8864,
+      "step": 53
+    },
+    {
+      "epoch": 0.09288325091378198,
+      "grad_norm": 6.20143461227417,
+      "learning_rate": 5.400000000000001e-06,
+      "loss": 45.7776,
+      "step": 54
+    },
+    {
+      "epoch": 0.09460331111588906,
+      "grad_norm": 5.544395446777344,
+      "learning_rate": 5.500000000000001e-06,
+      "loss": 46.3334,
+      "step": 55
+    },
+    {
+      "epoch": 0.09632337131799613,
+      "grad_norm": 5.583694934844971,
+      "learning_rate": 5.600000000000001e-06,
+      "loss": 45.5733,
+      "step": 56
+    },
+    {
+      "epoch": 0.0980434315201032,
+      "grad_norm": 5.751343250274658,
+      "learning_rate": 5.7e-06,
+      "loss": 45.2805,
+      "step": 57
+    },
+    {
+      "epoch": 0.09976349172221027,
+      "grad_norm": 6.024663925170898,
+      "learning_rate": 5.8e-06,
+      "loss": 45.6569,
+      "step": 58
+    },
+    {
+      "epoch": 0.10148355192431735,
+      "grad_norm": 5.834673881530762,
+      "learning_rate": 5.9e-06,
+      "loss": 45.7854,
+      "step": 59
+    },
+    {
+      "epoch": 0.10320361212642443,
+      "grad_norm": 6.796127796173096,
+      "learning_rate": 6e-06,
+      "loss": 46.4184,
+      "step": 60
+    },
+    {
+      "epoch": 0.1049236723285315,
+      "grad_norm": 6.303890705108643,
+      "learning_rate": 6.1e-06,
+      "loss": 46.0877,
+      "step": 61
+    },
+    {
+      "epoch": 0.10664373253063857,
+      "grad_norm": 6.320569038391113,
+      "learning_rate": 6.200000000000001e-06,
+      "loss": 45.8802,
+      "step": 62
+    },
+    {
+      "epoch": 0.10836379273274564,
+      "grad_norm": 6.015869617462158,
+      "learning_rate": 6.300000000000001e-06,
+      "loss": 45.6733,
+      "step": 63
+    },
+    {
+      "epoch": 0.11008385293485272,
+      "grad_norm": 6.314846515655518,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 45.8531,
+      "step": 64
+    },
+    {
+      "epoch": 0.1118039131369598,
+      "grad_norm": 6.106888294219971,
+      "learning_rate": 6.5000000000000004e-06,
+      "loss": 45.2966,
+      "step": 65
+    },
+    {
+      "epoch": 0.11352397333906687,
+      "grad_norm": 6.093003273010254,
+      "learning_rate": 6.600000000000001e-06,
+      "loss": 46.0552,
+      "step": 66
+    },
+    {
+      "epoch": 0.11524403354117393,
+      "grad_norm": 5.634897708892822,
+      "learning_rate": 6.700000000000001e-06,
+      "loss": 45.3367,
+      "step": 67
+    },
+    {
+      "epoch": 0.11696409374328101,
+      "grad_norm": 6.261721134185791,
+      "learning_rate": 6.800000000000001e-06,
+      "loss": 45.7655,
+      "step": 68
+    },
+    {
+      "epoch": 0.11868415394538809,
+      "grad_norm": 5.926329135894775,
+      "learning_rate": 6.9e-06,
+      "loss": 45.1391,
+      "step": 69
+    },
+    {
+      "epoch": 0.12040421414749516,
+      "grad_norm": 5.887923717498779,
+      "learning_rate": 7e-06,
+      "loss": 45.43,
+      "step": 70
+    },
+    {
+      "epoch": 0.12212427434960224,
+      "grad_norm": 5.79194450378418,
+      "learning_rate": 7.100000000000001e-06,
+      "loss": 45.763,
+      "step": 71
+    },
+    {
+      "epoch": 0.1238443345517093,
+      "grad_norm": 6.2670722007751465,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 45.137,
+      "step": 72
+    },
+    {
+      "epoch": 0.1255643947538164,
+      "grad_norm": 5.60247278213501,
+      "learning_rate": 7.3e-06,
+      "loss": 45.7123,
+      "step": 73
+    },
+    {
+      "epoch": 0.12728445495592347,
+      "grad_norm": 7.346188545227051,
+      "learning_rate": 7.4e-06,
+      "loss": 45.5004,
+      "step": 74
+    },
+    {
+      "epoch": 0.12900451515803052,
+      "grad_norm": 6.078243255615234,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 45.5731,
+      "step": 75
+    },
+    {
+      "epoch": 0.1307245753601376,
+      "grad_norm": 7.248181343078613,
+      "learning_rate": 7.600000000000001e-06,
+      "loss": 45.7506,
+      "step": 76
+    },
+    {
+      "epoch": 0.13244463556224467,
+      "grad_norm": 5.837612628936768,
+      "learning_rate": 7.7e-06,
+      "loss": 45.7486,
+      "step": 77
+    },
+    {
+      "epoch": 0.13416469576435175,
+      "grad_norm": 7.661535739898682,
+      "learning_rate": 7.800000000000002e-06,
+      "loss": 45.5743,
+      "step": 78
+    },
+    {
+      "epoch": 0.13588475596645883,
+      "grad_norm": 5.925168991088867,
+      "learning_rate": 7.9e-06,
+      "loss": 45.0878,
+      "step": 79
+    },
+    {
+      "epoch": 0.1376048161685659,
+      "grad_norm": 6.47566032409668,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 45.4567,
+      "step": 80
+    },
+    {
+      "epoch": 0.13932487637067298,
+      "grad_norm": 5.9641852378845215,
+      "learning_rate": 8.1e-06,
+      "loss": 45.0619,
+      "step": 81
+    },
+    {
+      "epoch": 0.14104493657278006,
+      "grad_norm": 6.4052324295043945,
+      "learning_rate": 8.2e-06,
+      "loss": 45.9375,
+      "step": 82
+    },
+    {
+      "epoch": 0.14276499677488713,
+      "grad_norm": 6.149839401245117,
+      "learning_rate": 8.3e-06,
+      "loss": 45.5695,
+      "step": 83
+    },
+    {
+      "epoch": 0.1444850569769942,
+      "grad_norm": 6.546520233154297,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 45.8453,
+      "step": 84
+    },
+    {
+      "epoch": 0.14620511717910126,
+      "grad_norm": 6.221002578735352,
+      "learning_rate": 8.5e-06,
+      "loss": 44.9717,
+      "step": 85
+    },
+    {
+      "epoch": 0.14792517738120833,
+      "grad_norm": 6.772427558898926,
+      "learning_rate": 8.6e-06,
+      "loss": 45.8218,
+      "step": 86
+    },
+    {
+      "epoch": 0.1496452375833154,
+      "grad_norm": 7.193835258483887,
+      "learning_rate": 8.700000000000001e-06,
+      "loss": 45.4445,
+      "step": 87
+    },
+    {
+      "epoch": 0.1513652977854225,
+      "grad_norm": 6.1274895668029785,
+      "learning_rate": 8.8e-06,
+      "loss": 45.127,
+      "step": 88
+    },
+    {
+      "epoch": 0.15308535798752956,
+      "grad_norm": 7.984500885009766,
+      "learning_rate": 8.900000000000001e-06,
+      "loss": 44.9928,
+      "step": 89
+    },
+    {
+      "epoch": 0.15480541818963664,
+      "grad_norm": 6.346322059631348,
+      "learning_rate": 9e-06,
+      "loss": 45.4616,
+      "step": 90
+    },
+    {
+      "epoch": 0.15652547839174372,
+      "grad_norm": 7.460362911224365,
+      "learning_rate": 9.100000000000001e-06,
+      "loss": 45.2488,
+      "step": 91
+    },
+    {
+      "epoch": 0.1582455385938508,
+      "grad_norm": 7.510838508605957,
+      "learning_rate": 9.200000000000002e-06,
+      "loss": 45.0087,
+      "step": 92
+    },
+    {
+      "epoch": 0.15996559879595787,
+      "grad_norm": 6.567174911499023,
+      "learning_rate": 9.3e-06,
+      "loss": 44.3216,
+      "step": 93
+    },
+    {
+      "epoch": 0.16168565899806495,
+      "grad_norm": 8.357160568237305,
+      "learning_rate": 9.4e-06,
+      "loss": 45.1102,
+      "step": 94
+    },
+    {
+      "epoch": 0.163405719200172,
+      "grad_norm": 6.190718650817871,
+      "learning_rate": 9.5e-06,
+      "loss": 45.0187,
+      "step": 95
+    },
+    {
+      "epoch": 0.16512577940227907,
+      "grad_norm": 6.47848653793335,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 45.9968,
+      "step": 96
+    },
+    {
+      "epoch": 0.16684583960438615,
+      "grad_norm": 6.271050453186035,
+      "learning_rate": 9.7e-06,
+      "loss": 45.6172,
+      "step": 97
+    },
+    {
+      "epoch": 0.16856589980649322,
+      "grad_norm": 7.013180255889893,
+      "learning_rate": 9.800000000000001e-06,
+      "loss": 45.1628,
+      "step": 98
+    },
+    {
+      "epoch": 0.1702859600086003,
+      "grad_norm": 6.151601314544678,
+      "learning_rate": 9.9e-06,
+      "loss": 45.3331,
+      "step": 99
+    },
+    {
+      "epoch": 0.17200602021070738,
+      "grad_norm": 6.522064685821533,
+      "learning_rate": 1e-05,
+      "loss": 45.01,
+      "step": 100
+    },
+    {
+      "epoch": 0.17372608041281445,
+      "grad_norm": 6.936015605926514,
+      "learning_rate": 9.999990859614876e-06,
+      "loss": 44.7838,
+      "step": 101
+    },
+    {
+      "epoch": 0.17544614061492153,
+      "grad_norm": 6.696622371673584,
+      "learning_rate": 9.99996343849292e-06,
+      "loss": 44.4096,
+      "step": 102
+    },
+    {
+      "epoch": 0.1771662008170286,
+      "grad_norm": 6.770718574523926,
+      "learning_rate": 9.999917736734387e-06,
+      "loss": 44.6306,
+      "step": 103
+    },
+    {
+      "epoch": 0.17888626101913566,
+      "grad_norm": 8.121861457824707,
+      "learning_rate": 9.999853754506375e-06,
+      "loss": 45.3421,
+      "step": 104
+    },
+    {
+      "epoch": 0.18060632122124273,
+      "grad_norm": 7.101470947265625,
+      "learning_rate": 9.999771492042807e-06,
+      "loss": 45.415,
+      "step": 105
+    },
+    {
+      "epoch": 0.1823263814233498,
+      "grad_norm": 8.22966194152832,
+      "learning_rate": 9.99967094964445e-06,
+      "loss": 45.3149,
+      "step": 106
+    },
+    {
+      "epoch": 0.18404644162545689,
+      "grad_norm": 10.975841522216797,
+      "learning_rate": 9.9995521276789e-06,
+      "loss": 45.2939,
+      "step": 107
+    },
+    {
+      "epoch": 0.18576650182756396,
+      "grad_norm": 7.771969795227051,
+      "learning_rate": 9.999415026580592e-06,
+      "loss": 45.2433,
+      "step": 108
+    },
+    {
+      "epoch": 0.18748656202967104,
+      "grad_norm": 10.387533187866211,
+      "learning_rate": 9.999259646850787e-06,
+      "loss": 45.2594,
+      "step": 109
+    },
+    {
+      "epoch": 0.18920662223177812,
+      "grad_norm": 8.546263694763184,
+      "learning_rate": 9.999085989057578e-06,
+      "loss": 45.0587,
+      "step": 110
+    },
+    {
+      "epoch": 0.1909266824338852,
+      "grad_norm": 9.245797157287598,
+      "learning_rate": 9.998894053835883e-06,
+      "loss": 45.2273,
+      "step": 111
+    },
+    {
+      "epoch": 0.19264674263599227,
+      "grad_norm": 9.295574188232422,
+      "learning_rate": 9.998683841887449e-06,
+      "loss": 45.5471,
+      "step": 112
+    },
+    {
+      "epoch": 0.19436680283809935,
+      "grad_norm": 8.650074005126953,
+      "learning_rate": 9.99845535398084e-06,
+      "loss": 45.2802,
+      "step": 113
+    },
+    {
+      "epoch": 0.1960868630402064,
+      "grad_norm": 11.102032661437988,
+      "learning_rate": 9.998208590951449e-06,
+      "loss": 45.3066,
+      "step": 114
+    },
+    {
+      "epoch": 0.19780692324231347,
+      "grad_norm": 6.938677787780762,
+      "learning_rate": 9.99794355370147e-06,
+      "loss": 45.1454,
+      "step": 115
+    },
+    {
+      "epoch": 0.19952698344442055,
+      "grad_norm": 16.639892578125,
+      "learning_rate": 9.997660243199928e-06,
+      "loss": 45.6125,
+      "step": 116
+    },
+    {
+      "epoch": 0.20124704364652762,
+      "grad_norm": 11.0328369140625,
+      "learning_rate": 9.99735866048265e-06,
+      "loss": 45.4087,
+      "step": 117
+    },
+    {
+      "epoch": 0.2029671038486347,
+      "grad_norm": 13.419515609741211,
+      "learning_rate": 9.997038806652264e-06,
+      "loss": 45.8139,
+      "step": 118
+    },
+    {
+      "epoch": 0.20468716405074178,
+      "grad_norm": 13.69460678100586,
+      "learning_rate": 9.996700682878206e-06,
+      "loss": 45.304,
+      "step": 119
+    },
+    {
+      "epoch": 0.20640722425284885,
+      "grad_norm": 7.643240928649902,
+      "learning_rate": 9.996344290396713e-06,
+      "loss": 45.1699,
+      "step": 120
+    },
+    {
+      "epoch": 0.20812728445495593,
+      "grad_norm": 9.692140579223633,
+      "learning_rate": 9.995969630510805e-06,
+      "loss": 44.7309,
+      "step": 121
+    },
+    {
+      "epoch": 0.209847344657063,
+      "grad_norm": 7.550359725952148,
+      "learning_rate": 9.995576704590299e-06,
+      "loss": 44.6299,
+      "step": 122
+    },
+    {
+      "epoch": 0.21156740485917008,
+      "grad_norm": 6.425361156463623,
+      "learning_rate": 9.995165514071793e-06,
+      "loss": 44.8296,
+      "step": 123
+    },
+    {
+      "epoch": 0.21328746506127713,
+      "grad_norm": 8.089837074279785,
+      "learning_rate": 9.994736060458665e-06,
+      "loss": 45.2622,
+      "step": 124
+    },
+    {
+      "epoch": 0.2150075252633842,
+      "grad_norm": 6.491065979003906,
+      "learning_rate": 9.994288345321059e-06,
+      "loss": 44.3288,
+      "step": 125
+    },
+    {
+      "epoch": 0.21672758546549128,
+      "grad_norm": 5.423704147338867,
+      "learning_rate": 9.993822370295892e-06,
+      "loss": 44.7233,
+      "step": 126
+    },
+    {
+      "epoch": 0.21844764566759836,
+      "grad_norm": 7.901766777038574,
+      "learning_rate": 9.993338137086841e-06,
+      "loss": 45.0159,
+      "step": 127
+    },
+    {
+      "epoch": 0.22016770586970544,
+      "grad_norm": 7.50944709777832,
+      "learning_rate": 9.992835647464339e-06,
+      "loss": 44.8785,
+      "step": 128
+    },
+    {
+      "epoch": 0.22188776607181251,
+      "grad_norm": 7.400681972503662,
+      "learning_rate": 9.992314903265561e-06,
+      "loss": 45.3053,
+      "step": 129
+    },
+    {
+      "epoch": 0.2236078262739196,
+      "grad_norm": 8.214972496032715,
+      "learning_rate": 9.991775906394434e-06,
+      "loss": 44.8066,
+      "step": 130
+    },
+    {
+      "epoch": 0.22532788647602667,
+      "grad_norm": 7.00181770324707,
+      "learning_rate": 9.991218658821609e-06,
+      "loss": 44.5013,
+      "step": 131
+    },
+    {
+      "epoch": 0.22704794667813374,
+      "grad_norm": 6.811863422393799,
+      "learning_rate": 9.990643162584467e-06,
+      "loss": 44.7327,
+      "step": 132
+    },
+    {
+      "epoch": 0.22876800688024082,
+      "grad_norm": 8.760518074035645,
+      "learning_rate": 9.99004941978712e-06,
+      "loss": 44.6706,
+      "step": 133
+    },
+    {
+      "epoch": 0.23048806708234787,
+      "grad_norm": 7.204503536224365,
+      "learning_rate": 9.989437432600373e-06,
+      "loss": 44.8113,
+      "step": 134
+    },
+    {
+      "epoch": 0.23220812728445495,
+      "grad_norm": 8.635207176208496,
+      "learning_rate": 9.988807203261752e-06,
+      "loss": 45.3002,
+      "step": 135
+    },
+    {
+      "epoch": 0.23392818748656202,
+      "grad_norm": 7.5076823234558105,
+      "learning_rate": 9.988158734075468e-06,
+      "loss": 44.919,
+      "step": 136
+    },
+    {
+      "epoch": 0.2356482476886691,
+      "grad_norm": 8.329002380371094,
+      "learning_rate": 9.98749202741243e-06,
+      "loss": 44.9676,
+      "step": 137
+    },
+    {
+      "epoch": 0.23736830789077618,
+      "grad_norm": 7.5155158042907715,
+      "learning_rate": 9.986807085710213e-06,
+      "loss": 45.167,
+      "step": 138
+    },
+    {
+      "epoch": 0.23908836809288325,
+      "grad_norm": 8.341093063354492,
+      "learning_rate": 9.986103911473075e-06,
+      "loss": 45.1884,
+      "step": 139
+    },
+    {
+      "epoch": 0.24080842829499033,
+      "grad_norm": 7.4146342277526855,
+      "learning_rate": 9.985382507271928e-06,
+      "loss": 44.8973,
+      "step": 140
+    },
+    {
+      "epoch": 0.2425284884970974,
+      "grad_norm": 8.28603458404541,
+      "learning_rate": 9.984642875744338e-06,
+      "loss": 44.6793,
+      "step": 141
+    },
+    {
+      "epoch": 0.24424854869920448,
+      "grad_norm": 7.7817583084106445,
+      "learning_rate": 9.983885019594506e-06,
+      "loss": 44.7361,
+      "step": 142
+    },
+    {
+      "epoch": 0.24596860890131156,
+      "grad_norm": 8.71711254119873,
+      "learning_rate": 9.983108941593277e-06,
+      "loss": 44.514,
+      "step": 143
+    },
+    {
+      "epoch": 0.2476886691034186,
+      "grad_norm": 8.746935844421387,
+      "learning_rate": 9.982314644578111e-06,
+      "loss": 45.4915,
+      "step": 144
+    },
+    {
+      "epoch": 0.24940872930552568,
+      "grad_norm": 8.616266250610352,
+      "learning_rate": 9.981502131453077e-06,
+      "loss": 44.4955,
+      "step": 145
+    },
+    {
+      "epoch": 0.2511287895076328,
+      "grad_norm": 8.73255443572998,
+      "learning_rate": 9.980671405188852e-06,
+      "loss": 44.7087,
+      "step": 146
+    },
+    {
+      "epoch": 0.25284884970973986,
+      "grad_norm": 7.063310623168945,
+      "learning_rate": 9.979822468822696e-06,
+      "loss": 44.7221,
+      "step": 147
+    },
+    {
+      "epoch": 0.25456890991184694,
+      "grad_norm": 8.635971069335938,
+      "learning_rate": 9.978955325458453e-06,
+      "loss": 44.5993,
+      "step": 148
+    },
+    {
+      "epoch": 0.25628897011395396,
+      "grad_norm": 7.771418571472168,
+      "learning_rate": 9.978069978266534e-06,
+      "loss": 45.1399,
+      "step": 149
+    },
+    {
+      "epoch": 0.25800903031606104,
+      "grad_norm": 9.317761421203613,
+      "learning_rate": 9.9771664304839e-06,
+      "loss": 44.8672,
+      "step": 150
+    },
+    {
+      "epoch": 0.2597290905181681,
+      "grad_norm": 8.102662086486816,
+      "learning_rate": 9.976244685414065e-06,
+      "loss": 45.4121,
+      "step": 151
+    },
+    {
+      "epoch": 0.2614491507202752,
+      "grad_norm": 12.453920364379883,
+      "learning_rate": 9.97530474642707e-06,
+      "loss": 44.6091,
+      "step": 152
+    },
+    {
+      "epoch": 0.26316921092238227,
+      "grad_norm": 9.156878471374512,
+      "learning_rate": 9.974346616959476e-06,
+      "loss": 45.1213,
+      "step": 153
+    },
+    {
+      "epoch": 0.26488927112448934,
+      "grad_norm": 11.89529037475586,
+      "learning_rate": 9.973370300514353e-06,
+      "loss": 44.5162,
+      "step": 154
+    },
+    {
+      "epoch": 0.2666093313265964,
+      "grad_norm": 9.946409225463867,
+      "learning_rate": 9.972375800661264e-06,
+      "loss": 44.106,
+      "step": 155
+    },
+    {
+      "epoch": 0.2683293915287035,
+      "grad_norm": 10.854037284851074,
+      "learning_rate": 9.971363121036252e-06,
+      "loss": 44.9188,
+      "step": 156
+    },
+    {
+      "epoch": 0.2700494517308106,
+      "grad_norm": 9.196579933166504,
+      "learning_rate": 9.970332265341833e-06,
+      "loss": 44.6235,
+      "step": 157
+    },
+    {
+      "epoch": 0.27176951193291765,
+      "grad_norm": 11.79194164276123,
+      "learning_rate": 9.969283237346973e-06,
+      "loss": 45.7572,
+      "step": 158
+    },
+    {
+      "epoch": 0.2734895721350247,
+      "grad_norm": 7.789967060089111,
+      "learning_rate": 9.968216040887078e-06,
+      "loss": 44.8069,
+      "step": 159
+    },
+    {
+      "epoch": 0.2752096323371318,
+      "grad_norm": 10.540818214416504,
+      "learning_rate": 9.967130679863984e-06,
+      "loss": 44.7341,
+      "step": 160
+    },
+    {
+      "epoch": 0.2769296925392389,
+      "grad_norm": 7.4835405349731445,
+      "learning_rate": 9.966027158245939e-06,
+      "loss": 45.4445,
+      "step": 161
+    },
+    {
+      "epoch": 0.27864975274134596,
+      "grad_norm": 9.472413063049316,
+      "learning_rate": 9.964905480067585e-06,
+      "loss": 44.5498,
+      "step": 162
+    },
+    {
+      "epoch": 0.28036981294345303,
+      "grad_norm": 9.155404090881348,
+      "learning_rate": 9.963765649429954e-06,
+      "loss": 44.3932,
+      "step": 163
+    },
+    {
+      "epoch": 0.2820898731455601,
+      "grad_norm": 8.14942455291748,
+      "learning_rate": 9.962607670500442e-06,
+      "loss": 44.5368,
+      "step": 164
+    },
+    {
+      "epoch": 0.2838099333476672,
+      "grad_norm": 10.103939056396484,
+      "learning_rate": 9.961431547512794e-06,
+      "loss": 44.1877,
+      "step": 165
+    },
+    {
+      "epoch": 0.28552999354977426,
+      "grad_norm": 7.972369194030762,
+      "learning_rate": 9.960237284767103e-06,
+      "loss": 44.775,
+      "step": 166
+    },
+    {
+      "epoch": 0.28725005375188134,
+      "grad_norm": 13.627598762512207,
+      "learning_rate": 9.959024886629772e-06,
+      "loss": 44.3529,
+      "step": 167
+    },
+    {
+      "epoch": 0.2889701139539884,
+      "grad_norm": 9.467161178588867,
+      "learning_rate": 9.957794357533518e-06,
+      "loss": 44.7574,
+      "step": 168
+    },
+    {
+      "epoch": 0.29069017415609544,
+      "grad_norm": 12.242958068847656,
+      "learning_rate": 9.956545701977347e-06,
+      "loss": 44.5439,
+      "step": 169
+    },
+    {
+      "epoch": 0.2924102343582025,
+      "grad_norm": 10.142478942871094,
+      "learning_rate": 9.955278924526532e-06,
+      "loss": 44.4951,
+      "step": 170
+    },
+    {
+      "epoch": 0.2941302945603096,
+      "grad_norm": 12.610379219055176,
+      "learning_rate": 9.95399402981261e-06,
+      "loss": 44.7777,
+      "step": 171
+    },
+    {
+      "epoch": 0.29585035476241667,
+      "grad_norm": 9.542120933532715,
+      "learning_rate": 9.952691022533352e-06,
+      "loss": 44.5877,
+      "step": 172
+    },
+    {
+      "epoch": 0.29757041496452374,
+      "grad_norm": 13.538004875183105,
+      "learning_rate": 9.951369907452752e-06,
+      "loss": 44.4161,
+      "step": 173
+    },
+    {
+      "epoch": 0.2992904751666308,
+      "grad_norm": 11.80547046661377,
+      "learning_rate": 9.950030689401014e-06,
+      "loss": 44.8214,
+      "step": 174
+    },
+    {
+      "epoch": 0.3010105353687379,
+      "grad_norm": 11.983033180236816,
+      "learning_rate": 9.948673373274523e-06,
+      "loss": 44.7073,
+      "step": 175
+    },
+    {
+      "epoch": 0.302730595570845,
+      "grad_norm": 10.844096183776855,
+      "learning_rate": 9.947297964035837e-06,
+      "loss": 44.4097,
+      "step": 176
+    },
+    {
+      "epoch": 0.30445065577295205,
+      "grad_norm": 11.820161819458008,
+      "learning_rate": 9.94590446671366e-06,
+      "loss": 44.7555,
+      "step": 177
+    },
+    {
+      "epoch": 0.3061707159750591,
+      "grad_norm": 11.587363243103027,
+      "learning_rate": 9.94449288640284e-06,
+      "loss": 44.2532,
+      "step": 178
+    },
+    {
+      "epoch": 0.3078907761771662,
+      "grad_norm": 10.775352478027344,
+      "learning_rate": 9.943063228264327e-06,
+      "loss": 44.8222,
+      "step": 179
+    },
+    {
+      "epoch": 0.3096108363792733,
+      "grad_norm": 12.077004432678223,
+      "learning_rate": 9.941615497525172e-06,
+      "loss": 45.0694,
+      "step": 180
+    },
+    {
+      "epoch": 0.31133089658138036,
+      "grad_norm": 9.174921035766602,
+      "learning_rate": 9.940149699478502e-06,
+      "loss": 44.4424,
+      "step": 181
+    },
+    {
+      "epoch": 0.31305095678348743,
+      "grad_norm": 9.949912071228027,
+      "learning_rate": 9.938665839483503e-06,
+      "loss": 44.82,
+      "step": 182
+    },
+    {
+      "epoch": 0.3147710169855945,
+      "grad_norm": 10.341442108154297,
+      "learning_rate": 9.937163922965394e-06,
+      "loss": 43.993,
+      "step": 183
+    },
+    {
+      "epoch": 0.3164910771877016,
+      "grad_norm": 7.373370170593262,
+      "learning_rate": 9.93564395541541e-06,
+      "loss": 44.6727,
+      "step": 184
+    },
+    {
+      "epoch": 0.31821113738980866,
+      "grad_norm": 11.212616920471191,
+      "learning_rate": 9.93410594239079e-06,
+      "loss": 44.3283,
+      "step": 185
+    },
+    {
+      "epoch": 0.31993119759191574,
+      "grad_norm": 8.05331802368164,
+      "learning_rate": 9.932549889514747e-06,
+      "loss": 44.6646,
+      "step": 186
+    },
+    {
+      "epoch": 0.3216512577940228,
+      "grad_norm": 11.395249366760254,
+      "learning_rate": 9.930975802476448e-06,
+      "loss": 43.9711,
+      "step": 187
+    },
+    {
+      "epoch": 0.3233713179961299,
+      "grad_norm": 7.730330944061279,
+      "learning_rate": 9.929383687030999e-06,
+      "loss": 45.2283,
+      "step": 188
+    },
+    {
+      "epoch": 0.3250913781982369,
+      "grad_norm": 11.927477836608887,
+      "learning_rate": 9.927773548999419e-06,
+      "loss": 44.723,
+      "step": 189
+    },
+    {
+      "epoch": 0.326811438400344,
+      "grad_norm": 8.933055877685547,
+      "learning_rate": 9.92614539426862e-06,
+      "loss": 44.5058,
+      "step": 190
+    },
+    {
+      "epoch": 0.32853149860245107,
+      "grad_norm": 13.985485076904297,
+      "learning_rate": 9.924499228791387e-06,
+      "loss": 44.8917,
+      "step": 191
+    },
+    {
+      "epoch": 0.33025155880455814,
+      "grad_norm": 13.561887741088867,
+      "learning_rate": 9.922835058586353e-06,
+      "loss": 44.6659,
+      "step": 192
+    },
+    {
+      "epoch": 0.3319716190066652,
+      "grad_norm": 8.802722930908203,
+      "learning_rate": 9.921152889737985e-06,
+      "loss": 43.9996,
+      "step": 193
+    },
+    {
+      "epoch": 0.3336916792087723,
+      "grad_norm": 10.387024879455566,
+      "learning_rate": 9.919452728396548e-06,
+      "loss": 44.7691,
+      "step": 194
+    },
+    {
+      "epoch": 0.3354117394108794,
+      "grad_norm": 8.743358612060547,
+      "learning_rate": 9.917734580778094e-06,
+      "loss": 45.2034,
+      "step": 195
+    },
+    {
+      "epoch": 0.33713179961298645,
+      "grad_norm": 7.972978591918945,
+      "learning_rate": 9.915998453164435e-06,
+      "loss": 45.5303,
+      "step": 196
+    },
+    {
+      "epoch": 0.3388518598150935,
+      "grad_norm": 8.630827903747559,
+      "learning_rate": 9.914244351903122e-06,
+      "loss": 44.7464,
+      "step": 197
+    },
+    {
+      "epoch": 0.3405719200172006,
+      "grad_norm": 6.4518609046936035,
+      "learning_rate": 9.912472283407421e-06,
+      "loss": 44.2976,
+      "step": 198
+    },
+    {
+      "epoch": 0.3422919802193077,
+      "grad_norm": 8.886955261230469,
+      "learning_rate": 9.910682254156284e-06,
+      "loss": 44.8556,
+      "step": 199
+    },
+    {
+      "epoch": 0.34401204042141476,
+      "grad_norm": 7.261415004730225,
+      "learning_rate": 9.908874270694337e-06,
+      "loss": 44.466,
+      "step": 200
+    },
+    {
+      "epoch": 0.34573210062352183,
+      "grad_norm": 8.031500816345215,
+      "learning_rate": 9.907048339631843e-06,
+      "loss": 44.4753,
+      "step": 201
+    },
+    {
+      "epoch": 0.3474521608256289,
+      "grad_norm": 7.2853569984436035,
+      "learning_rate": 9.905204467644688e-06,
+      "loss": 45.0669,
+      "step": 202
+    },
+    {
+      "epoch": 0.349172221027736,
+      "grad_norm": 8.290436744689941,
+      "learning_rate": 9.903342661474355e-06,
+      "loss": 44.7901,
+      "step": 203
+    },
+    {
+      "epoch": 0.35089228122984306,
+      "grad_norm": 9.512991905212402,
+      "learning_rate": 9.901462927927891e-06,
+      "loss": 44.7863,
+      "step": 204
+    },
+    {
+      "epoch": 0.35261234143195014,
+      "grad_norm": 8.04033374786377,
+      "learning_rate": 9.899565273877892e-06,
+      "loss": 44.9776,
+      "step": 205
+    },
+    {
+      "epoch": 0.3543324016340572,
+      "grad_norm": 11.140141487121582,
+      "learning_rate": 9.897649706262474e-06,
+      "loss": 45.1048,
+      "step": 206
+    },
+    {
+      "epoch": 0.3560524618361643,
+      "grad_norm": 8.036794662475586,
+      "learning_rate": 9.895716232085247e-06,
+      "loss": 44.763,
+      "step": 207
+    },
+    {
+      "epoch": 0.3577725220382713,
+      "grad_norm": 11.508170127868652,
+      "learning_rate": 9.89376485841529e-06,
+      "loss": 44.7094,
+      "step": 208
+    },
+    {
+      "epoch": 0.3594925822403784,
+      "grad_norm": 8.577386856079102,
+      "learning_rate": 9.891795592387127e-06,
+      "loss": 44.5892,
+      "step": 209
+    },
+    {
+      "epoch": 0.36121264244248547,
+      "grad_norm": 11.701868057250977,
+      "learning_rate": 9.889808441200697e-06,
+      "loss": 44.4664,
+      "step": 210
+    },
+    {
+      "epoch": 0.36293270264459254,
+      "grad_norm": 7.955048561096191,
+      "learning_rate": 9.887803412121331e-06,
+      "loss": 44.5424,
+      "step": 211
+    },
+    {
+      "epoch": 0.3646527628466996,
+      "grad_norm": 11.340240478515625,
+      "learning_rate": 9.885780512479725e-06,
+      "loss": 44.3322,
+      "step": 212
+    },
+    {
+      "epoch": 0.3663728230488067,
+      "grad_norm": 8.020219802856445,
+      "learning_rate": 9.88373974967191e-06,
+      "loss": 44.2046,
+      "step": 213
+    },
+    {
+      "epoch": 0.36809288325091377,
+      "grad_norm": 10.230839729309082,
+      "learning_rate": 9.881681131159232e-06,
+      "loss": 44.2038,
+      "step": 214
+    },
+    {
+      "epoch": 0.36981294345302085,
+      "grad_norm": 8.711820602416992,
+      "learning_rate": 9.879604664468315e-06,
+      "loss": 44.0065,
+      "step": 215
+    },
+    {
+      "epoch": 0.3715330036551279,
+      "grad_norm": 10.792113304138184,
+      "learning_rate": 9.877510357191042e-06,
+      "loss": 44.3272,
+      "step": 216
+    },
+    {
+      "epoch": 0.373253063857235,
+      "grad_norm": 8.520330429077148,
+      "learning_rate": 9.875398216984521e-06,
+      "loss": 44.403,
+      "step": 217
+    },
+    {
+      "epoch": 0.3749731240593421,
+      "grad_norm": 12.095443725585938,
+      "learning_rate": 9.873268251571065e-06,
+      "loss": 44.8801,
+      "step": 218
+    },
+    {
+      "epoch": 0.37669318426144915,
+      "grad_norm": 10.553284645080566,
+      "learning_rate": 9.871120468738156e-06,
+      "loss": 44.1855,
+      "step": 219
+    },
+    {
+      "epoch": 0.37841324446355623,
+      "grad_norm": 10.677011489868164,
+      "learning_rate": 9.868954876338414e-06,
+      "loss": 44.9765,
+      "step": 220
+    },
+    {
+      "epoch": 0.3801333046656633,
+      "grad_norm": 10.681044578552246,
+      "learning_rate": 9.866771482289585e-06,
+      "loss": 44.3767,
+      "step": 221
+    },
+    {
+      "epoch": 0.3818533648677704,
+      "grad_norm": 8.905034065246582,
+      "learning_rate": 9.86457029457449e-06,
+      "loss": 44.5735,
+      "step": 222
+    },
+    {
+      "epoch": 0.38357342506987746,
+      "grad_norm": 11.448368072509766,
+      "learning_rate": 9.86235132124101e-06,
+      "loss": 44.3242,
+      "step": 223
+    },
+    {
+      "epoch": 0.38529348527198454,
+      "grad_norm": 8.515959739685059,
+      "learning_rate": 9.860114570402055e-06,
+      "loss": 45.0004,
+      "step": 224
+    },
+    {
+      "epoch": 0.3870135454740916,
+      "grad_norm": 11.419853210449219,
+      "learning_rate": 9.85786005023553e-06,
+      "loss": 44.7329,
+      "step": 225
+    },
+    {
+      "epoch": 0.3887336056761987,
+      "grad_norm": 10.274301528930664,
+      "learning_rate": 9.855587768984308e-06,
+      "loss": 44.5693,
+      "step": 226
+    },
+    {
+      "epoch": 0.39045366587830577,
+      "grad_norm": 9.408884048461914,
+      "learning_rate": 9.8532977349562e-06,
+      "loss": 44.2397,
+      "step": 227
+    },
+    {
+      "epoch": 0.3921737260804128,
+      "grad_norm": 8.234622955322266,
+      "learning_rate": 9.850989956523922e-06,
+      "loss": 44.3023,
+      "step": 228
+    },
+    {
+      "epoch": 0.39389378628251986,
+      "grad_norm": 10.358318328857422,
+      "learning_rate": 9.848664442125068e-06,
+      "loss": 44.6614,
+      "step": 229
+    },
+    {
+      "epoch": 0.39561384648462694,
+      "grad_norm": 9.553954124450684,
+      "learning_rate": 9.846321200262079e-06,
+      "loss": 44.8496,
+      "step": 230
+    },
+    {
+      "epoch": 0.397333906686734,
+      "grad_norm": 9.689641952514648,
+      "learning_rate": 9.843960239502205e-06,
+      "loss": 44.1763,
+      "step": 231
+    },
+    {
+      "epoch": 0.3990539668888411,
+      "grad_norm": 8.881220817565918,
+      "learning_rate": 9.841581568477483e-06,
+      "loss": 44.5679,
+      "step": 232
+    },
+    {
+      "epoch": 0.40077402709094817,
+      "grad_norm": 8.874262809753418,
+      "learning_rate": 9.839185195884702e-06,
+      "loss": 45.0497,
+      "step": 233
+    },
+    {
+      "epoch": 0.40249408729305525,
+      "grad_norm": 8.762188911437988,
+      "learning_rate": 9.836771130485367e-06,
+      "loss": 44.415,
+      "step": 234
+    },
+    {
+      "epoch": 0.4042141474951623,
+      "grad_norm": 7.233129501342773,
+      "learning_rate": 9.834339381105676e-06,
+      "loss": 44.4637,
+      "step": 235
+    },
+    {
+      "epoch": 0.4059342076972694,
+      "grad_norm": 10.540146827697754,
+      "learning_rate": 9.831889956636478e-06,
+      "loss": 44.525,
+      "step": 236
+    },
+    {
+      "epoch": 0.4076542678993765,
+      "grad_norm": 5.345295429229736,
+      "learning_rate": 9.829422866033246e-06,
+      "loss": 43.6553,
+      "step": 237
+    },
+    {
+      "epoch": 0.40937432810148355,
+      "grad_norm": 9.131731033325195,
+      "learning_rate": 9.826938118316044e-06,
+      "loss": 44.2395,
+      "step": 238
+    },
+    {
+      "epoch": 0.41109438830359063,
+      "grad_norm": 6.4219560623168945,
+      "learning_rate": 9.82443572256949e-06,
+      "loss": 44.541,
+      "step": 239
+    },
+    {
+      "epoch": 0.4128144485056977,
+      "grad_norm": 7.843993663787842,
+      "learning_rate": 9.821915687942729e-06,
+      "loss": 44.6975,
+      "step": 240
+    },
+    {
+      "epoch": 0.4145345087078048,
+      "grad_norm": 7.926616668701172,
+      "learning_rate": 9.8193780236494e-06,
+      "loss": 43.983,
+      "step": 241
+    },
+    {
+      "epoch": 0.41625456890991186,
+      "grad_norm": 7.5367045402526855,
+      "learning_rate": 9.81682273896759e-06,
+      "loss": 43.9646,
+      "step": 242
+    },
+    {
+      "epoch": 0.41797462911201894,
+      "grad_norm": 10.298775672912598,
+      "learning_rate": 9.814249843239816e-06,
+      "loss": 44.0679,
+      "step": 243
+    },
+    {
+      "epoch": 0.419694689314126,
+      "grad_norm": 8.142918586730957,
+      "learning_rate": 9.811659345872979e-06,
+      "loss": 44.9597,
+      "step": 244
+    },
+    {
+      "epoch": 0.4214147495162331,
+      "grad_norm": 11.297587394714355,
+      "learning_rate": 9.809051256338338e-06,
+      "loss": 44.3569,
+      "step": 245
+    },
+    {
+      "epoch": 0.42313480971834017,
+      "grad_norm": 8.743409156799316,
+      "learning_rate": 9.806425584171468e-06,
+      "loss": 43.6217,
+      "step": 246
+    },
+    {
+      "epoch": 0.42485486992044724,
+      "grad_norm": 9.554738998413086,
+      "learning_rate": 9.803782338972235e-06,
+      "loss": 44.3762,
+      "step": 247
+    },
+    {
+      "epoch": 0.42657493012255426,
+      "grad_norm": 8.766114234924316,
+      "learning_rate": 9.801121530404746e-06,
+      "loss": 44.1824,
+      "step": 248
+    },
+    {
+      "epoch": 0.42829499032466134,
+      "grad_norm": 8.465466499328613,
+      "learning_rate": 9.798443168197332e-06,
+      "loss": 44.0283,
+      "step": 249
+    },
+    {
+      "epoch": 0.4300150505267684,
+      "grad_norm": 8.999267578125,
+      "learning_rate": 9.795747262142494e-06,
+      "loss": 44.1171,
+      "step": 250
+    },
+    {
+      "epoch": 0.4317351107288755,
+      "grad_norm": 6.678277492523193,
+      "learning_rate": 9.79303382209688e-06,
+      "loss": 44.7172,
+      "step": 251
+    },
+    {
+      "epoch": 0.43345517093098257,
+      "grad_norm": 9.430837631225586,
+      "learning_rate": 9.790302857981247e-06,
+      "loss": 44.3632,
+      "step": 252
+    },
+    {
+      "epoch": 0.43517523113308965,
+      "grad_norm": 6.532567501068115,
+      "learning_rate": 9.787554379780417e-06,
+      "loss": 44.2348,
+      "step": 253
+    },
+    {
+      "epoch": 0.4368952913351967,
+      "grad_norm": 9.008966445922852,
+      "learning_rate": 9.784788397543254e-06,
+      "loss": 43.9189,
+      "step": 254
+    },
+    {
+      "epoch": 0.4386153515373038,
+      "grad_norm": 7.171030521392822,
+      "learning_rate": 9.782004921382612e-06,
+      "loss": 44.719,
+      "step": 255
+    },
+    {
+      "epoch": 0.4403354117394109,
+      "grad_norm": 8.457947731018066,
+      "learning_rate": 9.77920396147531e-06,
+      "loss": 44.3203,
+      "step": 256
+    },
+    {
+      "epoch": 0.44205547194151795,
+      "grad_norm": 8.303704261779785,
+      "learning_rate": 9.77638552806209e-06,
+      "loss": 44.6251,
+      "step": 257
+    },
+    {
+      "epoch": 0.44377553214362503,
+      "grad_norm": 7.793144702911377,
+      "learning_rate": 9.773549631447576e-06,
+      "loss": 44.4527,
+      "step": 258
+    },
+    {
+      "epoch": 0.4454955923457321,
+      "grad_norm": 9.074666976928711,
+      "learning_rate": 9.770696282000245e-06,
+      "loss": 44.4602,
+      "step": 259
+    },
+    {
+      "epoch": 0.4472156525478392,
+      "grad_norm": 7.790366172790527,
+      "learning_rate": 9.767825490152381e-06,
+      "loss": 44.0525,
+      "step": 260
+    },
+    {
+      "epoch": 0.44893571274994626,
+      "grad_norm": 6.790820598602295,
+      "learning_rate": 9.764937266400042e-06,
+      "loss": 44.3677,
+      "step": 261
+    },
+    {
+      "epoch": 0.45065577295205334,
+      "grad_norm": 7.48856782913208,
+      "learning_rate": 9.76203162130302e-06,
+      "loss": 44.281,
+      "step": 262
+    },
+    {
+      "epoch": 0.4523758331541604,
+      "grad_norm": 6.4276814460754395,
+      "learning_rate": 9.759108565484796e-06,
+      "loss": 44.5151,
+      "step": 263
+    },
+    {
+      "epoch": 0.4540958933562675,
+      "grad_norm": 7.4707350730896,
+      "learning_rate": 9.756168109632519e-06,
+      "loss": 44.4653,
+      "step": 264
+    },
+    {
+      "epoch": 0.45581595355837456,
+      "grad_norm": 6.78653621673584,
+      "learning_rate": 9.753210264496943e-06,
+      "loss": 45.1088,
+      "step": 265
+    },
+    {
+      "epoch": 0.45753601376048164,
+      "grad_norm": 6.732356548309326,
+      "learning_rate": 9.75023504089241e-06,
+      "loss": 44.3737,
+      "step": 266
+    },
+    {
+      "epoch": 0.4592560739625887,
+      "grad_norm": 6.069101810455322,
+      "learning_rate": 9.747242449696794e-06,
+      "loss": 44.5619,
+      "step": 267
+    },
+    {
+      "epoch": 0.46097613416469574,
+      "grad_norm": 6.123370170593262,
+      "learning_rate": 9.74423250185147e-06,
+      "loss": 44.3021,
+      "step": 268
+    },
+    {
+      "epoch": 0.4626961943668028,
+      "grad_norm": 6.104144096374512,
+      "learning_rate": 9.74120520836127e-06,
+      "loss": 44.2148,
+      "step": 269
+    },
+    {
+      "epoch": 0.4644162545689099,
+      "grad_norm": 8.752744674682617,
+      "learning_rate": 9.738160580294444e-06,
+      "loss": 44.6205,
+      "step": 270
+    },
+    {
+      "epoch": 0.46613631477101697,
+      "grad_norm": 7.354405879974365,
+      "learning_rate": 9.735098628782624e-06,
+      "loss": 44.459,
+      "step": 271
+    },
+    {
+      "epoch": 0.46785637497312405,
+      "grad_norm": 8.801732063293457,
+      "learning_rate": 9.732019365020778e-06,
+      "loss": 44.635,
+      "step": 272
+    },
+    {
+      "epoch": 0.4695764351752311,
+      "grad_norm": 7.110331058502197,
+      "learning_rate": 9.728922800267162e-06,
+      "loss": 44.3402,
+      "step": 273
+    },
+    {
+      "epoch": 0.4712964953773382,
+      "grad_norm": 8.147953033447266,
+      "learning_rate": 9.7258089458433e-06,
+      "loss": 44.6291,
+      "step": 274
+    },
+    {
+      "epoch": 0.4730165555794453,
+      "grad_norm": 7.08845853805542,
+      "learning_rate": 9.722677813133921e-06,
+      "loss": 45.0577,
+      "step": 275
+    },
+    {
+      "epoch": 0.47473661578155235,
+      "grad_norm": 6.4095025062561035,
+      "learning_rate": 9.719529413586928e-06,
+      "loss": 43.9258,
+      "step": 276
+    },
+    {
+      "epoch": 0.47645667598365943,
+      "grad_norm": 7.270499229431152,
+      "learning_rate": 9.716363758713357e-06,
+      "loss": 44.8198,
+      "step": 277
+    },
+    {
+      "epoch": 0.4781767361857665,
+      "grad_norm": 7.040411949157715,
+      "learning_rate": 9.713180860087328e-06,
+      "loss": 44.1966,
+      "step": 278
+    },
+    {
+      "epoch": 0.4798967963878736,
+      "grad_norm": 6.831189155578613,
+      "learning_rate": 9.709980729346009e-06,
+      "loss": 44.88,
+      "step": 279
+    },
+    {
+      "epoch": 0.48161685658998066,
+      "grad_norm": 7.017982482910156,
+      "learning_rate": 9.706763378189571e-06,
+      "loss": 44.3914,
+      "step": 280
+    },
+    {
+      "epoch": 0.48333691679208773,
+      "grad_norm": 7.631150245666504,
+      "learning_rate": 9.703528818381144e-06,
+      "loss": 44.3798,
+      "step": 281
+    },
+    {
+      "epoch": 0.4850569769941948,
+      "grad_norm": 6.717972278594971,
+      "learning_rate": 9.70027706174678e-06,
+      "loss": 43.8038,
+      "step": 282
+    },
+    {
+      "epoch": 0.4867770371963019,
+      "grad_norm": 8.252500534057617,
+      "learning_rate": 9.697008120175402e-06,
+      "loss": 44.4889,
+      "step": 283
+    },
+    {
+      "epoch": 0.48849709739840896,
+      "grad_norm": 7.6612420082092285,
+      "learning_rate": 9.693722005618763e-06,
+      "loss": 44.2001,
+      "step": 284
+    },
+    {
+      "epoch": 0.49021715760051604,
+      "grad_norm": 6.846263408660889,
+      "learning_rate": 9.690418730091403e-06,
+      "loss": 44.6985,
+      "step": 285
+    },
+    {
+      "epoch": 0.4919372178026231,
+      "grad_norm": 6.461937427520752,
+      "learning_rate": 9.687098305670606e-06,
+      "loss": 44.2506,
+      "step": 286
+    },
+    {
+      "epoch": 0.49365727800473014,
+      "grad_norm": 6.650141716003418,
+      "learning_rate": 9.683760744496356e-06,
+      "loss": 44.3858,
+      "step": 287
+    },
+    {
+      "epoch": 0.4953773382068372,
+      "grad_norm": 6.5828986167907715,
+      "learning_rate": 9.68040605877129e-06,
+      "loss": 43.6322,
+      "step": 288
+    },
+    {
+      "epoch": 0.4970973984089443,
+      "grad_norm": 6.385183334350586,
+      "learning_rate": 9.677034260760658e-06,
+      "loss": 44.7745,
+      "step": 289
+    },
+    {
+      "epoch": 0.49881745861105137,
+      "grad_norm": 7.130415916442871,
+      "learning_rate": 9.673645362792273e-06,
+      "loss": 44.1543,
+      "step": 290
+    },
+    {
+      "epoch": 0.5005375188131584,
+      "grad_norm": 6.580416202545166,
+      "learning_rate": 9.670239377256467e-06,
+      "loss": 43.8422,
+      "step": 291
+    },
+    {
+      "epoch": 0.5022575790152656,
+      "grad_norm": 7.959731101989746,
+      "learning_rate": 9.666816316606044e-06,
+      "loss": 44.4367,
+      "step": 292
+    },
+    {
+      "epoch": 0.5039776392173726,
+      "grad_norm": 6.089702606201172,
+      "learning_rate": 9.663376193356249e-06,
+      "loss": 43.9484,
+      "step": 293
+    },
+    {
+      "epoch": 0.5056976994194797,
+      "grad_norm": 8.458806037902832,
+      "learning_rate": 9.659919020084695e-06,
+      "loss": 44.1408,
+      "step": 294
+    },
+    {
+      "epoch": 0.5074177596215868,
+      "grad_norm": 6.733780860900879,
+      "learning_rate": 9.656444809431344e-06,
+      "loss": 43.9267,
+      "step": 295
+    },
+    {
+      "epoch": 0.5091378198236939,
+      "grad_norm": 8.281241416931152,
+      "learning_rate": 9.652953574098444e-06,
+      "loss": 44.7447,
+      "step": 296
+    },
+    {
+      "epoch": 0.5108578800258009,
+      "grad_norm": 7.2893195152282715,
+      "learning_rate": 9.649445326850491e-06,
+      "loss": 44.1749,
+      "step": 297
+    },
+    {
+      "epoch": 0.5125779402279079,
+      "grad_norm": 8.188138961791992,
+      "learning_rate": 9.645920080514176e-06,
+      "loss": 44.5725,
+      "step": 298
+    },
+    {
+      "epoch": 0.5142980004300151,
+      "grad_norm": 8.285508155822754,
+      "learning_rate": 9.642377847978343e-06,
+      "loss": 44.4519,
+      "step": 299
+    },
+    {
+      "epoch": 0.5160180606321221,
+      "grad_norm": 12.107803344726562,
+      "learning_rate": 9.638818642193939e-06,
+      "loss": 43.6642,
+      "step": 300
+    },
+    {
+      "epoch": 0.5177381208342292,
+      "grad_norm": 10.988150596618652,
+      "learning_rate": 9.63524247617397e-06,
+      "loss": 43.9385,
+      "step": 301
+    },
+    {
+      "epoch": 0.5194581810363362,
+      "grad_norm": 12.65985107421875,
+      "learning_rate": 9.631649362993447e-06,
+      "loss": 44.304,
+      "step": 302
+    },
+    {
+      "epoch": 0.5211782412384434,
+      "grad_norm": 12.63979721069336,
+      "learning_rate": 9.62803931578935e-06,
+      "loss": 44.2028,
+      "step": 303
+    },
+    {
+      "epoch": 0.5228983014405504,
+      "grad_norm": 7.90657377243042,
+      "learning_rate": 9.624412347760564e-06,
+      "loss": 44.1649,
+      "step": 304
+    },
+    {
+      "epoch": 0.5246183616426575,
+      "grad_norm": 9.31624698638916,
+      "learning_rate": 9.620768472167844e-06,
+      "loss": 43.996,
+      "step": 305
+    },
+    {
+      "epoch": 0.5263384218447645,
+      "grad_norm": 8.557055473327637,
+      "learning_rate": 9.61710770233376e-06,
+      "loss": 44.3358,
+      "step": 306
+    },
+    {
+      "epoch": 0.5280584820468717,
+      "grad_norm": 7.057743549346924,
+      "learning_rate": 9.613430051642652e-06,
+      "loss": 44.583,
+      "step": 307
+    },
+    {
+      "epoch": 0.5297785422489787,
+      "grad_norm": 7.244456768035889,
+      "learning_rate": 9.609735533540576e-06,
+      "loss": 43.7423,
+      "step": 308
+    },
+    {
+      "epoch": 0.5314986024510858,
+      "grad_norm": 6.6239333152771,
+      "learning_rate": 9.606024161535261e-06,
+      "loss": 43.719,
+      "step": 309
+    },
+    {
+      "epoch": 0.5332186626531928,
+      "grad_norm": 7.109512805938721,
+      "learning_rate": 9.602295949196052e-06,
+      "loss": 43.8263,
+      "step": 310
+    },
+    {
+      "epoch": 0.5349387228553,
+      "grad_norm": 7.938423156738281,
+      "learning_rate": 9.59855091015387e-06,
+      "loss": 43.513,
+      "step": 311
+    },
+    {
+      "epoch": 0.536658783057407,
+      "grad_norm": 7.159519195556641,
+      "learning_rate": 9.594789058101154e-06,
+      "loss": 43.7476,
+      "step": 312
+    },
+    {
+      "epoch": 0.5383788432595141,
+      "grad_norm": 9.051861763000488,
+      "learning_rate": 9.591010406791814e-06,
+      "loss": 44.6808,
+      "step": 313
+    },
+    {
+      "epoch": 0.5400989034616211,
+      "grad_norm": 7.437441825866699,
+      "learning_rate": 9.587214970041181e-06,
+      "loss": 44.2018,
+      "step": 314
+    },
+    {
+      "epoch": 0.5418189636637283,
+      "grad_norm": 8.533609390258789,
+      "learning_rate": 9.58340276172596e-06,
+      "loss": 44.35,
+      "step": 315
+    },
+    {
+      "epoch": 0.5435390238658353,
+      "grad_norm": 7.41975736618042,
+      "learning_rate": 9.579573795784167e-06,
+      "loss": 44.1627,
+      "step": 316
+    },
+    {
+      "epoch": 0.5452590840679423,
+      "grad_norm": 6.5013580322265625,
+      "learning_rate": 9.575728086215093e-06,
+      "loss": 44.0411,
+      "step": 317
+    },
+    {
+      "epoch": 0.5469791442700495,
+      "grad_norm": 8.311059951782227,
+      "learning_rate": 9.571865647079246e-06,
+      "loss": 44.6953,
+      "step": 318
+    },
+    {
+      "epoch": 0.5486992044721565,
+      "grad_norm": 5.960739612579346,
+      "learning_rate": 9.567986492498299e-06,
+      "loss": 44.1261,
+      "step": 319
+    },
+    {
+      "epoch": 0.5504192646742636,
+      "grad_norm": 7.7150959968566895,
+      "learning_rate": 9.564090636655033e-06,
+      "loss": 44.0052,
+      "step": 320
+    },
+    {
+      "epoch": 0.5521393248763706,
+      "grad_norm": 7.0516815185546875,
+      "learning_rate": 9.560178093793304e-06,
+      "loss": 44.4024,
+      "step": 321
+    },
+    {
+      "epoch": 0.5538593850784778,
+      "grad_norm": 6.510403633117676,
+      "learning_rate": 9.55624887821797e-06,
+      "loss": 44.3171,
+      "step": 322
+    },
+    {
+      "epoch": 0.5555794452805848,
+      "grad_norm": 6.586174488067627,
+      "learning_rate": 9.552303004294845e-06,
+      "loss": 44.1694,
+      "step": 323
+    },
+    {
+      "epoch": 0.5572995054826919,
+      "grad_norm": 7.093349456787109,
+      "learning_rate": 9.548340486450656e-06,
+      "loss": 43.9714,
+      "step": 324
+    },
+    {
+      "epoch": 0.5590195656847989,
+      "grad_norm": 5.638337135314941,
+      "learning_rate": 9.544361339172976e-06,
+      "loss": 43.9597,
+      "step": 325
+    },
+    {
+      "epoch": 0.5607396258869061,
+      "grad_norm": 6.339056491851807,
+      "learning_rate": 9.54036557701018e-06,
+      "loss": 43.9718,
+      "step": 326
+    },
+    {
+      "epoch": 0.5624596860890131,
+      "grad_norm": 6.259551525115967,
+      "learning_rate": 9.536353214571393e-06,
+      "loss": 44.1765,
+      "step": 327
+    },
+    {
+      "epoch": 0.5641797462911202,
+      "grad_norm": 6.071033477783203,
+      "learning_rate": 9.53232426652643e-06,
+      "loss": 44.1298,
+      "step": 328
+    },
+    {
+      "epoch": 0.5658998064932272,
+      "grad_norm": 5.966522216796875,
+      "learning_rate": 9.528278747605741e-06,
+      "loss": 43.5899,
+      "step": 329
+    },
+    {
+      "epoch": 0.5676198666953344,
+      "grad_norm": 6.373861312866211,
+      "learning_rate": 9.52421667260037e-06,
+      "loss": 43.9747,
+      "step": 330
+    },
+    {
+      "epoch": 0.5693399268974414,
+      "grad_norm": 7.79583215713501,
+      "learning_rate": 9.52013805636189e-06,
+      "loss": 44.1957,
+      "step": 331
+    },
+    {
+      "epoch": 0.5710599870995485,
+      "grad_norm": 6.730131149291992,
+      "learning_rate": 9.516042913802349e-06,
+      "loss": 44.5105,
+      "step": 332
+    },
+    {
+      "epoch": 0.5727800473016555,
+      "grad_norm": 6.649820804595947,
+      "learning_rate": 9.511931259894219e-06,
+      "loss": 43.4763,
+      "step": 333
+    },
+    {
+      "epoch": 0.5745001075037627,
+      "grad_norm": 8.12640380859375,
+      "learning_rate": 9.507803109670337e-06,
+      "loss": 43.5421,
+      "step": 334
+    },
+    {
+      "epoch": 0.5762201677058697,
+      "grad_norm": 7.1897783279418945,
+      "learning_rate": 9.503658478223862e-06,
+      "loss": 43.661,
+      "step": 335
+    },
+    {
+      "epoch": 0.5779402279079768,
+      "grad_norm": 6.837246417999268,
+      "learning_rate": 9.499497380708202e-06,
+      "loss": 43.7997,
+      "step": 336
+    },
+    {
+      "epoch": 0.5796602881100839,
+      "grad_norm": 8.631741523742676,
+      "learning_rate": 9.495319832336969e-06,
+      "loss": 43.8287,
+      "step": 337
+    },
+    {
+      "epoch": 0.5813803483121909,
+      "grad_norm": 8.00960636138916,
+      "learning_rate": 9.491125848383926e-06,
+      "loss": 43.6861,
+      "step": 338
+    },
+    {
+      "epoch": 0.583100408514298,
+      "grad_norm": 6.487185001373291,
+      "learning_rate": 9.486915444182926e-06,
+      "loss": 43.8275,
+      "step": 339
+    },
+    {
+      "epoch": 0.584820468716405,
+      "grad_norm": 7.411306381225586,
+      "learning_rate": 9.482688635127849e-06,
+      "loss": 43.4639,
+      "step": 340
+    },
+    {
+      "epoch": 0.5865405289185122,
+      "grad_norm": 6.86546516418457,
+      "learning_rate": 9.478445436672566e-06,
+      "loss": 44.094,
+      "step": 341
+    },
+    {
+      "epoch": 0.5882605891206192,
+      "grad_norm": 8.094916343688965,
+      "learning_rate": 9.474185864330861e-06,
+      "loss": 43.8569,
+      "step": 342
+    },
+    {
+      "epoch": 0.5899806493227263,
+      "grad_norm": 7.7384138107299805,
+      "learning_rate": 9.469909933676388e-06,
+      "loss": 43.8937,
+      "step": 343
+    },
+    {
+      "epoch": 0.5917007095248333,
+      "grad_norm": 8.301685333251953,
+      "learning_rate": 9.46561766034261e-06,
+      "loss": 44.288,
+      "step": 344
+    },
+    {
+      "epoch": 0.5934207697269405,
+      "grad_norm": 10.100611686706543,
+      "learning_rate": 9.461309060022737e-06,
+      "loss": 43.9889,
+      "step": 345
+    },
+    {
+      "epoch": 0.5951408299290475,
+      "grad_norm": 6.778660774230957,
+      "learning_rate": 9.456984148469674e-06,
+      "loss": 43.697,
+      "step": 346
+    },
+    {
+      "epoch": 0.5968608901311546,
+      "grad_norm": 11.071110725402832,
+      "learning_rate": 9.452642941495967e-06,
+      "loss": 44.3392,
+      "step": 347
+    },
+    {
+      "epoch": 0.5985809503332616,
+      "grad_norm": 7.321798324584961,
+      "learning_rate": 9.448285454973739e-06,
+      "loss": 43.5908,
+      "step": 348
+    },
+    {
+      "epoch": 0.6003010105353688,
+      "grad_norm": 10.584439277648926,
+      "learning_rate": 9.443911704834624e-06,
+      "loss": 44.094,
+      "step": 349
+    },
+    {
+      "epoch": 0.6020210707374758,
+      "grad_norm": 7.291213035583496,
+      "learning_rate": 9.439521707069737e-06,
+      "loss": 43.7771,
+      "step": 350
+    },
+    {
+      "epoch": 0.6037411309395829,
+      "grad_norm": 10.645991325378418,
+      "learning_rate": 9.435115477729577e-06,
+      "loss": 44.2799,
+      "step": 351
+    },
+    {
+      "epoch": 0.60546119114169,
+      "grad_norm": 8.463363647460938,
+      "learning_rate": 9.430693032924003e-06,
+      "loss": 43.9206,
+      "step": 352
+    },
+    {
+      "epoch": 0.6071812513437971,
+      "grad_norm": 11.510597229003906,
+      "learning_rate": 9.426254388822152e-06,
+      "loss": 43.8839,
+      "step": 353
+    },
+    {
+      "epoch": 0.6089013115459041,
+      "grad_norm": 9.736458778381348,
+      "learning_rate": 9.421799561652391e-06,
+      "loss": 44.6627,
+      "step": 354
+    },
+    {
+      "epoch": 0.6106213717480112,
+      "grad_norm": 11.647321701049805,
+      "learning_rate": 9.417328567702256e-06,
+      "loss": 43.6068,
+      "step": 355
+    },
+    {
+      "epoch": 0.6123414319501183,
+      "grad_norm": 12.150289535522461,
+      "learning_rate": 9.412841423318386e-06,
+      "loss": 43.6395,
+      "step": 356
+    },
+    {
+      "epoch": 0.6140614921522253,
+      "grad_norm": 9.2246675491333,
+      "learning_rate": 9.408338144906475e-06,
+      "loss": 44.1518,
+      "step": 357
+    },
+    {
+      "epoch": 0.6157815523543324,
+      "grad_norm": 10.310601234436035,
+      "learning_rate": 9.403818748931201e-06,
+      "loss": 43.9735,
+      "step": 358
+    },
+    {
+      "epoch": 0.6175016125564394,
+      "grad_norm": 8.350564002990723,
+      "learning_rate": 9.399283251916174e-06,
+      "loss": 43.7255,
+      "step": 359
+    },
+    {
+      "epoch": 0.6192216727585466,
+      "grad_norm": 9.116833686828613,
+      "learning_rate": 9.394731670443869e-06,
+      "loss": 43.6732,
+      "step": 360
+    },
+    {
+      "epoch": 0.6209417329606536,
+      "grad_norm": 8.664816856384277,
+      "learning_rate": 9.390164021155568e-06,
+      "loss": 43.6788,
+      "step": 361
+    },
+    {
+      "epoch": 0.6226617931627607,
+      "grad_norm": 8.421276092529297,
+      "learning_rate": 9.385580320751301e-06,
+      "loss": 43.7357,
+      "step": 362
+    },
+    {
+      "epoch": 0.6243818533648677,
+      "grad_norm": 7.4578680992126465,
+      "learning_rate": 9.380980585989782e-06,
+      "loss": 43.7452,
+      "step": 363
+    },
+    {
+      "epoch": 0.6261019135669749,
+      "grad_norm": 8.058576583862305,
+      "learning_rate": 9.376364833688352e-06,
+      "loss": 43.839,
+      "step": 364
+    },
+    {
+      "epoch": 0.6278219737690819,
+      "grad_norm": 7.826845169067383,
+      "learning_rate": 9.371733080722911e-06,
+      "loss": 43.8903,
+      "step": 365
+    },
+    {
+      "epoch": 0.629542033971189,
+      "grad_norm": 7.76792573928833,
+      "learning_rate": 9.367085344027862e-06,
+      "loss": 44.3217,
+      "step": 366
+    },
+    {
+      "epoch": 0.631262094173296,
+      "grad_norm": 8.252405166625977,
+      "learning_rate": 9.362421640596044e-06,
+      "loss": 44.0905,
+      "step": 367
+    },
+    {
+      "epoch": 0.6329821543754032,
+      "grad_norm": 7.431006908416748,
+      "learning_rate": 9.35774198747868e-06,
+      "loss": 43.864,
+      "step": 368
+    },
+    {
+      "epoch": 0.6347022145775102,
+      "grad_norm": 9.33600902557373,
+      "learning_rate": 9.353046401785297e-06,
+      "loss": 43.0342,
+      "step": 369
+    },
+    {
+      "epoch": 0.6364222747796173,
+      "grad_norm": 7.454495906829834,
+      "learning_rate": 9.348334900683685e-06,
+      "loss": 43.7442,
+      "step": 370
+    },
+    {
+      "epoch": 0.6381423349817243,
+      "grad_norm": 8.429414749145508,
+      "learning_rate": 9.343607501399812e-06,
+      "loss": 43.7836,
+      "step": 371
+    },
+    {
+      "epoch": 0.6398623951838315,
+      "grad_norm": 7.481090545654297,
+      "learning_rate": 9.338864221217783e-06,
+      "loss": 43.9994,
+      "step": 372
+    },
+    {
+      "epoch": 0.6415824553859385,
+      "grad_norm": 7.106781482696533,
+      "learning_rate": 9.33410507747976e-06,
+      "loss": 43.6922,
+      "step": 373
+    },
+    {
+      "epoch": 0.6433025155880456,
+      "grad_norm": 7.0425615310668945,
+      "learning_rate": 9.329330087585905e-06,
+      "loss": 44.5278,
+      "step": 374
+    },
+    {
+      "epoch": 0.6450225757901527,
+      "grad_norm": 7.197376728057861,
+      "learning_rate": 9.324539268994317e-06,
+      "loss": 43.8955,
+      "step": 375
+    },
+    {
+      "epoch": 0.6467426359922598,
+      "grad_norm": 6.9038286209106445,
+      "learning_rate": 9.319732639220965e-06,
+      "loss": 43.9268,
+      "step": 376
+    },
+    {
+      "epoch": 0.6484626961943668,
+      "grad_norm": 7.032724380493164,
+      "learning_rate": 9.31491021583963e-06,
+      "loss": 43.5221,
+      "step": 377
+    },
+    {
+      "epoch": 0.6501827563964738,
+      "grad_norm": 7.234856605529785,
+      "learning_rate": 9.310072016481832e-06,
+      "loss": 43.4951,
+      "step": 378
+    },
+    {
+      "epoch": 0.651902816598581,
+      "grad_norm": 6.546868801116943,
+      "learning_rate": 9.305218058836778e-06,
+      "loss": 44.0876,
+      "step": 379
+    },
+    {
+      "epoch": 0.653622876800688,
+      "grad_norm": 6.904932498931885,
+      "learning_rate": 9.300348360651282e-06,
+      "loss": 44.5423,
+      "step": 380
+    },
+    {
+      "epoch": 0.6553429370027951,
+      "grad_norm": 7.505612850189209,
+      "learning_rate": 9.295462939729711e-06,
+      "loss": 43.9986,
+      "step": 381
+    },
+    {
+      "epoch": 0.6570629972049021,
+      "grad_norm": 5.859259605407715,
+      "learning_rate": 9.290561813933916e-06,
+      "loss": 43.8683,
+      "step": 382
+    },
+    {
+      "epoch": 0.6587830574070093,
+      "grad_norm": 8.047765731811523,
+      "learning_rate": 9.285645001183167e-06,
+      "loss": 44.4658,
+      "step": 383
+    },
+    {
+      "epoch": 0.6605031176091163,
+      "grad_norm": 6.570570945739746,
+      "learning_rate": 9.280712519454092e-06,
+      "loss": 43.6115,
+      "step": 384
+    },
+    {
+      "epoch": 0.6622231778112234,
+      "grad_norm": 6.266587734222412,
+      "learning_rate": 9.2757643867806e-06,
+      "loss": 43.6822,
+      "step": 385
+    },
+    {
+      "epoch": 0.6639432380133304,
+      "grad_norm": 7.281513214111328,
+      "learning_rate": 9.270800621253833e-06,
+      "loss": 43.8285,
+      "step": 386
+    },
+    {
+      "epoch": 0.6656632982154376,
+      "grad_norm": 6.563234806060791,
+      "learning_rate": 9.265821241022074e-06,
+      "loss": 43.6976,
+      "step": 387
+    },
+    {
+      "epoch": 0.6673833584175446,
+      "grad_norm": 6.870432376861572,
+      "learning_rate": 9.26082626429071e-06,
+      "loss": 43.9566,
+      "step": 388
+    },
+    {
+      "epoch": 0.6691034186196517,
+      "grad_norm": 8.11976432800293,
+      "learning_rate": 9.255815709322142e-06,
+      "loss": 43.8613,
+      "step": 389
+    },
+    {
+      "epoch": 0.6708234788217587,
+      "grad_norm": 7.040714740753174,
+      "learning_rate": 9.250789594435735e-06,
+      "loss": 43.3387,
+      "step": 390
+    },
+    {
+      "epoch": 0.6725435390238659,
+      "grad_norm": 6.891185283660889,
+      "learning_rate": 9.245747938007734e-06,
+      "loss": 43.596,
+      "step": 391
+    },
+    {
+      "epoch": 0.6742635992259729,
+      "grad_norm": 7.045391082763672,
+      "learning_rate": 9.240690758471216e-06,
+      "loss": 43.1001,
+      "step": 392
+    },
+    {
+      "epoch": 0.67598365942808,
+      "grad_norm": 6.838486194610596,
+      "learning_rate": 9.235618074316005e-06,
+      "loss": 44.0918,
+      "step": 393
+    },
+    {
+      "epoch": 0.677703719630187,
+      "grad_norm": 8.006799697875977,
+      "learning_rate": 9.230529904088621e-06,
+      "loss": 43.4563,
+      "step": 394
+    },
+    {
+      "epoch": 0.6794237798322942,
+      "grad_norm": 7.786087989807129,
+      "learning_rate": 9.225426266392191e-06,
+      "loss": 44.1002,
+      "step": 395
+    },
+    {
+      "epoch": 0.6811438400344012,
+      "grad_norm": 7.782168388366699,
+      "learning_rate": 9.220307179886408e-06,
+      "loss": 44.15,
+      "step": 396
+    },
+    {
+      "epoch": 0.6828639002365082,
+      "grad_norm": 7.179986953735352,
+      "learning_rate": 9.215172663287435e-06,
+      "loss": 43.8326,
+      "step": 397
+    },
+    {
+      "epoch": 0.6845839604386154,
+      "grad_norm": 8.045145988464355,
+      "learning_rate": 9.210022735367857e-06,
+      "loss": 43.3196,
+      "step": 398
+    },
+    {
+      "epoch": 0.6863040206407224,
+      "grad_norm": 7.906603813171387,
+      "learning_rate": 9.204857414956606e-06,
+      "loss": 44.152,
+      "step": 399
+    },
+    {
+      "epoch": 0.6880240808428295,
+      "grad_norm": 8.266923904418945,
+      "learning_rate": 9.199676720938886e-06,
+      "loss": 44.158,
+      "step": 400
+    },
+    {
+      "epoch": 0.6897441410449365,
+      "grad_norm": 7.465760231018066,
+      "learning_rate": 9.194480672256117e-06,
+      "loss": 43.9078,
+      "step": 401
+    },
+    {
+      "epoch": 0.6914642012470437,
+      "grad_norm": 7.522243499755859,
+      "learning_rate": 9.189269287905849e-06,
+      "loss": 43.3097,
+      "step": 402
+    },
+    {
+      "epoch": 0.6931842614491507,
+      "grad_norm": 6.484007358551025,
+      "learning_rate": 9.184042586941708e-06,
+      "loss": 43.9014,
+      "step": 403
+    },
+    {
+      "epoch": 0.6949043216512578,
+      "grad_norm": 6.548778533935547,
+      "learning_rate": 9.178800588473317e-06,
+      "loss": 43.9104,
+      "step": 404
+    },
+    {
+      "epoch": 0.6966243818533648,
+      "grad_norm": 6.578863620758057,
+      "learning_rate": 9.17354331166623e-06,
+      "loss": 44.0022,
+      "step": 405
+    },
+    {
+      "epoch": 0.698344442055472,
+      "grad_norm": 7.083658695220947,
+      "learning_rate": 9.168270775741863e-06,
+      "loss": 43.7902,
+      "step": 406
+    },
+    {
+      "epoch": 0.700064502257579,
+      "grad_norm": 7.241711139678955,
+      "learning_rate": 9.162982999977417e-06,
+      "loss": 43.9464,
+      "step": 407
+    },
+    {
+      "epoch": 0.7017845624596861,
+      "grad_norm": 6.329436779022217,
+      "learning_rate": 9.157680003705816e-06,
+      "loss": 44.1072,
+      "step": 408
+    },
+    {
+      "epoch": 0.7035046226617931,
+      "grad_norm": 6.435650825500488,
+      "learning_rate": 9.15236180631563e-06,
+      "loss": 43.392,
+      "step": 409
+    },
+    {
+      "epoch": 0.7052246828639003,
+      "grad_norm": 5.681223392486572,
+      "learning_rate": 9.14702842725101e-06,
+      "loss": 44.3148,
+      "step": 410
+    },
+    {
+      "epoch": 0.7069447430660073,
+      "grad_norm": 6.71289587020874,
+      "learning_rate": 9.14167988601161e-06,
+      "loss": 43.8893,
+      "step": 411
+    },
+    {
+      "epoch": 0.7086648032681144,
+      "grad_norm": 6.497440814971924,
+      "learning_rate": 9.13631620215252e-06,
+      "loss": 44.1776,
+      "step": 412
+    },
+    {
+      "epoch": 0.7103848634702215,
+      "grad_norm": 7.291422367095947,
+      "learning_rate": 9.130937395284199e-06,
+      "loss": 43.8195,
+      "step": 413
+    },
+    {
+      "epoch": 0.7121049236723286,
+      "grad_norm": 6.935153961181641,
+      "learning_rate": 9.125543485072386e-06,
+      "loss": 43.9977,
+      "step": 414
+    },
+    {
+      "epoch": 0.7138249838744356,
+      "grad_norm": 6.302245140075684,
+      "learning_rate": 9.120134491238054e-06,
+      "loss": 43.677,
+      "step": 415
+    },
+    {
+      "epoch": 0.7155450440765426,
+      "grad_norm": 6.205868244171143,
+      "learning_rate": 9.114710433557314e-06,
+      "loss": 43.8423,
+      "step": 416
+    },
+    {
+      "epoch": 0.7172651042786498,
+      "grad_norm": 5.34831428527832,
+      "learning_rate": 9.109271331861361e-06,
+      "loss": 43.6707,
+      "step": 417
+    },
+    {
+      "epoch": 0.7189851644807568,
+      "grad_norm": 7.174152374267578,
+      "learning_rate": 9.103817206036383e-06,
+      "loss": 43.3579,
+      "step": 418
+    },
+    {
+      "epoch": 0.7207052246828639,
+      "grad_norm": 6.666977882385254,
+      "learning_rate": 9.098348076023506e-06,
+      "loss": 43.8424,
+      "step": 419
+    },
+    {
+      "epoch": 0.7224252848849709,
+      "grad_norm": 7.491025924682617,
+      "learning_rate": 9.092863961818715e-06,
+      "loss": 44.4333,
+      "step": 420
+    },
+    {
+      "epoch": 0.7241453450870781,
+      "grad_norm": 6.508261680603027,
+      "learning_rate": 9.087364883472774e-06,
+      "loss": 43.7001,
+      "step": 421
+    },
+    {
+      "epoch": 0.7258654052891851,
+      "grad_norm": 7.418080806732178,
+      "learning_rate": 9.08185086109116e-06,
+      "loss": 44.0001,
+      "step": 422
+    },
+    {
+      "epoch": 0.7275854654912922,
+      "grad_norm": 6.983603000640869,
+      "learning_rate": 9.076321914833988e-06,
+      "loss": 44.6241,
+      "step": 423
+    },
+    {
+      "epoch": 0.7293055256933992,
+      "grad_norm": 8.667305946350098,
+      "learning_rate": 9.070778064915937e-06,
+      "loss": 44.1089,
+      "step": 424
+    },
+    {
+      "epoch": 0.7310255858955064,
+      "grad_norm": 7.419984340667725,
+      "learning_rate": 9.065219331606182e-06,
+      "loss": 43.9046,
+      "step": 425
+    },
+    {
+      "epoch": 0.7327456460976134,
+      "grad_norm": 7.34318733215332,
+      "learning_rate": 9.0596457352283e-06,
+      "loss": 43.794,
+      "step": 426
+    },
+    {
+      "epoch": 0.7344657062997205,
+      "grad_norm": 7.931493759155273,
+      "learning_rate": 9.054057296160221e-06,
+      "loss": 44.6317,
+      "step": 427
+    },
+    {
+      "epoch": 0.7361857665018275,
+      "grad_norm": 6.583981037139893,
+      "learning_rate": 9.048454034834143e-06,
+      "loss": 43.5199,
+      "step": 428
+    },
+    {
+      "epoch": 0.7379058267039347,
+      "grad_norm": 8.499653816223145,
+      "learning_rate": 9.042835971736446e-06,
+      "loss": 43.8616,
+      "step": 429
+    },
+    {
+      "epoch": 0.7396258869060417,
+      "grad_norm": 6.757936000823975,
+      "learning_rate": 9.037203127407642e-06,
+      "loss": 44.0385,
+      "step": 430
+    },
+    {
+      "epoch": 0.7413459471081488,
+      "grad_norm": 8.30978012084961,
+      "learning_rate": 9.031555522442268e-06,
+      "loss": 43.4628,
+      "step": 431
+    },
+    {
+      "epoch": 0.7430660073102558,
+      "grad_norm": 7.357321262359619,
+      "learning_rate": 9.025893177488848e-06,
+      "loss": 43.6677,
+      "step": 432
+    },
+    {
+      "epoch": 0.744786067512363,
+      "grad_norm": 8.4613676071167,
+      "learning_rate": 9.02021611324978e-06,
+      "loss": 43.2542,
+      "step": 433
+    },
+    {
+      "epoch": 0.74650612771447,
+      "grad_norm": 8.782477378845215,
+      "learning_rate": 9.014524350481287e-06,
+      "loss": 44.0515,
+      "step": 434
+    },
+    {
+      "epoch": 0.7482261879165771,
+      "grad_norm": 6.701351165771484,
+      "learning_rate": 9.008817909993332e-06,
+      "loss": 44.0643,
+      "step": 435
+    },
+    {
+      "epoch": 0.7499462481186842,
+      "grad_norm": 7.1624884605407715,
+      "learning_rate": 9.00309681264954e-06,
+      "loss": 44.0639,
+      "step": 436
+    },
+    {
+      "epoch": 0.7516663083207912,
+      "grad_norm": 6.229190349578857,
+      "learning_rate": 8.997361079367124e-06,
+      "loss": 43.9093,
+      "step": 437
+    },
+    {
+      "epoch": 0.7533863685228983,
+      "grad_norm": 7.180543422698975,
+      "learning_rate": 8.991610731116808e-06,
+      "loss": 44.2652,
+      "step": 438
+    },
+    {
+      "epoch": 0.7551064287250053,
+      "grad_norm": 6.3346781730651855,
+      "learning_rate": 8.985845788922753e-06,
+      "loss": 43.2561,
+      "step": 439
+    },
+    {
+      "epoch": 0.7568264889271125,
+      "grad_norm": 7.270414352416992,
+      "learning_rate": 8.980066273862473e-06,
+      "loss": 43.9074,
+      "step": 440
+    },
+    {
+      "epoch": 0.7585465491292195,
+      "grad_norm": 6.431163311004639,
+      "learning_rate": 8.974272207066767e-06,
+      "loss": 43.9343,
+      "step": 441
+    },
+    {
+      "epoch": 0.7602666093313266,
+      "grad_norm": 6.415679931640625,
+      "learning_rate": 8.968463609719636e-06,
+      "loss": 44.2067,
+      "step": 442
+    },
+    {
+      "epoch": 0.7619866695334336,
+      "grad_norm": 6.083033561706543,
+      "learning_rate": 8.962640503058206e-06,
+      "loss": 43.7967,
+      "step": 443
+    },
+    {
+      "epoch": 0.7637067297355408,
+      "grad_norm": 6.942599773406982,
+      "learning_rate": 8.956802908372652e-06,
+      "loss": 43.4928,
+      "step": 444
+    },
+    {
+      "epoch": 0.7654267899376478,
+      "grad_norm": 6.515557765960693,
+      "learning_rate": 8.95095084700612e-06,
+      "loss": 43.6577,
+      "step": 445
+    },
+    {
+      "epoch": 0.7671468501397549,
+      "grad_norm": 7.167238235473633,
+      "learning_rate": 8.945084340354646e-06,
+      "loss": 43.4542,
+      "step": 446
+    },
+    {
+      "epoch": 0.7688669103418619,
+      "grad_norm": 7.30296516418457,
+      "learning_rate": 8.939203409867084e-06,
+      "loss": 43.123,
+      "step": 447
+    },
+    {
+      "epoch": 0.7705869705439691,
+      "grad_norm": 7.387278079986572,
+      "learning_rate": 8.933308077045022e-06,
+      "loss": 43.6603,
+      "step": 448
+    },
+    {
+      "epoch": 0.7723070307460761,
+      "grad_norm": 7.026780128479004,
+      "learning_rate": 8.927398363442705e-06,
+      "loss": 44.0497,
+      "step": 449
+    },
+    {
+      "epoch": 0.7740270909481832,
+      "grad_norm": 7.03558874130249,
+      "learning_rate": 8.921474290666955e-06,
+      "loss": 43.1461,
+      "step": 450
+    },
+    {
+      "epoch": 0.7757471511502902,
+      "grad_norm": 8.48353099822998,
+      "learning_rate": 8.915535880377096e-06,
+      "loss": 43.6771,
+      "step": 451
+    },
+    {
+      "epoch": 0.7774672113523974,
+      "grad_norm": 7.250082015991211,
+      "learning_rate": 8.909583154284868e-06,
+      "loss": 43.9369,
+      "step": 452
+    },
+    {
+      "epoch": 0.7791872715545044,
+      "grad_norm": 8.958197593688965,
+      "learning_rate": 8.90361613415436e-06,
+      "loss": 43.7911,
+      "step": 453
+    },
+    {
+      "epoch": 0.7809073317566115,
+      "grad_norm": 8.168319702148438,
+      "learning_rate": 8.897634841801911e-06,
+      "loss": 43.3905,
+      "step": 454
+    },
+    {
+      "epoch": 0.7826273919587186,
+      "grad_norm": 8.520408630371094,
+      "learning_rate": 8.891639299096051e-06,
+      "loss": 43.3708,
+      "step": 455
+    },
+    {
+      "epoch": 0.7843474521608256,
+      "grad_norm": 8.194758415222168,
+      "learning_rate": 8.885629527957407e-06,
+      "loss": 43.3692,
+      "step": 456
+    },
+    {
+      "epoch": 0.7860675123629327,
+      "grad_norm": 7.554206848144531,
+      "learning_rate": 8.879605550358627e-06,
+      "loss": 43.5693,
+      "step": 457
+    },
+    {
+      "epoch": 0.7877875725650397,
+      "grad_norm": 7.146202087402344,
+      "learning_rate": 8.873567388324302e-06,
+      "loss": 43.8261,
+      "step": 458
+    },
+    {
+      "epoch": 0.7895076327671469,
+      "grad_norm": 6.808493137359619,
+      "learning_rate": 8.867515063930881e-06,
+      "loss": 43.3648,
+      "step": 459
+    },
+    {
+      "epoch": 0.7912276929692539,
+      "grad_norm": 7.165658950805664,
+      "learning_rate": 8.861448599306597e-06,
+      "loss": 43.0367,
+      "step": 460
+    },
+    {
+      "epoch": 0.792947753171361,
+      "grad_norm": 6.527984142303467,
+      "learning_rate": 8.855368016631377e-06,
+      "loss": 43.4491,
+      "step": 461
+    },
+    {
+      "epoch": 0.794667813373468,
+      "grad_norm": 6.912752628326416,
+      "learning_rate": 8.849273338136772e-06,
+      "loss": 43.6405,
+      "step": 462
+    },
+    {
+      "epoch": 0.7963878735755752,
+      "grad_norm": 6.334918975830078,
+      "learning_rate": 8.84316458610586e-06,
+      "loss": 44.4637,
+      "step": 463
+    },
+    {
+      "epoch": 0.7981079337776822,
+      "grad_norm": 7.22133207321167,
+      "learning_rate": 8.837041782873182e-06,
+      "loss": 43.2829,
+      "step": 464
+    },
+    {
+      "epoch": 0.7998279939797893,
+      "grad_norm": 6.233572006225586,
+      "learning_rate": 8.83090495082465e-06,
+      "loss": 43.3993,
+      "step": 465
+    },
+    {
+      "epoch": 0.8015480541818963,
+      "grad_norm": 6.681156635284424,
+      "learning_rate": 8.824754112397467e-06,
+      "loss": 43.2356,
+      "step": 466
+    },
+    {
+      "epoch": 0.8032681143840035,
+      "grad_norm": 7.240959167480469,
+      "learning_rate": 8.818589290080043e-06,
+      "loss": 42.8966,
+      "step": 467
+    },
+    {
+      "epoch": 0.8049881745861105,
+      "grad_norm": 6.253081798553467,
+      "learning_rate": 8.812410506411925e-06,
+      "loss": 43.9822,
+      "step": 468
+    },
+    {
+      "epoch": 0.8067082347882176,
+      "grad_norm": 7.541505813598633,
+      "learning_rate": 8.806217783983693e-06,
+      "loss": 43.9604,
+      "step": 469
+    },
+    {
+      "epoch": 0.8084282949903246,
+      "grad_norm": 7.4928483963012695,
+      "learning_rate": 8.800011145436893e-06,
+      "loss": 43.8446,
+      "step": 470
+    },
+    {
+      "epoch": 0.8101483551924318,
+      "grad_norm": 6.140499591827393,
+      "learning_rate": 8.793790613463956e-06,
+      "loss": 43.6913,
+      "step": 471
+    },
+    {
+      "epoch": 0.8118684153945388,
+      "grad_norm": 7.944373607635498,
+      "learning_rate": 8.787556210808101e-06,
+      "loss": 43.4474,
+      "step": 472
+    },
+    {
+      "epoch": 0.8135884755966459,
+      "grad_norm": 6.9422101974487305,
+      "learning_rate": 8.781307960263267e-06,
+      "loss": 43.293,
+      "step": 473
+    },
+    {
+      "epoch": 0.815308535798753,
+      "grad_norm": 6.664095878601074,
+      "learning_rate": 8.77504588467402e-06,
+      "loss": 43.8227,
+      "step": 474
+    },
+    {
+      "epoch": 0.81702859600086,
+      "grad_norm": 7.298461437225342,
+      "learning_rate": 8.768770006935475e-06,
+      "loss": 43.7175,
+      "step": 475
+    },
+    {
+      "epoch": 0.8187486562029671,
+      "grad_norm": 6.43251895904541,
+      "learning_rate": 8.762480349993204e-06,
+      "loss": 43.143,
+      "step": 476
+    },
+    {
+      "epoch": 0.8204687164050741,
+      "grad_norm": 6.303859233856201,
+      "learning_rate": 8.756176936843161e-06,
+      "loss": 43.7655,
+      "step": 477
+    },
+    {
+      "epoch": 0.8221887766071813,
+      "grad_norm": 6.824503421783447,
+      "learning_rate": 8.749859790531601e-06,
+      "loss": 43.5909,
+      "step": 478
+    },
+    {
+      "epoch": 0.8239088368092883,
+      "grad_norm": 6.232965469360352,
+      "learning_rate": 8.743528934154982e-06,
+      "loss": 43.6798,
+      "step": 479
+    },
+    {
+      "epoch": 0.8256288970113954,
+      "grad_norm": 6.288873672485352,
+      "learning_rate": 8.737184390859887e-06,
+      "loss": 43.4713,
+      "step": 480
+    },
+    {
+      "epoch": 0.8273489572135024,
+      "grad_norm": 6.1072306632995605,
+      "learning_rate": 8.730826183842947e-06,
+      "loss": 43.4521,
+      "step": 481
+    },
+    {
+      "epoch": 0.8290690174156096,
+      "grad_norm": 7.3213701248168945,
+      "learning_rate": 8.724454336350742e-06,
+      "loss": 43.9662,
+      "step": 482
+    },
+    {
+      "epoch": 0.8307890776177166,
+      "grad_norm": 6.282354354858398,
+      "learning_rate": 8.718068871679735e-06,
+      "loss": 44.1781,
+      "step": 483
+    },
+    {
+      "epoch": 0.8325091378198237,
+      "grad_norm": 7.692941188812256,
+      "learning_rate": 8.711669813176165e-06,
+      "loss": 43.5585,
+      "step": 484
+    },
+    {
+      "epoch": 0.8342291980219307,
+      "grad_norm": 6.070176124572754,
+      "learning_rate": 8.705257184235973e-06,
+      "loss": 43.843,
+      "step": 485
+    },
+    {
+      "epoch": 0.8359492582240379,
+      "grad_norm": 7.584023952484131,
+      "learning_rate": 8.698831008304723e-06,
+      "loss": 43.5888,
+      "step": 486
+    },
+    {
+      "epoch": 0.8376693184261449,
+      "grad_norm": 8.037973403930664,
+      "learning_rate": 8.6923913088775e-06,
+      "loss": 43.4765,
+      "step": 487
+    },
+    {
+      "epoch": 0.839389378628252,
+      "grad_norm": 6.745630741119385,
+      "learning_rate": 8.685938109498839e-06,
+      "loss": 44.0438,
+      "step": 488
+    },
+    {
+      "epoch": 0.841109438830359,
+      "grad_norm": 6.660660743713379,
+      "learning_rate": 8.679471433762633e-06,
+      "loss": 43.5884,
+      "step": 489
+    },
+    {
+      "epoch": 0.8428294990324662,
+      "grad_norm": 7.205166339874268,
+      "learning_rate": 8.672991305312042e-06,
+      "loss": 43.5902,
+      "step": 490
+    },
+    {
+      "epoch": 0.8445495592345732,
+      "grad_norm": 6.969662666320801,
+      "learning_rate": 8.666497747839413e-06,
+      "loss": 43.339,
+      "step": 491
+    },
+    {
+      "epoch": 0.8462696194366803,
+      "grad_norm": 6.3066229820251465,
+      "learning_rate": 8.659990785086195e-06,
+      "loss": 43.8102,
+      "step": 492
+    },
+    {
+      "epoch": 0.8479896796387874,
+      "grad_norm": 7.804117202758789,
+      "learning_rate": 8.653470440842847e-06,
+      "loss": 43.7162,
+      "step": 493
+    },
+    {
+      "epoch": 0.8497097398408945,
+      "grad_norm": 6.339798450469971,
+      "learning_rate": 8.646936738948747e-06,
+      "loss": 43.3229,
+      "step": 494
+    },
+    {
+      "epoch": 0.8514298000430015,
+      "grad_norm": 8.31767749786377,
+      "learning_rate": 8.64038970329212e-06,
+      "loss": 43.8772,
+      "step": 495
+    },
+    {
+      "epoch": 0.8531498602451085,
+      "grad_norm": 7.365615367889404,
+      "learning_rate": 8.633829357809937e-06,
+      "loss": 43.2881,
+      "step": 496
+    },
+    {
+      "epoch": 0.8548699204472157,
+      "grad_norm": 8.474952697753906,
+      "learning_rate": 8.627255726487831e-06,
+      "loss": 43.7939,
+      "step": 497
+    },
+    {
+      "epoch": 0.8565899806493227,
+      "grad_norm": 6.960323333740234,
+      "learning_rate": 8.620668833360009e-06,
+      "loss": 43.499,
+      "step": 498
+    },
+    {
+      "epoch": 0.8583100408514298,
+      "grad_norm": 7.6085734367370605,
+      "learning_rate": 8.614068702509169e-06,
+      "loss": 43.2025,
+      "step": 499
+    },
+    {
+      "epoch": 0.8600301010535368,
+      "grad_norm": 5.805634498596191,
+      "learning_rate": 8.607455358066404e-06,
+      "loss": 44.0489,
+      "step": 500
+    },
+    {
+      "epoch": 0.861750161255644,
+      "grad_norm": 7.452625274658203,
+      "learning_rate": 8.600828824211122e-06,
+      "loss": 43.147,
+      "step": 501
+    },
+    {
+      "epoch": 0.863470221457751,
+      "grad_norm": 6.4528584480285645,
+      "learning_rate": 8.594189125170952e-06,
+      "loss": 43.5228,
+      "step": 502
+    },
+    {
+      "epoch": 0.8651902816598581,
+      "grad_norm": 7.375026702880859,
+      "learning_rate": 8.587536285221656e-06,
+      "loss": 44.272,
+      "step": 503
+    },
+    {
+      "epoch": 0.8669103418619651,
+      "grad_norm": 6.547872543334961,
+      "learning_rate": 8.580870328687041e-06,
+      "loss": 43.4759,
+      "step": 504
+    },
+    {
+      "epoch": 0.8686304020640723,
+      "grad_norm": 7.088862419128418,
+      "learning_rate": 8.574191279938872e-06,
+      "loss": 43.5528,
+      "step": 505
+    },
+    {
+      "epoch": 0.8703504622661793,
+      "grad_norm": 6.6691083908081055,
+      "learning_rate": 8.567499163396777e-06,
+      "loss": 43.4488,
+      "step": 506
+    },
+    {
+      "epoch": 0.8720705224682864,
+      "grad_norm": 7.416652202606201,
+      "learning_rate": 8.560794003528171e-06,
+      "loss": 43.5291,
+      "step": 507
+    },
+    {
+      "epoch": 0.8737905826703934,
+      "grad_norm": 5.969050884246826,
+      "learning_rate": 8.554075824848146e-06,
+      "loss": 43.5905,
+      "step": 508
+    },
+    {
+      "epoch": 0.8755106428725006,
+      "grad_norm": 7.501400470733643,
+      "learning_rate": 8.5473446519194e-06,
+      "loss": 44.0266,
+      "step": 509
+    },
+    {
+      "epoch": 0.8772307030746076,
+      "grad_norm": 6.1612548828125,
+      "learning_rate": 8.540600509352139e-06,
+      "loss": 44.2164,
+      "step": 510
+    },
+    {
+      "epoch": 0.8789507632767147,
+      "grad_norm": 7.144975662231445,
+      "learning_rate": 8.533843421803985e-06,
+      "loss": 43.1628,
+      "step": 511
+    },
+    {
+      "epoch": 0.8806708234788218,
+      "grad_norm": 6.905309200286865,
+      "learning_rate": 8.527073413979894e-06,
+      "loss": 43.8717,
+      "step": 512
+    },
+    {
+      "epoch": 0.8823908836809289,
+      "grad_norm": 7.095192909240723,
+      "learning_rate": 8.520290510632055e-06,
+      "loss": 43.612,
+      "step": 513
+    },
+    {
+      "epoch": 0.8841109438830359,
+      "grad_norm": 6.804945468902588,
+      "learning_rate": 8.51349473655981e-06,
+      "loss": 43.4807,
+      "step": 514
+    },
+    {
+      "epoch": 0.8858310040851429,
+      "grad_norm": 8.927321434020996,
+      "learning_rate": 8.506686116609553e-06,
+      "loss": 43.5119,
+      "step": 515
+    },
+    {
+      "epoch": 0.8875510642872501,
+      "grad_norm": 6.946136474609375,
+      "learning_rate": 8.499864675674648e-06,
+      "loss": 43.1227,
+      "step": 516
+    },
+    {
+      "epoch": 0.8892711244893571,
+      "grad_norm": 11.003009796142578,
+      "learning_rate": 8.493030438695336e-06,
+      "loss": 43.2844,
+      "step": 517
+    },
+    {
+      "epoch": 0.8909911846914642,
+      "grad_norm": 9.151321411132812,
+      "learning_rate": 8.486183430658639e-06,
+      "loss": 43.9351,
+      "step": 518
+    },
+    {
+      "epoch": 0.8927112448935712,
+      "grad_norm": 10.419197082519531,
+      "learning_rate": 8.479323676598271e-06,
+      "loss": 43.4703,
+      "step": 519
+    },
+    {
+      "epoch": 0.8944313050956784,
+      "grad_norm": 9.973617553710938,
+      "learning_rate": 8.472451201594556e-06,
+      "loss": 43.2196,
+      "step": 520
+    },
+    {
+      "epoch": 0.8961513652977854,
+      "grad_norm": 7.706507682800293,
+      "learning_rate": 8.465566030774314e-06,
+      "loss": 43.3206,
+      "step": 521
+    },
+    {
+      "epoch": 0.8978714254998925,
+      "grad_norm": 7.455542087554932,
+      "learning_rate": 8.458668189310793e-06,
+      "loss": 43.2529,
+      "step": 522
+    },
+    {
+      "epoch": 0.8995914857019995,
+      "grad_norm": 7.595444202423096,
+      "learning_rate": 8.451757702423566e-06,
+      "loss": 43.5217,
+      "step": 523
+    },
+    {
+      "epoch": 0.9013115459041067,
+      "grad_norm": 7.0413899421691895,
+      "learning_rate": 8.444834595378434e-06,
+      "loss": 43.6686,
+      "step": 524
+    },
+    {
+      "epoch": 0.9030316061062137,
+      "grad_norm": 7.505328178405762,
+      "learning_rate": 8.437898893487345e-06,
+      "loss": 43.508,
+      "step": 525
+    },
+    {
+      "epoch": 0.9047516663083208,
+      "grad_norm": 6.636236667633057,
+      "learning_rate": 8.430950622108292e-06,
+      "loss": 43.6455,
+      "step": 526
+    },
+    {
+      "epoch": 0.9064717265104278,
+      "grad_norm": 8.014334678649902,
+      "learning_rate": 8.42398980664523e-06,
+      "loss": 43.9419,
+      "step": 527
+    },
+    {
+      "epoch": 0.908191786712535,
+      "grad_norm": 6.861055374145508,
+      "learning_rate": 8.417016472547968e-06,
+      "loss": 44.0091,
+      "step": 528
+    },
+    {
+      "epoch": 0.909911846914642,
+      "grad_norm": 7.528046607971191,
+      "learning_rate": 8.41003064531209e-06,
+      "loss": 43.9412,
+      "step": 529
+    },
+    {
+      "epoch": 0.9116319071167491,
+      "grad_norm": 6.380741596221924,
+      "learning_rate": 8.403032350478857e-06,
+      "loss": 43.1688,
+      "step": 530
+    },
+    {
+      "epoch": 0.9133519673188562,
+      "grad_norm": 8.248329162597656,
+      "learning_rate": 8.396021613635116e-06,
+      "loss": 43.241,
+      "step": 531
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 1743,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 59,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.448505275235225e+19,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}