Upload initial model

.gitattributes

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

README.md

@@ -0,0 +1,11 @@
+datasets:
+  - GX-XinGao/GRA
+language:
+  - en
+base_model:
+  - Qwen/Qwen2.5-7B
+license: apache-2.0
+metrics:
+  - accuracy
+pipeline_tag: text-generation
+library_name: transformers

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
+}

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9969306322897483,
+    "total_flos": 2.835499283468255e+18,
+    "train_loss": 0.40513286713896124,
+    "train_runtime": 4111.0607,
+    "train_samples_per_second": 12.677,
+    "train_steps_per_second": 0.049
+}

config.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "/mnt/petrelfs/gaoxin/Qwen2.5-7B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "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": "4.46.1",
+  "use_cache": false,
+  "use_mrope": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.46.1"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d158dc99abb2b853d206a1525af49ba18b51e85c49f9d41b5c532eaa7c27e67
+size 4877660776

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f095fe1cfb4d87f02905d7712f084edae8207c4408320096763ecc5bde65c576
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d87e64356d2182fa0b1b445013153c2919c6cb594bcf94fc065385797229c789
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcd80cd928930108b4d98126766ae12af68c1bc544141001e7f7dc2f2eee54a8
+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.json

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

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": 4096,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9969306322897483,
+    "total_flos": 2.835499283468255e+18,
+    "train_loss": 0.40513286713896124,
+    "train_runtime": 4111.0607,
+    "train_samples_per_second": 12.677,
+    "train_steps_per_second": 0.049
+}

trainer_log.jsonl

@@ -0,0 +1,204 @@
+{"current_steps": 1, "total_steps": 203, "loss": 0.5458, "lr": 7.142857142857143e-07, "epoch": 0.004910988336402701, "percentage": 0.49, "elapsed_time": "0:00:22", "remaining_time": "1:14:32"}
+{"current_steps": 2, "total_steps": 203, "loss": 0.4966, "lr": 1.4285714285714286e-06, "epoch": 0.009821976672805401, "percentage": 0.99, "elapsed_time": "0:00:44", "remaining_time": "1:14:41"}
+{"current_steps": 3, "total_steps": 203, "loss": 0.5349, "lr": 2.1428571428571427e-06, "epoch": 0.014732965009208104, "percentage": 1.48, "elapsed_time": "0:01:06", "remaining_time": "1:13:53"}
+{"current_steps": 4, "total_steps": 203, "loss": 0.5184, "lr": 2.8571428571428573e-06, "epoch": 0.019643953345610803, "percentage": 1.97, "elapsed_time": "0:01:27", "remaining_time": "1:12:25"}
+{"current_steps": 5, "total_steps": 203, "loss": 0.5031, "lr": 3.5714285714285718e-06, "epoch": 0.024554941682013505, "percentage": 2.46, "elapsed_time": "0:01:52", "remaining_time": "1:14:09"}
+{"current_steps": 6, "total_steps": 203, "loss": 0.5246, "lr": 4.2857142857142855e-06, "epoch": 0.029465930018416207, "percentage": 2.96, "elapsed_time": "0:02:10", "remaining_time": "1:11:31"}
+{"current_steps": 7, "total_steps": 203, "loss": 0.5022, "lr": 5e-06, "epoch": 0.03437691835481891, "percentage": 3.45, "elapsed_time": "0:02:28", "remaining_time": "1:09:22"}
+{"current_steps": 8, "total_steps": 203, "loss": 0.4882, "lr": 4.999678864499828e-06, "epoch": 0.039287906691221605, "percentage": 3.94, "elapsed_time": "0:02:46", "remaining_time": "1:07:38"}
+{"current_steps": 9, "total_steps": 203, "loss": 0.4695, "lr": 4.99871554050172e-06, "epoch": 0.04419889502762431, "percentage": 4.43, "elapsed_time": "0:03:09", "remaining_time": "1:07:57"}
+{"current_steps": 10, "total_steps": 203, "loss": 0.4773, "lr": 4.997110275491702e-06, "epoch": 0.04910988336402701, "percentage": 4.93, "elapsed_time": "0:03:29", "remaining_time": "1:07:24"}
+{"current_steps": 11, "total_steps": 203, "loss": 0.469, "lr": 4.994863481875842e-06, "epoch": 0.05402087170042971, "percentage": 5.42, "elapsed_time": "0:03:51", "remaining_time": "1:07:15"}
+{"current_steps": 12, "total_steps": 203, "loss": 0.4836, "lr": 4.9919757368742895e-06, "epoch": 0.058931860036832415, "percentage": 5.91, "elapsed_time": "0:04:09", "remaining_time": "1:06:06"}
+{"current_steps": 13, "total_steps": 203, "loss": 0.4308, "lr": 4.988447782372996e-06, "epoch": 0.06384284837323512, "percentage": 6.4, "elapsed_time": "0:04:29", "remaining_time": "1:05:43"}
+{"current_steps": 14, "total_steps": 203, "loss": 0.4609, "lr": 4.984280524733107e-06, "epoch": 0.06875383670963782, "percentage": 6.9, "elapsed_time": "0:04:47", "remaining_time": "1:04:39"}
+{"current_steps": 15, "total_steps": 203, "loss": 0.4194, "lr": 4.979475034558115e-06, "epoch": 0.07366482504604052, "percentage": 7.39, "elapsed_time": "0:05:05", "remaining_time": "1:03:43"}
+{"current_steps": 16, "total_steps": 203, "loss": 0.4366, "lr": 4.974032546418816e-06, "epoch": 0.07857581338244321, "percentage": 7.88, "elapsed_time": "0:05:26", "remaining_time": "1:03:38"}
+{"current_steps": 17, "total_steps": 203, "loss": 0.444, "lr": 4.967954458536126e-06, "epoch": 0.08348680171884591, "percentage": 8.37, "elapsed_time": "0:05:47", "remaining_time": "1:03:27"}
+{"current_steps": 18, "total_steps": 203, "loss": 0.4538, "lr": 4.9612423324218816e-06, "epoch": 0.08839779005524862, "percentage": 8.87, "elapsed_time": "0:06:08", "remaining_time": "1:03:06"}
+{"current_steps": 19, "total_steps": 203, "loss": 0.4498, "lr": 4.953897892477664e-06, "epoch": 0.09330877839165132, "percentage": 9.36, "elapsed_time": "0:06:29", "remaining_time": "1:02:55"}
+{"current_steps": 20, "total_steps": 203, "loss": 0.4537, "lr": 4.945923025551789e-06, "epoch": 0.09821976672805402, "percentage": 9.85, "elapsed_time": "0:06:48", "remaining_time": "1:02:13"}
+{"current_steps": 21, "total_steps": 203, "loss": 0.4524, "lr": 4.937319780454559e-06, "epoch": 0.10313075506445672, "percentage": 10.34, "elapsed_time": "0:07:05", "remaining_time": "1:01:25"}
+{"current_steps": 22, "total_steps": 203, "loss": 0.4174, "lr": 4.92809036743191e-06, "epoch": 0.10804174340085942, "percentage": 10.84, "elapsed_time": "0:07:29", "remaining_time": "1:01:35"}
+{"current_steps": 23, "total_steps": 203, "loss": 0.4343, "lr": 4.918237157597574e-06, "epoch": 0.11295273173726213, "percentage": 11.33, "elapsed_time": "0:07:49", "remaining_time": "1:01:17"}
+{"current_steps": 24, "total_steps": 203, "loss": 0.4104, "lr": 4.907762682323926e-06, "epoch": 0.11786372007366483, "percentage": 11.82, "elapsed_time": "0:08:10", "remaining_time": "1:00:59"}
+{"current_steps": 25, "total_steps": 203, "loss": 0.433, "lr": 4.896669632591652e-06, "epoch": 0.12277470841006753, "percentage": 12.32, "elapsed_time": "0:08:29", "remaining_time": "1:00:26"}
+{"current_steps": 26, "total_steps": 203, "loss": 0.4344, "lr": 4.88496085829841e-06, "epoch": 0.12768569674647023, "percentage": 12.81, "elapsed_time": "0:08:46", "remaining_time": "0:59:41"}
+{"current_steps": 27, "total_steps": 203, "loss": 0.4349, "lr": 4.872639367526672e-06, "epoch": 0.13259668508287292, "percentage": 13.3, "elapsed_time": "0:09:04", "remaining_time": "0:59:11"}
+{"current_steps": 28, "total_steps": 203, "loss": 0.4284, "lr": 4.859708325770919e-06, "epoch": 0.13750767341927564, "percentage": 13.79, "elapsed_time": "0:09:23", "remaining_time": "0:58:40"}
+{"current_steps": 29, "total_steps": 203, "loss": 0.4298, "lr": 4.846171055124401e-06, "epoch": 0.14241866175567833, "percentage": 14.29, "elapsed_time": "0:09:42", "remaining_time": "0:58:12"}
+{"current_steps": 30, "total_steps": 203, "loss": 0.4325, "lr": 4.832031033425663e-06, "epoch": 0.14732965009208104, "percentage": 14.78, "elapsed_time": "0:10:01", "remaining_time": "0:57:46"}
+{"current_steps": 31, "total_steps": 203, "loss": 0.4317, "lr": 4.817291893365055e-06, "epoch": 0.15224063842848373, "percentage": 15.27, "elapsed_time": "0:10:18", "remaining_time": "0:57:13"}
+{"current_steps": 32, "total_steps": 203, "loss": 0.4297, "lr": 4.8019574215514705e-06, "epoch": 0.15715162676488642, "percentage": 15.76, "elapsed_time": "0:10:36", "remaining_time": "0:56:41"}
+{"current_steps": 33, "total_steps": 203, "loss": 0.4429, "lr": 4.786031557539532e-06, "epoch": 0.16206261510128914, "percentage": 16.26, "elapsed_time": "0:10:55", "remaining_time": "0:56:16"}
+{"current_steps": 34, "total_steps": 203, "loss": 0.3938, "lr": 4.7695183928174804e-06, "epoch": 0.16697360343769183, "percentage": 16.75, "elapsed_time": "0:11:15", "remaining_time": "0:55:57"}
+{"current_steps": 35, "total_steps": 203, "loss": 0.4069, "lr": 4.752422169756048e-06, "epoch": 0.17188459177409454, "percentage": 17.24, "elapsed_time": "0:11:34", "remaining_time": "0:55:33"}
+{"current_steps": 36, "total_steps": 203, "loss": 0.4297, "lr": 4.734747280518549e-06, "epoch": 0.17679558011049723, "percentage": 17.73, "elapsed_time": "0:11:54", "remaining_time": "0:55:14"}
+{"current_steps": 37, "total_steps": 203, "loss": 0.3899, "lr": 4.716498265932501e-06, "epoch": 0.18170656844689995, "percentage": 18.23, "elapsed_time": "0:12:17", "remaining_time": "0:55:09"}
+{"current_steps": 38, "total_steps": 203, "loss": 0.4362, "lr": 4.697679814323044e-06, "epoch": 0.18661755678330263, "percentage": 18.72, "elapsed_time": "0:12:36", "remaining_time": "0:54:45"}
+{"current_steps": 39, "total_steps": 203, "loss": 0.4164, "lr": 4.678296760308474e-06, "epoch": 0.19152854511970535, "percentage": 19.21, "elapsed_time": "0:12:56", "remaining_time": "0:54:24"}
+{"current_steps": 40, "total_steps": 203, "loss": 0.3879, "lr": 4.6583540835581885e-06, "epoch": 0.19643953345610804, "percentage": 19.7, "elapsed_time": "0:13:16", "remaining_time": "0:54:06"}
+{"current_steps": 41, "total_steps": 203, "loss": 0.4174, "lr": 4.637856907513366e-06, "epoch": 0.20135052179251076, "percentage": 20.2, "elapsed_time": "0:13:39", "remaining_time": "0:53:57"}
+{"current_steps": 42, "total_steps": 203, "loss": 0.4419, "lr": 4.6168104980707105e-06, "epoch": 0.20626151012891344, "percentage": 20.69, "elapsed_time": "0:14:00", "remaining_time": "0:53:40"}
+{"current_steps": 43, "total_steps": 203, "loss": 0.3967, "lr": 4.5952202622296015e-06, "epoch": 0.21117249846531613, "percentage": 21.18, "elapsed_time": "0:14:20", "remaining_time": "0:53:22"}
+{"current_steps": 44, "total_steps": 203, "loss": 0.3988, "lr": 4.573091746702988e-06, "epoch": 0.21608348680171885, "percentage": 21.67, "elapsed_time": "0:14:40", "remaining_time": "0:53:03"}
+{"current_steps": 45, "total_steps": 203, "loss": 0.3921, "lr": 4.55043063649239e-06, "epoch": 0.22099447513812154, "percentage": 22.17, "elapsed_time": "0:14:59", "remaining_time": "0:52:38"}
+{"current_steps": 46, "total_steps": 203, "loss": 0.4017, "lr": 4.527242753427378e-06, "epoch": 0.22590546347452425, "percentage": 22.66, "elapsed_time": "0:15:20", "remaining_time": "0:52:22"}
+{"current_steps": 47, "total_steps": 203, "loss": 0.3999, "lr": 4.5035340546698915e-06, "epoch": 0.23081645181092694, "percentage": 23.15, "elapsed_time": "0:15:42", "remaining_time": "0:52:09"}
+{"current_steps": 48, "total_steps": 203, "loss": 0.3997, "lr": 4.4793106311838e-06, "epoch": 0.23572744014732966, "percentage": 23.65, "elapsed_time": "0:16:01", "remaining_time": "0:51:45"}
+{"current_steps": 49, "total_steps": 203, "loss": 0.4198, "lr": 4.454578706170075e-06, "epoch": 0.24063842848373235, "percentage": 24.14, "elapsed_time": "0:16:20", "remaining_time": "0:51:23"}
+{"current_steps": 50, "total_steps": 203, "loss": 0.4077, "lr": 4.429344633468005e-06, "epoch": 0.24554941682013506, "percentage": 24.63, "elapsed_time": "0:16:42", "remaining_time": "0:51:07"}
+{"current_steps": 51, "total_steps": 203, "loss": 0.4201, "lr": 4.4036148959228365e-06, "epoch": 0.2504604051565378, "percentage": 25.12, "elapsed_time": "0:17:00", "remaining_time": "0:50:40"}
+{"current_steps": 52, "total_steps": 203, "loss": 0.4245, "lr": 4.3773961037202784e-06, "epoch": 0.25537139349294047, "percentage": 25.62, "elapsed_time": "0:17:19", "remaining_time": "0:50:17"}
+{"current_steps": 53, "total_steps": 203, "loss": 0.3993, "lr": 4.350694992688289e-06, "epoch": 0.26028238182934316, "percentage": 26.11, "elapsed_time": "0:17:39", "remaining_time": "0:49:59"}
+{"current_steps": 54, "total_steps": 203, "loss": 0.3734, "lr": 4.323518422566586e-06, "epoch": 0.26519337016574585, "percentage": 26.6, "elapsed_time": "0:18:01", "remaining_time": "0:49:43"}
+{"current_steps": 55, "total_steps": 203, "loss": 0.4114, "lr": 4.295873375244319e-06, "epoch": 0.27010435850214853, "percentage": 27.09, "elapsed_time": "0:18:20", "remaining_time": "0:49:20"}
+{"current_steps": 56, "total_steps": 203, "loss": 0.4113, "lr": 4.267766952966369e-06, "epoch": 0.2750153468385513, "percentage": 27.59, "elapsed_time": "0:18:40", "remaining_time": "0:49:00"}
+{"current_steps": 57, "total_steps": 203, "loss": 0.3883, "lr": 4.239206376508716e-06, "epoch": 0.27992633517495397, "percentage": 28.08, "elapsed_time": "0:18:58", "remaining_time": "0:48:36"}
+{"current_steps": 58, "total_steps": 203, "loss": 0.3747, "lr": 4.210198983323366e-06, "epoch": 0.28483732351135665, "percentage": 28.57, "elapsed_time": "0:19:17", "remaining_time": "0:48:14"}
+{"current_steps": 59, "total_steps": 203, "loss": 0.4376, "lr": 4.1807522256532925e-06, "epoch": 0.28974831184775934, "percentage": 29.06, "elapsed_time": "0:19:35", "remaining_time": "0:47:48"}
+{"current_steps": 60, "total_steps": 203, "loss": 0.4346, "lr": 4.150873668617899e-06, "epoch": 0.2946593001841621, "percentage": 29.56, "elapsed_time": "0:19:53", "remaining_time": "0:47:25"}
+{"current_steps": 61, "total_steps": 203, "loss": 0.3967, "lr": 4.120570988269472e-06, "epoch": 0.2995702885205648, "percentage": 30.05, "elapsed_time": "0:20:11", "remaining_time": "0:46:59"}
+{"current_steps": 62, "total_steps": 203, "loss": 0.3882, "lr": 4.089851969621138e-06, "epoch": 0.30448127685696746, "percentage": 30.54, "elapsed_time": "0:20:30", "remaining_time": "0:46:38"}
+{"current_steps": 63, "total_steps": 203, "loss": 0.4036, "lr": 4.058724504646834e-06, "epoch": 0.30939226519337015, "percentage": 31.03, "elapsed_time": "0:20:51", "remaining_time": "0:46:21"}
+{"current_steps": 64, "total_steps": 203, "loss": 0.4039, "lr": 4.027196590253786e-06, "epoch": 0.31430325352977284, "percentage": 31.53, "elapsed_time": "0:21:13", "remaining_time": "0:46:06"}
+{"current_steps": 65, "total_steps": 203, "loss": 0.4365, "lr": 3.99527632622804e-06, "epoch": 0.3192142418661756, "percentage": 32.02, "elapsed_time": "0:21:32", "remaining_time": "0:45:43"}
+{"current_steps": 66, "total_steps": 203, "loss": 0.4037, "lr": 3.9629719131535595e-06, "epoch": 0.3241252302025783, "percentage": 32.51, "elapsed_time": "0:21:51", "remaining_time": "0:45:21"}
+{"current_steps": 67, "total_steps": 203, "loss": 0.4069, "lr": 3.930291650305424e-06, "epoch": 0.32903621853898096, "percentage": 33.0, "elapsed_time": "0:22:12", "remaining_time": "0:45:03"}
+{"current_steps": 68, "total_steps": 203, "loss": 0.4031, "lr": 3.897243933517679e-06, "epoch": 0.33394720687538365, "percentage": 33.5, "elapsed_time": "0:22:31", "remaining_time": "0:44:43"}
+{"current_steps": 69, "total_steps": 203, "loss": 0.4066, "lr": 3.863837253026372e-06, "epoch": 0.3388581952117864, "percentage": 33.99, "elapsed_time": "0:22:49", "remaining_time": "0:44:20"}
+{"current_steps": 70, "total_steps": 203, "loss": 0.389, "lr": 3.830080191288342e-06, "epoch": 0.3437691835481891, "percentage": 34.48, "elapsed_time": "0:23:08", "remaining_time": "0:43:57"}
+{"current_steps": 71, "total_steps": 203, "loss": 0.396, "lr": 3.7959814207763134e-06, "epoch": 0.34868017188459177, "percentage": 34.98, "elapsed_time": "0:23:29", "remaining_time": "0:43:41"}
+{"current_steps": 72, "total_steps": 203, "loss": 0.3878, "lr": 3.761549701750865e-06, "epoch": 0.35359116022099446, "percentage": 35.47, "elapsed_time": "0:23:47", "remaining_time": "0:43:16"}
+{"current_steps": 73, "total_steps": 203, "loss": 0.4148, "lr": 3.7267938800098454e-06, "epoch": 0.35850214855739715, "percentage": 35.96, "elapsed_time": "0:24:07", "remaining_time": "0:42:56"}
+{"current_steps": 74, "total_steps": 203, "loss": 0.4048, "lr": 3.691722884615814e-06, "epoch": 0.3634131368937999, "percentage": 36.45, "elapsed_time": "0:24:26", "remaining_time": "0:42:35"}
+{"current_steps": 75, "total_steps": 203, "loss": 0.3923, "lr": 3.656345725602089e-06, "epoch": 0.3683241252302026, "percentage": 36.95, "elapsed_time": "0:24:44", "remaining_time": "0:42:13"}
+{"current_steps": 76, "total_steps": 203, "loss": 0.4059, "lr": 3.6206714916579925e-06, "epoch": 0.37323511356660527, "percentage": 37.44, "elapsed_time": "0:25:01", "remaining_time": "0:41:49"}
+{"current_steps": 77, "total_steps": 203, "loss": 0.4082, "lr": 3.5847093477938955e-06, "epoch": 0.37814610190300796, "percentage": 37.93, "elapsed_time": "0:25:20", "remaining_time": "0:41:27"}
+{"current_steps": 78, "total_steps": 203, "loss": 0.4119, "lr": 3.5484685329866424e-06, "epoch": 0.3830570902394107, "percentage": 38.42, "elapsed_time": "0:25:39", "remaining_time": "0:41:06"}
+{"current_steps": 79, "total_steps": 203, "loss": 0.4001, "lr": 3.5119583578059845e-06, "epoch": 0.3879680785758134, "percentage": 38.92, "elapsed_time": "0:25:57", "remaining_time": "0:40:44"}
+{"current_steps": 80, "total_steps": 203, "loss": 0.4347, "lr": 3.4751882020226174e-06, "epoch": 0.3928790669122161, "percentage": 39.41, "elapsed_time": "0:26:13", "remaining_time": "0:40:19"}
+{"current_steps": 81, "total_steps": 203, "loss": 0.3889, "lr": 3.438167512198436e-06, "epoch": 0.39779005524861877, "percentage": 39.9, "elapsed_time": "0:26:31", "remaining_time": "0:39:57"}
+{"current_steps": 82, "total_steps": 203, "loss": 0.3909, "lr": 3.400905799259634e-06, "epoch": 0.4027010435850215, "percentage": 40.39, "elapsed_time": "0:26:51", "remaining_time": "0:39:38"}
+{"current_steps": 83, "total_steps": 203, "loss": 0.4161, "lr": 3.3634126360532694e-06, "epoch": 0.4076120319214242, "percentage": 40.89, "elapsed_time": "0:27:11", "remaining_time": "0:39:18"}
+{"current_steps": 84, "total_steps": 203, "loss": 0.3894, "lr": 3.3256976548879183e-06, "epoch": 0.4125230202578269, "percentage": 41.38, "elapsed_time": "0:27:30", "remaining_time": "0:38:58"}
+{"current_steps": 85, "total_steps": 203, "loss": 0.4025, "lr": 3.2877705450590525e-06, "epoch": 0.4174340085942296, "percentage": 41.87, "elapsed_time": "0:27:48", "remaining_time": "0:38:35"}
+{"current_steps": 86, "total_steps": 203, "loss": 0.3824, "lr": 3.249641050359779e-06, "epoch": 0.42234499693063227, "percentage": 42.36, "elapsed_time": "0:28:07", "remaining_time": "0:38:16"}
+{"current_steps": 87, "total_steps": 203, "loss": 0.4145, "lr": 3.2113189665775812e-06, "epoch": 0.427255985267035, "percentage": 42.86, "elapsed_time": "0:28:24", "remaining_time": "0:37:53"}
+{"current_steps": 88, "total_steps": 203, "loss": 0.4059, "lr": 3.172814138977692e-06, "epoch": 0.4321669736034377, "percentage": 43.35, "elapsed_time": "0:28:42", "remaining_time": "0:37:31"}
+{"current_steps": 89, "total_steps": 203, "loss": 0.4219, "lr": 3.1341364597737684e-06, "epoch": 0.4370779619398404, "percentage": 43.84, "elapsed_time": "0:29:00", "remaining_time": "0:37:09"}
+{"current_steps": 90, "total_steps": 203, "loss": 0.4042, "lr": 3.0952958655864957e-06, "epoch": 0.4419889502762431, "percentage": 44.33, "elapsed_time": "0:29:23", "remaining_time": "0:36:54"}
+{"current_steps": 91, "total_steps": 203, "loss": 0.3886, "lr": 3.056302334890786e-06, "epoch": 0.4468999386126458, "percentage": 44.83, "elapsed_time": "0:29:42", "remaining_time": "0:36:34"}
+{"current_steps": 92, "total_steps": 203, "loss": 0.3966, "lr": 3.0171658854522274e-06, "epoch": 0.4518109269490485, "percentage": 45.32, "elapsed_time": "0:30:04", "remaining_time": "0:36:17"}
+{"current_steps": 93, "total_steps": 203, "loss": 0.3676, "lr": 2.9778965717534314e-06, "epoch": 0.4567219152854512, "percentage": 45.81, "elapsed_time": "0:30:25", "remaining_time": "0:35:59"}
+{"current_steps": 94, "total_steps": 203, "loss": 0.3805, "lr": 2.9385044824109544e-06, "epoch": 0.4616329036218539, "percentage": 46.31, "elapsed_time": "0:30:44", "remaining_time": "0:35:38"}
+{"current_steps": 95, "total_steps": 203, "loss": 0.4195, "lr": 2.8989997375834485e-06, "epoch": 0.4665438919582566, "percentage": 46.8, "elapsed_time": "0:31:02", "remaining_time": "0:35:17"}
+{"current_steps": 96, "total_steps": 203, "loss": 0.3787, "lr": 2.859392486371705e-06, "epoch": 0.4714548802946593, "percentage": 47.29, "elapsed_time": "0:31:20", "remaining_time": "0:34:56"}
+{"current_steps": 97, "total_steps": 203, "loss": 0.3883, "lr": 2.8196929042112652e-06, "epoch": 0.476365868631062, "percentage": 47.78, "elapsed_time": "0:31:39", "remaining_time": "0:34:35"}
+{"current_steps": 98, "total_steps": 203, "loss": 0.362, "lr": 2.7799111902582697e-06, "epoch": 0.4812768569674647, "percentage": 48.28, "elapsed_time": "0:31:57", "remaining_time": "0:34:14"}
+{"current_steps": 99, "total_steps": 203, "loss": 0.3904, "lr": 2.7400575647692046e-06, "epoch": 0.4861878453038674, "percentage": 48.77, "elapsed_time": "0:32:16", "remaining_time": "0:33:54"}
+{"current_steps": 100, "total_steps": 203, "loss": 0.3682, "lr": 2.7001422664752338e-06, "epoch": 0.4910988336402701, "percentage": 49.26, "elapsed_time": "0:32:36", "remaining_time": "0:33:35"}
+{"current_steps": 101, "total_steps": 203, "loss": 0.3912, "lr": 2.6601755499517826e-06, "epoch": 0.4960098219766728, "percentage": 49.75, "elapsed_time": "0:32:56", "remaining_time": "0:33:16"}
+{"current_steps": 102, "total_steps": 203, "loss": 0.3942, "lr": 2.620167682984052e-06, "epoch": 0.5009208103130756, "percentage": 50.25, "elapsed_time": "0:33:17", "remaining_time": "0:32:57"}
+{"current_steps": 103, "total_steps": 203, "loss": 0.3893, "lr": 2.580128943929139e-06, "epoch": 0.5058317986494782, "percentage": 50.74, "elapsed_time": "0:33:41", "remaining_time": "0:32:42"}
+{"current_steps": 104, "total_steps": 203, "loss": 0.4, "lr": 2.5400696190754347e-06, "epoch": 0.5107427869858809, "percentage": 51.23, "elapsed_time": "0:33:59", "remaining_time": "0:32:21"}
+{"current_steps": 105, "total_steps": 203, "loss": 0.3997, "lr": 2.5e-06, "epoch": 0.5156537753222836, "percentage": 51.72, "elapsed_time": "0:34:18", "remaining_time": "0:32:01"}
+{"current_steps": 106, "total_steps": 203, "loss": 0.3896, "lr": 2.459930380924566e-06, "epoch": 0.5205647636586863, "percentage": 52.22, "elapsed_time": "0:34:36", "remaining_time": "0:31:40"}
+{"current_steps": 107, "total_steps": 203, "loss": 0.3933, "lr": 2.4198710560708623e-06, "epoch": 0.525475751995089, "percentage": 52.71, "elapsed_time": "0:34:55", "remaining_time": "0:31:20"}
+{"current_steps": 108, "total_steps": 203, "loss": 0.398, "lr": 2.3798323170159487e-06, "epoch": 0.5303867403314917, "percentage": 53.2, "elapsed_time": "0:35:13", "remaining_time": "0:30:59"}
+{"current_steps": 109, "total_steps": 203, "loss": 0.386, "lr": 2.339824450048218e-06, "epoch": 0.5352977286678944, "percentage": 53.69, "elapsed_time": "0:35:31", "remaining_time": "0:30:38"}
+{"current_steps": 110, "total_steps": 203, "loss": 0.3705, "lr": 2.299857733524767e-06, "epoch": 0.5402087170042971, "percentage": 54.19, "elapsed_time": "0:35:49", "remaining_time": "0:30:17"}
+{"current_steps": 111, "total_steps": 203, "loss": 0.3953, "lr": 2.2599424352307958e-06, "epoch": 0.5451197053406999, "percentage": 54.68, "elapsed_time": "0:36:06", "remaining_time": "0:29:56"}
+{"current_steps": 112, "total_steps": 203, "loss": 0.379, "lr": 2.2200888097417308e-06, "epoch": 0.5500306936771026, "percentage": 55.17, "elapsed_time": "0:36:27", "remaining_time": "0:29:37"}
+{"current_steps": 113, "total_steps": 203, "loss": 0.4005, "lr": 2.1803070957887348e-06, "epoch": 0.5549416820135052, "percentage": 55.67, "elapsed_time": "0:36:44", "remaining_time": "0:29:15"}
+{"current_steps": 114, "total_steps": 203, "loss": 0.3904, "lr": 2.140607513628296e-06, "epoch": 0.5598526703499079, "percentage": 56.16, "elapsed_time": "0:37:08", "remaining_time": "0:29:00"}
+{"current_steps": 115, "total_steps": 203, "loss": 0.3659, "lr": 2.1010002624165528e-06, "epoch": 0.5647636586863106, "percentage": 56.65, "elapsed_time": "0:37:28", "remaining_time": "0:28:40"}
+{"current_steps": 116, "total_steps": 203, "loss": 0.4218, "lr": 2.0614955175890464e-06, "epoch": 0.5696746470227133, "percentage": 57.14, "elapsed_time": "0:37:47", "remaining_time": "0:28:20"}
+{"current_steps": 117, "total_steps": 203, "loss": 0.4194, "lr": 2.02210342824657e-06, "epoch": 0.574585635359116, "percentage": 57.64, "elapsed_time": "0:38:05", "remaining_time": "0:27:59"}
+{"current_steps": 118, "total_steps": 203, "loss": 0.4065, "lr": 1.982834114547773e-06, "epoch": 0.5794966236955187, "percentage": 58.13, "elapsed_time": "0:38:23", "remaining_time": "0:27:39"}
+{"current_steps": 119, "total_steps": 203, "loss": 0.3913, "lr": 1.9436976651092143e-06, "epoch": 0.5844076120319214, "percentage": 58.62, "elapsed_time": "0:38:41", "remaining_time": "0:27:18"}
+{"current_steps": 120, "total_steps": 203, "loss": 0.4051, "lr": 1.9047041344135045e-06, "epoch": 0.5893186003683242, "percentage": 59.11, "elapsed_time": "0:38:58", "remaining_time": "0:26:57"}
+{"current_steps": 121, "total_steps": 203, "loss": 0.3745, "lr": 1.865863540226232e-06, "epoch": 0.5942295887047269, "percentage": 59.61, "elapsed_time": "0:39:17", "remaining_time": "0:26:37"}
+{"current_steps": 122, "total_steps": 203, "loss": 0.4129, "lr": 1.827185861022308e-06, "epoch": 0.5991405770411296, "percentage": 60.1, "elapsed_time": "0:39:35", "remaining_time": "0:26:16"}
+{"current_steps": 123, "total_steps": 203, "loss": 0.35, "lr": 1.7886810334224192e-06, "epoch": 0.6040515653775322, "percentage": 60.59, "elapsed_time": "0:39:53", "remaining_time": "0:25:56"}
+{"current_steps": 124, "total_steps": 203, "loss": 0.4009, "lr": 1.750358949640221e-06, "epoch": 0.6089625537139349, "percentage": 61.08, "elapsed_time": "0:40:11", "remaining_time": "0:25:36"}
+{"current_steps": 125, "total_steps": 203, "loss": 0.4074, "lr": 1.7122294549409486e-06, "epoch": 0.6138735420503376, "percentage": 61.58, "elapsed_time": "0:40:31", "remaining_time": "0:25:17"}
+{"current_steps": 126, "total_steps": 203, "loss": 0.4062, "lr": 1.6743023451120831e-06, "epoch": 0.6187845303867403, "percentage": 62.07, "elapsed_time": "0:40:54", "remaining_time": "0:24:59"}
+{"current_steps": 127, "total_steps": 203, "loss": 0.387, "lr": 1.6365873639467314e-06, "epoch": 0.623695518723143, "percentage": 62.56, "elapsed_time": "0:41:16", "remaining_time": "0:24:41"}
+{"current_steps": 128, "total_steps": 203, "loss": 0.3793, "lr": 1.599094200740367e-06, "epoch": 0.6286065070595457, "percentage": 63.05, "elapsed_time": "0:41:37", "remaining_time": "0:24:23"}
+{"current_steps": 129, "total_steps": 203, "loss": 0.4056, "lr": 1.561832487801565e-06, "epoch": 0.6335174953959485, "percentage": 63.55, "elapsed_time": "0:41:56", "remaining_time": "0:24:03"}
+{"current_steps": 130, "total_steps": 203, "loss": 0.4064, "lr": 1.524811797977383e-06, "epoch": 0.6384284837323512, "percentage": 64.04, "elapsed_time": "0:42:13", "remaining_time": "0:23:42"}
+{"current_steps": 131, "total_steps": 203, "loss": 0.4017, "lr": 1.4880416421940155e-06, "epoch": 0.6433394720687539, "percentage": 64.53, "elapsed_time": "0:42:31", "remaining_time": "0:23:22"}
+{"current_steps": 132, "total_steps": 203, "loss": 0.4165, "lr": 1.4515314670133582e-06, "epoch": 0.6482504604051565, "percentage": 65.02, "elapsed_time": "0:42:50", "remaining_time": "0:23:02"}
+{"current_steps": 133, "total_steps": 203, "loss": 0.3883, "lr": 1.415290652206105e-06, "epoch": 0.6531614487415592, "percentage": 65.52, "elapsed_time": "0:43:09", "remaining_time": "0:22:42"}
+{"current_steps": 134, "total_steps": 203, "loss": 0.3851, "lr": 1.3793285083420077e-06, "epoch": 0.6580724370779619, "percentage": 66.01, "elapsed_time": "0:43:30", "remaining_time": "0:22:24"}
+{"current_steps": 135, "total_steps": 203, "loss": 0.4047, "lr": 1.3436542743979125e-06, "epoch": 0.6629834254143646, "percentage": 66.5, "elapsed_time": "0:43:50", "remaining_time": "0:22:04"}
+{"current_steps": 136, "total_steps": 203, "loss": 0.3868, "lr": 1.3082771153841872e-06, "epoch": 0.6678944137507673, "percentage": 67.0, "elapsed_time": "0:44:09", "remaining_time": "0:21:45"}
+{"current_steps": 137, "total_steps": 203, "loss": 0.3799, "lr": 1.2732061199901563e-06, "epoch": 0.67280540208717, "percentage": 67.49, "elapsed_time": "0:44:30", "remaining_time": "0:21:26"}
+{"current_steps": 138, "total_steps": 203, "loss": 0.3884, "lr": 1.2384502982491359e-06, "epoch": 0.6777163904235728, "percentage": 67.98, "elapsed_time": "0:44:47", "remaining_time": "0:21:05"}
+{"current_steps": 139, "total_steps": 203, "loss": 0.3931, "lr": 1.2040185792236874e-06, "epoch": 0.6826273787599755, "percentage": 68.47, "elapsed_time": "0:45:05", "remaining_time": "0:20:45"}
+{"current_steps": 140, "total_steps": 203, "loss": 0.3888, "lr": 1.169919808711659e-06, "epoch": 0.6875383670963782, "percentage": 68.97, "elapsed_time": "0:45:25", "remaining_time": "0:20:26"}
+{"current_steps": 141, "total_steps": 203, "loss": 0.4002, "lr": 1.1361627469736286e-06, "epoch": 0.6924493554327809, "percentage": 69.46, "elapsed_time": "0:45:43", "remaining_time": "0:20:06"}
+{"current_steps": 142, "total_steps": 203, "loss": 0.4165, "lr": 1.1027560664823208e-06, "epoch": 0.6973603437691835, "percentage": 69.95, "elapsed_time": "0:46:01", "remaining_time": "0:19:46"}
+{"current_steps": 143, "total_steps": 203, "loss": 0.3974, "lr": 1.0697083496945766e-06, "epoch": 0.7022713321055862, "percentage": 70.44, "elapsed_time": "0:46:21", "remaining_time": "0:19:27"}
+{"current_steps": 144, "total_steps": 203, "loss": 0.3717, "lr": 1.0370280868464405e-06, "epoch": 0.7071823204419889, "percentage": 70.94, "elapsed_time": "0:46:38", "remaining_time": "0:19:06"}
+{"current_steps": 145, "total_steps": 203, "loss": 0.392, "lr": 1.00472367377196e-06, "epoch": 0.7120933087783916, "percentage": 71.43, "elapsed_time": "0:46:55", "remaining_time": "0:18:46"}
+{"current_steps": 146, "total_steps": 203, "loss": 0.3711, "lr": 9.728034097462144e-07, "epoch": 0.7170042971147943, "percentage": 71.92, "elapsed_time": "0:47:16", "remaining_time": "0:18:27"}
+{"current_steps": 147, "total_steps": 203, "loss": 0.3861, "lr": 9.412754953531664e-07, "epoch": 0.7219152854511971, "percentage": 72.41, "elapsed_time": "0:47:33", "remaining_time": "0:18:07"}
+{"current_steps": 148, "total_steps": 203, "loss": 0.3755, "lr": 9.101480303788623e-07, "epoch": 0.7268262737875998, "percentage": 72.91, "elapsed_time": "0:47:58", "remaining_time": "0:17:49"}
+{"current_steps": 149, "total_steps": 203, "loss": 0.3799, "lr": 8.794290117305296e-07, "epoch": 0.7317372621240025, "percentage": 73.4, "elapsed_time": "0:48:17", "remaining_time": "0:17:30"}
+{"current_steps": 150, "total_steps": 203, "loss": 0.3858, "lr": 8.491263313821021e-07, "epoch": 0.7366482504604052, "percentage": 73.89, "elapsed_time": "0:48:36", "remaining_time": "0:17:10"}
+{"current_steps": 151, "total_steps": 203, "loss": 0.3777, "lr": 8.192477743467078e-07, "epoch": 0.7415592387968079, "percentage": 74.38, "elapsed_time": "0:48:56", "remaining_time": "0:16:51"}
+{"current_steps": 152, "total_steps": 203, "loss": 0.4331, "lr": 7.898010166766348e-07, "epoch": 0.7464702271332105, "percentage": 74.88, "elapsed_time": "0:49:16", "remaining_time": "0:16:31"}
+{"current_steps": 153, "total_steps": 203, "loss": 0.3726, "lr": 7.607936234912841e-07, "epoch": 0.7513812154696132, "percentage": 75.37, "elapsed_time": "0:49:40", "remaining_time": "0:16:13"}
+{"current_steps": 154, "total_steps": 203, "loss": 0.383, "lr": 7.322330470336314e-07, "epoch": 0.7562922038060159, "percentage": 75.86, "elapsed_time": "0:50:01", "remaining_time": "0:15:55"}
+{"current_steps": 155, "total_steps": 203, "loss": 0.3809, "lr": 7.041266247556814e-07, "epoch": 0.7612031921424187, "percentage": 76.35, "elapsed_time": "0:50:21", "remaining_time": "0:15:35"}
+{"current_steps": 156, "total_steps": 203, "loss": 0.4129, "lr": 6.764815774334149e-07, "epoch": 0.7661141804788214, "percentage": 76.85, "elapsed_time": "0:50:40", "remaining_time": "0:15:16"}
+{"current_steps": 157, "total_steps": 203, "loss": 0.3691, "lr": 6.493050073117115e-07, "epoch": 0.7710251688152241, "percentage": 77.34, "elapsed_time": "0:51:02", "remaining_time": "0:14:57"}
+{"current_steps": 158, "total_steps": 203, "loss": 0.3751, "lr": 6.226038962797218e-07, "epoch": 0.7759361571516268, "percentage": 77.83, "elapsed_time": "0:51:21", "remaining_time": "0:14:37"}
+{"current_steps": 159, "total_steps": 203, "loss": 0.384, "lr": 5.963851040771639e-07, "epoch": 0.7808471454880295, "percentage": 78.33, "elapsed_time": "0:51:39", "remaining_time": "0:14:17"}
+{"current_steps": 160, "total_steps": 203, "loss": 0.4117, "lr": 5.706553665319955e-07, "epoch": 0.7857581338244322, "percentage": 78.82, "elapsed_time": "0:51:57", "remaining_time": "0:13:57"}
+{"current_steps": 161, "total_steps": 203, "loss": 0.3742, "lr": 5.454212938299256e-07, "epoch": 0.7906691221608348, "percentage": 79.31, "elapsed_time": "0:52:18", "remaining_time": "0:13:38"}
+{"current_steps": 162, "total_steps": 203, "loss": 0.3882, "lr": 5.20689368816201e-07, "epoch": 0.7955801104972375, "percentage": 79.8, "elapsed_time": "0:52:34", "remaining_time": "0:13:18"}
+{"current_steps": 163, "total_steps": 203, "loss": 0.3867, "lr": 4.964659453301088e-07, "epoch": 0.8004910988336402, "percentage": 80.3, "elapsed_time": "0:52:53", "remaining_time": "0:12:58"}
+{"current_steps": 164, "total_steps": 203, "loss": 0.3864, "lr": 4.7275724657262293e-07, "epoch": 0.805402087170043, "percentage": 80.79, "elapsed_time": "0:53:12", "remaining_time": "0:12:39"}
+{"current_steps": 165, "total_steps": 203, "loss": 0.3751, "lr": 4.495693635076101e-07, "epoch": 0.8103130755064457, "percentage": 81.28, "elapsed_time": "0:53:33", "remaining_time": "0:12:20"}
+{"current_steps": 166, "total_steps": 203, "loss": 0.3971, "lr": 4.2690825329701313e-07, "epoch": 0.8152240638428484, "percentage": 81.77, "elapsed_time": "0:53:53", "remaining_time": "0:12:00"}
+{"current_steps": 167, "total_steps": 203, "loss": 0.3625, "lr": 4.0477973777039854e-07, "epoch": 0.8201350521792511, "percentage": 82.27, "elapsed_time": "0:54:17", "remaining_time": "0:11:42"}
+{"current_steps": 168, "total_steps": 203, "loss": 0.4035, "lr": 3.831895019292897e-07, "epoch": 0.8250460405156538, "percentage": 82.76, "elapsed_time": "0:54:36", "remaining_time": "0:11:22"}
+{"current_steps": 169, "total_steps": 203, "loss": 0.39, "lr": 3.621430924866348e-07, "epoch": 0.8299570288520565, "percentage": 83.25, "elapsed_time": "0:54:57", "remaining_time": "0:11:03"}
+{"current_steps": 170, "total_steps": 203, "loss": 0.385, "lr": 3.4164591644181233e-07, "epoch": 0.8348680171884592, "percentage": 83.74, "elapsed_time": "0:55:16", "remaining_time": "0:10:43"}
+{"current_steps": 171, "total_steps": 203, "loss": 0.366, "lr": 3.217032396915265e-07, "epoch": 0.8397790055248618, "percentage": 84.24, "elapsed_time": "0:55:37", "remaining_time": "0:10:24"}
+{"current_steps": 172, "total_steps": 203, "loss": 0.352, "lr": 3.02320185676957e-07, "epoch": 0.8446899938612645, "percentage": 84.73, "elapsed_time": "0:56:01", "remaining_time": "0:10:05"}
+{"current_steps": 173, "total_steps": 203, "loss": 0.3883, "lr": 2.8350173406749975e-07, "epoch": 0.8496009821976673, "percentage": 85.22, "elapsed_time": "0:56:19", "remaining_time": "0:09:46"}
+{"current_steps": 174, "total_steps": 203, "loss": 0.3799, "lr": 2.652527194814511e-07, "epoch": 0.85451197053407, "percentage": 85.71, "elapsed_time": "0:56:40", "remaining_time": "0:09:26"}
+{"current_steps": 175, "total_steps": 203, "loss": 0.3845, "lr": 2.4757783024395244e-07, "epoch": 0.8594229588704727, "percentage": 86.21, "elapsed_time": "0:56:57", "remaining_time": "0:09:06"}
+{"current_steps": 176, "total_steps": 203, "loss": 0.3717, "lr": 2.3048160718252e-07, "epoch": 0.8643339472068754, "percentage": 86.7, "elapsed_time": "0:57:15", "remaining_time": "0:08:47"}
+{"current_steps": 177, "total_steps": 203, "loss": 0.4046, "lr": 2.1396844246046904e-07, "epoch": 0.8692449355432781, "percentage": 87.19, "elapsed_time": "0:57:33", "remaining_time": "0:08:27"}
+{"current_steps": 178, "total_steps": 203, "loss": 0.3988, "lr": 1.980425784485293e-07, "epoch": 0.8741559238796808, "percentage": 87.68, "elapsed_time": "0:57:50", "remaining_time": "0:08:07"}
+{"current_steps": 179, "total_steps": 203, "loss": 0.373, "lr": 1.8270810663494591e-07, "epoch": 0.8790669122160835, "percentage": 88.18, "elapsed_time": "0:58:13", "remaining_time": "0:07:48"}
+{"current_steps": 180, "total_steps": 203, "loss": 0.3609, "lr": 1.679689665743381e-07, "epoch": 0.8839779005524862, "percentage": 88.67, "elapsed_time": "0:58:35", "remaining_time": "0:07:29"}
+{"current_steps": 181, "total_steps": 203, "loss": 0.4122, "lr": 1.538289448755989e-07, "epoch": 0.8888888888888888, "percentage": 89.16, "elapsed_time": "0:58:53", "remaining_time": "0:07:09"}
+{"current_steps": 182, "total_steps": 203, "loss": 0.3752, "lr": 1.4029167422908107e-07, "epoch": 0.8937998772252916, "percentage": 89.66, "elapsed_time": "0:59:13", "remaining_time": "0:06:49"}
+{"current_steps": 183, "total_steps": 203, "loss": 0.3935, "lr": 1.273606324733284e-07, "epoch": 0.8987108655616943, "percentage": 90.15, "elapsed_time": "0:59:31", "remaining_time": "0:06:30"}
+{"current_steps": 184, "total_steps": 203, "loss": 0.3885, "lr": 1.1503914170159058e-07, "epoch": 0.903621853898097, "percentage": 90.64, "elapsed_time": "0:59:52", "remaining_time": "0:06:10"}
+{"current_steps": 185, "total_steps": 203, "loss": 0.3889, "lr": 1.0333036740834857e-07, "epoch": 0.9085328422344997, "percentage": 91.13, "elapsed_time": "1:00:11", "remaining_time": "0:05:51"}
+{"current_steps": 186, "total_steps": 203, "loss": 0.4182, "lr": 9.223731767607436e-08, "epoch": 0.9134438305709024, "percentage": 91.63, "elapsed_time": "1:00:28", "remaining_time": "0:05:31"}
+{"current_steps": 187, "total_steps": 203, "loss": 0.4103, "lr": 8.176284240242638e-08, "epoch": 0.9183548189073051, "percentage": 92.12, "elapsed_time": "1:00:45", "remaining_time": "0:05:11"}
+{"current_steps": 188, "total_steps": 203, "loss": 0.3801, "lr": 7.190963256809069e-08, "epoch": 0.9232658072437078, "percentage": 92.61, "elapsed_time": "1:01:05", "remaining_time": "0:04:52"}
+{"current_steps": 189, "total_steps": 203, "loss": 0.4097, "lr": 6.268021954544095e-08, "epoch": 0.9281767955801105, "percentage": 93.1, "elapsed_time": "1:01:22", "remaining_time": "0:04:32"}
+{"current_steps": 190, "total_steps": 203, "loss": 0.392, "lr": 5.407697444821169e-08, "epoch": 0.9330877839165131, "percentage": 93.6, "elapsed_time": "1:01:43", "remaining_time": "0:04:13"}
+{"current_steps": 191, "total_steps": 203, "loss": 0.4033, "lr": 4.61021075223364e-08, "epoch": 0.937998772252916, "percentage": 94.09, "elapsed_time": "1:02:01", "remaining_time": "0:03:53"}
+{"current_steps": 192, "total_steps": 203, "loss": 0.3782, "lr": 3.8757667578119e-08, "epoch": 0.9429097605893186, "percentage": 94.58, "elapsed_time": "1:02:21", "remaining_time": "0:03:34"}
+{"current_steps": 193, "total_steps": 203, "loss": 0.3531, "lr": 3.204554146387456e-08, "epoch": 0.9478207489257213, "percentage": 95.07, "elapsed_time": "1:02:46", "remaining_time": "0:03:15"}
+{"current_steps": 194, "total_steps": 203, "loss": 0.3826, "lr": 2.5967453581185187e-08, "epoch": 0.952731737262124, "percentage": 95.57, "elapsed_time": "1:03:06", "remaining_time": "0:02:55"}
+{"current_steps": 195, "total_steps": 203, "loss": 0.3778, "lr": 2.052496544188487e-08, "epoch": 0.9576427255985267, "percentage": 96.06, "elapsed_time": "1:03:24", "remaining_time": "0:02:36"}
+{"current_steps": 196, "total_steps": 203, "loss": 0.3513, "lr": 1.571947526689349e-08, "epoch": 0.9625537139349294, "percentage": 96.55, "elapsed_time": "1:03:46", "remaining_time": "0:02:16"}
+{"current_steps": 197, "total_steps": 203, "loss": 0.3937, "lr": 1.1552217627004426e-08, "epoch": 0.9674647022713321, "percentage": 97.04, "elapsed_time": "1:04:05", "remaining_time": "0:01:57"}
+{"current_steps": 198, "total_steps": 203, "loss": 0.3788, "lr": 8.024263125710751e-09, "epoch": 0.9723756906077348, "percentage": 97.54, "elapsed_time": "1:04:29", "remaining_time": "0:01:37"}
+{"current_steps": 199, "total_steps": 203, "loss": 0.4029, "lr": 5.136518124159162e-09, "epoch": 0.9772866789441375, "percentage": 98.03, "elapsed_time": "1:04:48", "remaining_time": "0:01:18"}
+{"current_steps": 200, "total_steps": 203, "loss": 0.3644, "lr": 2.8897245082978863e-09, "epoch": 0.9821976672805403, "percentage": 98.52, "elapsed_time": "1:05:09", "remaining_time": "0:00:58"}
+{"current_steps": 201, "total_steps": 203, "loss": 0.4098, "lr": 1.284459498280266e-09, "epoch": 0.9871086556169429, "percentage": 99.01, "elapsed_time": "1:05:28", "remaining_time": "0:00:39"}
+{"current_steps": 202, "total_steps": 203, "loss": 0.3883, "lr": 3.2113550017198734e-10, "epoch": 0.9920196439533456, "percentage": 99.51, "elapsed_time": "1:05:47", "remaining_time": "0:00:19"}
+{"current_steps": 203, "total_steps": 203, "loss": 0.396, "lr": 0.0, "epoch": 0.9969306322897483, "percentage": 100.0, "elapsed_time": "1:06:05", "remaining_time": "0:00:00"}
+{"current_steps": 203, "total_steps": 203, "epoch": 0.9969306322897483, "percentage": 100.0, "elapsed_time": "1:08:29", "remaining_time": "0:00:00"}

trainer_state.json

@@ -0,0 +1,1463 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9969306322897483,
+  "eval_steps": 500,
+  "global_step": 203,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.004910988336402701,
+      "grad_norm": 1.9743820428848267,
+      "learning_rate": 7.142857142857143e-07,
+      "loss": 0.5458,
+      "step": 1
+    },
+    {
+      "epoch": 0.009821976672805401,
+      "grad_norm": 1.6806402206420898,
+      "learning_rate": 1.4285714285714286e-06,
+      "loss": 0.4966,
+      "step": 2
+    },
+    {
+      "epoch": 0.014732965009208104,
+      "grad_norm": 1.864640474319458,
+      "learning_rate": 2.1428571428571427e-06,
+      "loss": 0.5349,
+      "step": 3
+    },
+    {
+      "epoch": 0.019643953345610803,
+      "grad_norm": 1.8146294355392456,
+      "learning_rate": 2.8571428571428573e-06,
+      "loss": 0.5184,
+      "step": 4
+    },
+    {
+      "epoch": 0.024554941682013505,
+      "grad_norm": 1.458012342453003,
+      "learning_rate": 3.5714285714285718e-06,
+      "loss": 0.5031,
+      "step": 5
+    },
+    {
+      "epoch": 0.029465930018416207,
+      "grad_norm": 1.426376223564148,
+      "learning_rate": 4.2857142857142855e-06,
+      "loss": 0.5246,
+      "step": 6
+    },
+    {
+      "epoch": 0.03437691835481891,
+      "grad_norm": 1.0303875207901,
+      "learning_rate": 5e-06,
+      "loss": 0.5022,
+      "step": 7
+    },
+    {
+      "epoch": 0.039287906691221605,
+      "grad_norm": 1.1609914302825928,
+      "learning_rate": 4.999678864499828e-06,
+      "loss": 0.4882,
+      "step": 8
+    },
+    {
+      "epoch": 0.04419889502762431,
+      "grad_norm": 1.8752422332763672,
+      "learning_rate": 4.99871554050172e-06,
+      "loss": 0.4695,
+      "step": 9
+    },
+    {
+      "epoch": 0.04910988336402701,
+      "grad_norm": 1.9186817407608032,
+      "learning_rate": 4.997110275491702e-06,
+      "loss": 0.4773,
+      "step": 10
+    },
+    {
+      "epoch": 0.05402087170042971,
+      "grad_norm": 1.7251890897750854,
+      "learning_rate": 4.994863481875842e-06,
+      "loss": 0.469,
+      "step": 11
+    },
+    {
+      "epoch": 0.058931860036832415,
+      "grad_norm": 1.2969812154769897,
+      "learning_rate": 4.9919757368742895e-06,
+      "loss": 0.4836,
+      "step": 12
+    },
+    {
+      "epoch": 0.06384284837323512,
+      "grad_norm": 0.9343390464782715,
+      "learning_rate": 4.988447782372996e-06,
+      "loss": 0.4308,
+      "step": 13
+    },
+    {
+      "epoch": 0.06875383670963782,
+      "grad_norm": 0.8766093850135803,
+      "learning_rate": 4.984280524733107e-06,
+      "loss": 0.4609,
+      "step": 14
+    },
+    {
+      "epoch": 0.07366482504604052,
+      "grad_norm": 0.7210026979446411,
+      "learning_rate": 4.979475034558115e-06,
+      "loss": 0.4194,
+      "step": 15
+    },
+    {
+      "epoch": 0.07857581338244321,
+      "grad_norm": 0.774637758731842,
+      "learning_rate": 4.974032546418816e-06,
+      "loss": 0.4366,
+      "step": 16
+    },
+    {
+      "epoch": 0.08348680171884591,
+      "grad_norm": 0.8075196146965027,
+      "learning_rate": 4.967954458536126e-06,
+      "loss": 0.444,
+      "step": 17
+    },
+    {
+      "epoch": 0.08839779005524862,
+      "grad_norm": 0.8209702968597412,
+      "learning_rate": 4.9612423324218816e-06,
+      "loss": 0.4538,
+      "step": 18
+    },
+    {
+      "epoch": 0.09330877839165132,
+      "grad_norm": 0.7099929451942444,
+      "learning_rate": 4.953897892477664e-06,
+      "loss": 0.4498,
+      "step": 19
+    },
+    {
+      "epoch": 0.09821976672805402,
+      "grad_norm": 0.6593644618988037,
+      "learning_rate": 4.945923025551789e-06,
+      "loss": 0.4537,
+      "step": 20
+    },
+    {
+      "epoch": 0.10313075506445672,
+      "grad_norm": 0.6309540271759033,
+      "learning_rate": 4.937319780454559e-06,
+      "loss": 0.4524,
+      "step": 21
+    },
+    {
+      "epoch": 0.10804174340085942,
+      "grad_norm": 0.6001752614974976,
+      "learning_rate": 4.92809036743191e-06,
+      "loss": 0.4174,
+      "step": 22
+    },
+    {
+      "epoch": 0.11295273173726213,
+      "grad_norm": 0.6278722286224365,
+      "learning_rate": 4.918237157597574e-06,
+      "loss": 0.4343,
+      "step": 23
+    },
+    {
+      "epoch": 0.11786372007366483,
+      "grad_norm": 0.5969433188438416,
+      "learning_rate": 4.907762682323926e-06,
+      "loss": 0.4104,
+      "step": 24
+    },
+    {
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.7083978056907654,
+      "learning_rate": 4.896669632591652e-06,
+      "loss": 0.433,
+      "step": 25
+    },
+    {
+      "epoch": 0.12768569674647023,
+      "grad_norm": 0.6708449125289917,
+      "learning_rate": 4.88496085829841e-06,
+      "loss": 0.4344,
+      "step": 26
+    },
+    {
+      "epoch": 0.13259668508287292,
+      "grad_norm": 0.6223589181900024,
+      "learning_rate": 4.872639367526672e-06,
+      "loss": 0.4349,
+      "step": 27
+    },
+    {
+      "epoch": 0.13750767341927564,
+      "grad_norm": 0.6295133829116821,
+      "learning_rate": 4.859708325770919e-06,
+      "loss": 0.4284,
+      "step": 28
+    },
+    {
+      "epoch": 0.14241866175567833,
+      "grad_norm": 0.5775576829910278,
+      "learning_rate": 4.846171055124401e-06,
+      "loss": 0.4298,
+      "step": 29
+    },
+    {
+      "epoch": 0.14732965009208104,
+      "grad_norm": 0.5659624934196472,
+      "learning_rate": 4.832031033425663e-06,
+      "loss": 0.4325,
+      "step": 30
+    },
+    {
+      "epoch": 0.15224063842848373,
+      "grad_norm": 0.5679330825805664,
+      "learning_rate": 4.817291893365055e-06,
+      "loss": 0.4317,
+      "step": 31
+    },
+    {
+      "epoch": 0.15715162676488642,
+      "grad_norm": 0.652981698513031,
+      "learning_rate": 4.8019574215514705e-06,
+      "loss": 0.4297,
+      "step": 32
+    },
+    {
+      "epoch": 0.16206261510128914,
+      "grad_norm": 0.6324057579040527,
+      "learning_rate": 4.786031557539532e-06,
+      "loss": 0.4429,
+      "step": 33
+    },
+    {
+      "epoch": 0.16697360343769183,
+      "grad_norm": 0.5611727833747864,
+      "learning_rate": 4.7695183928174804e-06,
+      "loss": 0.3938,
+      "step": 34
+    },
+    {
+      "epoch": 0.17188459177409454,
+      "grad_norm": 0.5474141240119934,
+      "learning_rate": 4.752422169756048e-06,
+      "loss": 0.4069,
+      "step": 35
+    },
+    {
+      "epoch": 0.17679558011049723,
+      "grad_norm": 0.5458659529685974,
+      "learning_rate": 4.734747280518549e-06,
+      "loss": 0.4297,
+      "step": 36
+    },
+    {
+      "epoch": 0.18170656844689995,
+      "grad_norm": 0.5720776319503784,
+      "learning_rate": 4.716498265932501e-06,
+      "loss": 0.3899,
+      "step": 37
+    },
+    {
+      "epoch": 0.18661755678330263,
+      "grad_norm": 0.5874093770980835,
+      "learning_rate": 4.697679814323044e-06,
+      "loss": 0.4362,
+      "step": 38
+    },
+    {
+      "epoch": 0.19152854511970535,
+      "grad_norm": 0.586660623550415,
+      "learning_rate": 4.678296760308474e-06,
+      "loss": 0.4164,
+      "step": 39
+    },
+    {
+      "epoch": 0.19643953345610804,
+      "grad_norm": 0.5384376049041748,
+      "learning_rate": 4.6583540835581885e-06,
+      "loss": 0.3879,
+      "step": 40
+    },
+    {
+      "epoch": 0.20135052179251076,
+      "grad_norm": 0.5448895692825317,
+      "learning_rate": 4.637856907513366e-06,
+      "loss": 0.4174,
+      "step": 41
+    },
+    {
+      "epoch": 0.20626151012891344,
+      "grad_norm": 0.5429976582527161,
+      "learning_rate": 4.6168104980707105e-06,
+      "loss": 0.4419,
+      "step": 42
+    },
+    {
+      "epoch": 0.21117249846531613,
+      "grad_norm": 0.4899347722530365,
+      "learning_rate": 4.5952202622296015e-06,
+      "loss": 0.3967,
+      "step": 43
+    },
+    {
+      "epoch": 0.21608348680171885,
+      "grad_norm": 0.5103813409805298,
+      "learning_rate": 4.573091746702988e-06,
+      "loss": 0.3988,
+      "step": 44
+    },
+    {
+      "epoch": 0.22099447513812154,
+      "grad_norm": 0.5400250554084778,
+      "learning_rate": 4.55043063649239e-06,
+      "loss": 0.3921,
+      "step": 45
+    },
+    {
+      "epoch": 0.22590546347452425,
+      "grad_norm": 0.5391291379928589,
+      "learning_rate": 4.527242753427378e-06,
+      "loss": 0.4017,
+      "step": 46
+    },
+    {
+      "epoch": 0.23081645181092694,
+      "grad_norm": 0.5237475633621216,
+      "learning_rate": 4.5035340546698915e-06,
+      "loss": 0.3999,
+      "step": 47
+    },
+    {
+      "epoch": 0.23572744014732966,
+      "grad_norm": 0.5172109007835388,
+      "learning_rate": 4.4793106311838e-06,
+      "loss": 0.3997,
+      "step": 48
+    },
+    {
+      "epoch": 0.24063842848373235,
+      "grad_norm": 0.5092684626579285,
+      "learning_rate": 4.454578706170075e-06,
+      "loss": 0.4198,
+      "step": 49
+    },
+    {
+      "epoch": 0.24554941682013506,
+      "grad_norm": 0.5523620843887329,
+      "learning_rate": 4.429344633468005e-06,
+      "loss": 0.4077,
+      "step": 50
+    },
+    {
+      "epoch": 0.2504604051565378,
+      "grad_norm": 0.5192026495933533,
+      "learning_rate": 4.4036148959228365e-06,
+      "loss": 0.4201,
+      "step": 51
+    },
+    {
+      "epoch": 0.25537139349294047,
+      "grad_norm": 0.5091439485549927,
+      "learning_rate": 4.3773961037202784e-06,
+      "loss": 0.4245,
+      "step": 52
+    },
+    {
+      "epoch": 0.26028238182934316,
+      "grad_norm": 0.5394258499145508,
+      "learning_rate": 4.350694992688289e-06,
+      "loss": 0.3993,
+      "step": 53
+    },
+    {
+      "epoch": 0.26519337016574585,
+      "grad_norm": 0.5097758173942566,
+      "learning_rate": 4.323518422566586e-06,
+      "loss": 0.3734,
+      "step": 54
+    },
+    {
+      "epoch": 0.27010435850214853,
+      "grad_norm": 0.5428377985954285,
+      "learning_rate": 4.295873375244319e-06,
+      "loss": 0.4114,
+      "step": 55
+    },
+    {
+      "epoch": 0.2750153468385513,
+      "grad_norm": 0.5589134693145752,
+      "learning_rate": 4.267766952966369e-06,
+      "loss": 0.4113,
+      "step": 56
+    },
+    {
+      "epoch": 0.27992633517495397,
+      "grad_norm": 0.5098860859870911,
+      "learning_rate": 4.239206376508716e-06,
+      "loss": 0.3883,
+      "step": 57
+    },
+    {
+      "epoch": 0.28483732351135665,
+      "grad_norm": 0.5086911916732788,
+      "learning_rate": 4.210198983323366e-06,
+      "loss": 0.3747,
+      "step": 58
+    },
+    {
+      "epoch": 0.28974831184775934,
+      "grad_norm": 0.5444743633270264,
+      "learning_rate": 4.1807522256532925e-06,
+      "loss": 0.4376,
+      "step": 59
+    },
+    {
+      "epoch": 0.2946593001841621,
+      "grad_norm": 0.528852641582489,
+      "learning_rate": 4.150873668617899e-06,
+      "loss": 0.4346,
+      "step": 60
+    },
+    {
+      "epoch": 0.2995702885205648,
+      "grad_norm": 0.5149591565132141,
+      "learning_rate": 4.120570988269472e-06,
+      "loss": 0.3967,
+      "step": 61
+    },
+    {
+      "epoch": 0.30448127685696746,
+      "grad_norm": 0.5277577638626099,
+      "learning_rate": 4.089851969621138e-06,
+      "loss": 0.3882,
+      "step": 62
+    },
+    {
+      "epoch": 0.30939226519337015,
+      "grad_norm": 0.5042468905448914,
+      "learning_rate": 4.058724504646834e-06,
+      "loss": 0.4036,
+      "step": 63
+    },
+    {
+      "epoch": 0.31430325352977284,
+      "grad_norm": 0.547420859336853,
+      "learning_rate": 4.027196590253786e-06,
+      "loss": 0.4039,
+      "step": 64
+    },
+    {
+      "epoch": 0.3192142418661756,
+      "grad_norm": 0.5304770469665527,
+      "learning_rate": 3.99527632622804e-06,
+      "loss": 0.4365,
+      "step": 65
+    },
+    {
+      "epoch": 0.3241252302025783,
+      "grad_norm": 0.5116031169891357,
+      "learning_rate": 3.9629719131535595e-06,
+      "loss": 0.4037,
+      "step": 66
+    },
+    {
+      "epoch": 0.32903621853898096,
+      "grad_norm": 0.5439778566360474,
+      "learning_rate": 3.930291650305424e-06,
+      "loss": 0.4069,
+      "step": 67
+    },
+    {
+      "epoch": 0.33394720687538365,
+      "grad_norm": 0.5249338746070862,
+      "learning_rate": 3.897243933517679e-06,
+      "loss": 0.4031,
+      "step": 68
+    },
+    {
+      "epoch": 0.3388581952117864,
+      "grad_norm": 0.5422383546829224,
+      "learning_rate": 3.863837253026372e-06,
+      "loss": 0.4066,
+      "step": 69
+    },
+    {
+      "epoch": 0.3437691835481891,
+      "grad_norm": 0.5329310894012451,
+      "learning_rate": 3.830080191288342e-06,
+      "loss": 0.389,
+      "step": 70
+    },
+    {
+      "epoch": 0.34868017188459177,
+      "grad_norm": 0.5688086748123169,
+      "learning_rate": 3.7959814207763134e-06,
+      "loss": 0.396,
+      "step": 71
+    },
+    {
+      "epoch": 0.35359116022099446,
+      "grad_norm": 0.5461124181747437,
+      "learning_rate": 3.761549701750865e-06,
+      "loss": 0.3878,
+      "step": 72
+    },
+    {
+      "epoch": 0.35850214855739715,
+      "grad_norm": 0.5189567804336548,
+      "learning_rate": 3.7267938800098454e-06,
+      "loss": 0.4148,
+      "step": 73
+    },
+    {
+      "epoch": 0.3634131368937999,
+      "grad_norm": 0.549319326877594,
+      "learning_rate": 3.691722884615814e-06,
+      "loss": 0.4048,
+      "step": 74
+    },
+    {
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.545100212097168,
+      "learning_rate": 3.656345725602089e-06,
+      "loss": 0.3923,
+      "step": 75
+    },
+    {
+      "epoch": 0.37323511356660527,
+      "grad_norm": 0.5365733504295349,
+      "learning_rate": 3.6206714916579925e-06,
+      "loss": 0.4059,
+      "step": 76
+    },
+    {
+      "epoch": 0.37814610190300796,
+      "grad_norm": 0.5217979550361633,
+      "learning_rate": 3.5847093477938955e-06,
+      "loss": 0.4082,
+      "step": 77
+    },
+    {
+      "epoch": 0.3830570902394107,
+      "grad_norm": 0.5529681444168091,
+      "learning_rate": 3.5484685329866424e-06,
+      "loss": 0.4119,
+      "step": 78
+    },
+    {
+      "epoch": 0.3879680785758134,
+      "grad_norm": 0.5729206800460815,
+      "learning_rate": 3.5119583578059845e-06,
+      "loss": 0.4001,
+      "step": 79
+    },
+    {
+      "epoch": 0.3928790669122161,
+      "grad_norm": 0.5517897605895996,
+      "learning_rate": 3.4751882020226174e-06,
+      "loss": 0.4347,
+      "step": 80
+    },
+    {
+      "epoch": 0.39779005524861877,
+      "grad_norm": 0.553726077079773,
+      "learning_rate": 3.438167512198436e-06,
+      "loss": 0.3889,
+      "step": 81
+    },
+    {
+      "epoch": 0.4027010435850215,
+      "grad_norm": 0.5306791663169861,
+      "learning_rate": 3.400905799259634e-06,
+      "loss": 0.3909,
+      "step": 82
+    },
+    {
+      "epoch": 0.4076120319214242,
+      "grad_norm": 0.5120440721511841,
+      "learning_rate": 3.3634126360532694e-06,
+      "loss": 0.4161,
+      "step": 83
+    },
+    {
+      "epoch": 0.4125230202578269,
+      "grad_norm": 0.5085780024528503,
+      "learning_rate": 3.3256976548879183e-06,
+      "loss": 0.3894,
+      "step": 84
+    },
+    {
+      "epoch": 0.4174340085942296,
+      "grad_norm": 0.5362045168876648,
+      "learning_rate": 3.2877705450590525e-06,
+      "loss": 0.4025,
+      "step": 85
+    },
+    {
+      "epoch": 0.42234499693063227,
+      "grad_norm": 0.5237392783164978,
+      "learning_rate": 3.249641050359779e-06,
+      "loss": 0.3824,
+      "step": 86
+    },
+    {
+      "epoch": 0.427255985267035,
+      "grad_norm": 0.5430594682693481,
+      "learning_rate": 3.2113189665775812e-06,
+      "loss": 0.4145,
+      "step": 87
+    },
+    {
+      "epoch": 0.4321669736034377,
+      "grad_norm": 0.5262824892997742,
+      "learning_rate": 3.172814138977692e-06,
+      "loss": 0.4059,
+      "step": 88
+    },
+    {
+      "epoch": 0.4370779619398404,
+      "grad_norm": 0.5239981412887573,
+      "learning_rate": 3.1341364597737684e-06,
+      "loss": 0.4219,
+      "step": 89
+    },
+    {
+      "epoch": 0.4419889502762431,
+      "grad_norm": 0.5327151417732239,
+      "learning_rate": 3.0952958655864957e-06,
+      "loss": 0.4042,
+      "step": 90
+    },
+    {
+      "epoch": 0.4468999386126458,
+      "grad_norm": 0.5039759874343872,
+      "learning_rate": 3.056302334890786e-06,
+      "loss": 0.3886,
+      "step": 91
+    },
+    {
+      "epoch": 0.4518109269490485,
+      "grad_norm": 0.5025433897972107,
+      "learning_rate": 3.0171658854522274e-06,
+      "loss": 0.3966,
+      "step": 92
+    },
+    {
+      "epoch": 0.4567219152854512,
+      "grad_norm": 0.4931771755218506,
+      "learning_rate": 2.9778965717534314e-06,
+      "loss": 0.3676,
+      "step": 93
+    },
+    {
+      "epoch": 0.4616329036218539,
+      "grad_norm": 0.5280553698539734,
+      "learning_rate": 2.9385044824109544e-06,
+      "loss": 0.3805,
+      "step": 94
+    },
+    {
+      "epoch": 0.4665438919582566,
+      "grad_norm": 0.5478957295417786,
+      "learning_rate": 2.8989997375834485e-06,
+      "loss": 0.4195,
+      "step": 95
+    },
+    {
+      "epoch": 0.4714548802946593,
+      "grad_norm": 0.49035656452178955,
+      "learning_rate": 2.859392486371705e-06,
+      "loss": 0.3787,
+      "step": 96
+    },
+    {
+      "epoch": 0.476365868631062,
+      "grad_norm": 0.509602963924408,
+      "learning_rate": 2.8196929042112652e-06,
+      "loss": 0.3883,
+      "step": 97
+    },
+    {
+      "epoch": 0.4812768569674647,
+      "grad_norm": 0.4914324879646301,
+      "learning_rate": 2.7799111902582697e-06,
+      "loss": 0.362,
+      "step": 98
+    },
+    {
+      "epoch": 0.4861878453038674,
+      "grad_norm": 0.5700349807739258,
+      "learning_rate": 2.7400575647692046e-06,
+      "loss": 0.3904,
+      "step": 99
+    },
+    {
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.5072741508483887,
+      "learning_rate": 2.7001422664752338e-06,
+      "loss": 0.3682,
+      "step": 100
+    },
+    {
+      "epoch": 0.4960098219766728,
+      "grad_norm": 0.4965267777442932,
+      "learning_rate": 2.6601755499517826e-06,
+      "loss": 0.3912,
+      "step": 101
+    },
+    {
+      "epoch": 0.5009208103130756,
+      "grad_norm": 0.48275378346443176,
+      "learning_rate": 2.620167682984052e-06,
+      "loss": 0.3942,
+      "step": 102
+    },
+    {
+      "epoch": 0.5058317986494782,
+      "grad_norm": 0.4831352233886719,
+      "learning_rate": 2.580128943929139e-06,
+      "loss": 0.3893,
+      "step": 103
+    },
+    {
+      "epoch": 0.5107427869858809,
+      "grad_norm": 0.5418382287025452,
+      "learning_rate": 2.5400696190754347e-06,
+      "loss": 0.4,
+      "step": 104
+    },
+    {
+      "epoch": 0.5156537753222836,
+      "grad_norm": 0.5543417930603027,
+      "learning_rate": 2.5e-06,
+      "loss": 0.3997,
+      "step": 105
+    },
+    {
+      "epoch": 0.5205647636586863,
+      "grad_norm": 0.49921315908432007,
+      "learning_rate": 2.459930380924566e-06,
+      "loss": 0.3896,
+      "step": 106
+    },
+    {
+      "epoch": 0.525475751995089,
+      "grad_norm": 0.529085636138916,
+      "learning_rate": 2.4198710560708623e-06,
+      "loss": 0.3933,
+      "step": 107
+    },
+    {
+      "epoch": 0.5303867403314917,
+      "grad_norm": 0.5602535605430603,
+      "learning_rate": 2.3798323170159487e-06,
+      "loss": 0.398,
+      "step": 108
+    },
+    {
+      "epoch": 0.5352977286678944,
+      "grad_norm": 0.5335884094238281,
+      "learning_rate": 2.339824450048218e-06,
+      "loss": 0.386,
+      "step": 109
+    },
+    {
+      "epoch": 0.5402087170042971,
+      "grad_norm": 0.5006181597709656,
+      "learning_rate": 2.299857733524767e-06,
+      "loss": 0.3705,
+      "step": 110
+    },
+    {
+      "epoch": 0.5451197053406999,
+      "grad_norm": 0.5093613862991333,
+      "learning_rate": 2.2599424352307958e-06,
+      "loss": 0.3953,
+      "step": 111
+    },
+    {
+      "epoch": 0.5500306936771026,
+      "grad_norm": 0.4948323667049408,
+      "learning_rate": 2.2200888097417308e-06,
+      "loss": 0.379,
+      "step": 112
+    },
+    {
+      "epoch": 0.5549416820135052,
+      "grad_norm": 0.5212745070457458,
+      "learning_rate": 2.1803070957887348e-06,
+      "loss": 0.4005,
+      "step": 113
+    },
+    {
+      "epoch": 0.5598526703499079,
+      "grad_norm": 0.5045080184936523,
+      "learning_rate": 2.140607513628296e-06,
+      "loss": 0.3904,
+      "step": 114
+    },
+    {
+      "epoch": 0.5647636586863106,
+      "grad_norm": 0.47190696001052856,
+      "learning_rate": 2.1010002624165528e-06,
+      "loss": 0.3659,
+      "step": 115
+    },
+    {
+      "epoch": 0.5696746470227133,
+      "grad_norm": 0.5220581293106079,
+      "learning_rate": 2.0614955175890464e-06,
+      "loss": 0.4218,
+      "step": 116
+    },
+    {
+      "epoch": 0.574585635359116,
+      "grad_norm": 0.5442221164703369,
+      "learning_rate": 2.02210342824657e-06,
+      "loss": 0.4194,
+      "step": 117
+    },
+    {
+      "epoch": 0.5794966236955187,
+      "grad_norm": 0.5178117752075195,
+      "learning_rate": 1.982834114547773e-06,
+      "loss": 0.4065,
+      "step": 118
+    },
+    {
+      "epoch": 0.5844076120319214,
+      "grad_norm": 0.5385434031486511,
+      "learning_rate": 1.9436976651092143e-06,
+      "loss": 0.3913,
+      "step": 119
+    },
+    {
+      "epoch": 0.5893186003683242,
+      "grad_norm": 0.5246177911758423,
+      "learning_rate": 1.9047041344135045e-06,
+      "loss": 0.4051,
+      "step": 120
+    },
+    {
+      "epoch": 0.5942295887047269,
+      "grad_norm": 0.4902905225753784,
+      "learning_rate": 1.865863540226232e-06,
+      "loss": 0.3745,
+      "step": 121
+    },
+    {
+      "epoch": 0.5991405770411296,
+      "grad_norm": 0.5408722162246704,
+      "learning_rate": 1.827185861022308e-06,
+      "loss": 0.4129,
+      "step": 122
+    },
+    {
+      "epoch": 0.6040515653775322,
+      "grad_norm": 0.5140423774719238,
+      "learning_rate": 1.7886810334224192e-06,
+      "loss": 0.35,
+      "step": 123
+    },
+    {
+      "epoch": 0.6089625537139349,
+      "grad_norm": 0.5432409048080444,
+      "learning_rate": 1.750358949640221e-06,
+      "loss": 0.4009,
+      "step": 124
+    },
+    {
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.5189937353134155,
+      "learning_rate": 1.7122294549409486e-06,
+      "loss": 0.4074,
+      "step": 125
+    },
+    {
+      "epoch": 0.6187845303867403,
+      "grad_norm": 0.5263358354568481,
+      "learning_rate": 1.6743023451120831e-06,
+      "loss": 0.4062,
+      "step": 126
+    },
+    {
+      "epoch": 0.623695518723143,
+      "grad_norm": 0.5967735052108765,
+      "learning_rate": 1.6365873639467314e-06,
+      "loss": 0.387,
+      "step": 127
+    },
+    {
+      "epoch": 0.6286065070595457,
+      "grad_norm": 0.5754578113555908,
+      "learning_rate": 1.599094200740367e-06,
+      "loss": 0.3793,
+      "step": 128
+    },
+    {
+      "epoch": 0.6335174953959485,
+      "grad_norm": 0.5088539123535156,
+      "learning_rate": 1.561832487801565e-06,
+      "loss": 0.4056,
+      "step": 129
+    },
+    {
+      "epoch": 0.6384284837323512,
+      "grad_norm": 0.534451961517334,
+      "learning_rate": 1.524811797977383e-06,
+      "loss": 0.4064,
+      "step": 130
+    },
+    {
+      "epoch": 0.6433394720687539,
+      "grad_norm": 0.529793381690979,
+      "learning_rate": 1.4880416421940155e-06,
+      "loss": 0.4017,
+      "step": 131
+    },
+    {
+      "epoch": 0.6482504604051565,
+      "grad_norm": 0.5185397863388062,
+      "learning_rate": 1.4515314670133582e-06,
+      "loss": 0.4165,
+      "step": 132
+    },
+    {
+      "epoch": 0.6531614487415592,
+      "grad_norm": 0.5312457084655762,
+      "learning_rate": 1.415290652206105e-06,
+      "loss": 0.3883,
+      "step": 133
+    },
+    {
+      "epoch": 0.6580724370779619,
+      "grad_norm": 0.5226163864135742,
+      "learning_rate": 1.3793285083420077e-06,
+      "loss": 0.3851,
+      "step": 134
+    },
+    {
+      "epoch": 0.6629834254143646,
+      "grad_norm": 0.5747395157814026,
+      "learning_rate": 1.3436542743979125e-06,
+      "loss": 0.4047,
+      "step": 135
+    },
+    {
+      "epoch": 0.6678944137507673,
+      "grad_norm": 0.5257447361946106,
+      "learning_rate": 1.3082771153841872e-06,
+      "loss": 0.3868,
+      "step": 136
+    },
+    {
+      "epoch": 0.67280540208717,
+      "grad_norm": 0.5245814919471741,
+      "learning_rate": 1.2732061199901563e-06,
+      "loss": 0.3799,
+      "step": 137
+    },
+    {
+      "epoch": 0.6777163904235728,
+      "grad_norm": 0.5371440052986145,
+      "learning_rate": 1.2384502982491359e-06,
+      "loss": 0.3884,
+      "step": 138
+    },
+    {
+      "epoch": 0.6826273787599755,
+      "grad_norm": 0.659095048904419,
+      "learning_rate": 1.2040185792236874e-06,
+      "loss": 0.3931,
+      "step": 139
+    },
+    {
+      "epoch": 0.6875383670963782,
+      "grad_norm": 0.5353334546089172,
+      "learning_rate": 1.169919808711659e-06,
+      "loss": 0.3888,
+      "step": 140
+    },
+    {
+      "epoch": 0.6924493554327809,
+      "grad_norm": 0.5111743211746216,
+      "learning_rate": 1.1361627469736286e-06,
+      "loss": 0.4002,
+      "step": 141
+    },
+    {
+      "epoch": 0.6973603437691835,
+      "grad_norm": 0.574226975440979,
+      "learning_rate": 1.1027560664823208e-06,
+      "loss": 0.4165,
+      "step": 142
+    },
+    {
+      "epoch": 0.7022713321055862,
+      "grad_norm": 0.5593245029449463,
+      "learning_rate": 1.0697083496945766e-06,
+      "loss": 0.3974,
+      "step": 143
+    },
+    {
+      "epoch": 0.7071823204419889,
+      "grad_norm": 0.549225389957428,
+      "learning_rate": 1.0370280868464405e-06,
+      "loss": 0.3717,
+      "step": 144
+    },
+    {
+      "epoch": 0.7120933087783916,
+      "grad_norm": 0.5222504138946533,
+      "learning_rate": 1.00472367377196e-06,
+      "loss": 0.392,
+      "step": 145
+    },
+    {
+      "epoch": 0.7170042971147943,
+      "grad_norm": 0.47291433811187744,
+      "learning_rate": 9.728034097462144e-07,
+      "loss": 0.3711,
+      "step": 146
+    },
+    {
+      "epoch": 0.7219152854511971,
+      "grad_norm": 0.49495929479599,
+      "learning_rate": 9.412754953531664e-07,
+      "loss": 0.3861,
+      "step": 147
+    },
+    {
+      "epoch": 0.7268262737875998,
+      "grad_norm": 0.5076296329498291,
+      "learning_rate": 9.101480303788623e-07,
+      "loss": 0.3755,
+      "step": 148
+    },
+    {
+      "epoch": 0.7317372621240025,
+      "grad_norm": 0.4925084710121155,
+      "learning_rate": 8.794290117305296e-07,
+      "loss": 0.3799,
+      "step": 149
+    },
+    {
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.5179643630981445,
+      "learning_rate": 8.491263313821021e-07,
+      "loss": 0.3858,
+      "step": 150
+    },
+    {
+      "epoch": 0.7415592387968079,
+      "grad_norm": 0.5128403902053833,
+      "learning_rate": 8.192477743467078e-07,
+      "loss": 0.3777,
+      "step": 151
+    },
+    {
+      "epoch": 0.7464702271332105,
+      "grad_norm": 0.5010114312171936,
+      "learning_rate": 7.898010166766348e-07,
+      "loss": 0.4331,
+      "step": 152
+    },
+    {
+      "epoch": 0.7513812154696132,
+      "grad_norm": 0.4962404668331146,
+      "learning_rate": 7.607936234912841e-07,
+      "loss": 0.3726,
+      "step": 153
+    },
+    {
+      "epoch": 0.7562922038060159,
+      "grad_norm": 0.5189750790596008,
+      "learning_rate": 7.322330470336314e-07,
+      "loss": 0.383,
+      "step": 154
+    },
+    {
+      "epoch": 0.7612031921424187,
+      "grad_norm": 0.5114901065826416,
+      "learning_rate": 7.041266247556814e-07,
+      "loss": 0.3809,
+      "step": 155
+    },
+    {
+      "epoch": 0.7661141804788214,
+      "grad_norm": 0.4884895980358124,
+      "learning_rate": 6.764815774334149e-07,
+      "loss": 0.4129,
+      "step": 156
+    },
+    {
+      "epoch": 0.7710251688152241,
+      "grad_norm": 0.46929967403411865,
+      "learning_rate": 6.493050073117115e-07,
+      "loss": 0.3691,
+      "step": 157
+    },
+    {
+      "epoch": 0.7759361571516268,
+      "grad_norm": 0.4958886504173279,
+      "learning_rate": 6.226038962797218e-07,
+      "loss": 0.3751,
+      "step": 158
+    },
+    {
+      "epoch": 0.7808471454880295,
+      "grad_norm": 0.4874633848667145,
+      "learning_rate": 5.963851040771639e-07,
+      "loss": 0.384,
+      "step": 159
+    },
+    {
+      "epoch": 0.7857581338244322,
+      "grad_norm": 0.5399041175842285,
+      "learning_rate": 5.706553665319955e-07,
+      "loss": 0.4117,
+      "step": 160
+    },
+    {
+      "epoch": 0.7906691221608348,
+      "grad_norm": 0.5015324354171753,
+      "learning_rate": 5.454212938299256e-07,
+      "loss": 0.3742,
+      "step": 161
+    },
+    {
+      "epoch": 0.7955801104972375,
+      "grad_norm": 0.5084665417671204,
+      "learning_rate": 5.20689368816201e-07,
+      "loss": 0.3882,
+      "step": 162
+    },
+    {
+      "epoch": 0.8004910988336402,
+      "grad_norm": 0.48962369561195374,
+      "learning_rate": 4.964659453301088e-07,
+      "loss": 0.3867,
+      "step": 163
+    },
+    {
+      "epoch": 0.805402087170043,
+      "grad_norm": 0.511327862739563,
+      "learning_rate": 4.7275724657262293e-07,
+      "loss": 0.3864,
+      "step": 164
+    },
+    {
+      "epoch": 0.8103130755064457,
+      "grad_norm": 0.5067377686500549,
+      "learning_rate": 4.495693635076101e-07,
+      "loss": 0.3751,
+      "step": 165
+    },
+    {
+      "epoch": 0.8152240638428484,
+      "grad_norm": 0.49379196763038635,
+      "learning_rate": 4.2690825329701313e-07,
+      "loss": 0.3971,
+      "step": 166
+    },
+    {
+      "epoch": 0.8201350521792511,
+      "grad_norm": 0.47511026263237,
+      "learning_rate": 4.0477973777039854e-07,
+      "loss": 0.3625,
+      "step": 167
+    },
+    {
+      "epoch": 0.8250460405156538,
+      "grad_norm": 0.5097380876541138,
+      "learning_rate": 3.831895019292897e-07,
+      "loss": 0.4035,
+      "step": 168
+    },
+    {
+      "epoch": 0.8299570288520565,
+      "grad_norm": 0.4871266484260559,
+      "learning_rate": 3.621430924866348e-07,
+      "loss": 0.39,
+      "step": 169
+    },
+    {
+      "epoch": 0.8348680171884592,
+      "grad_norm": 0.5044364333152771,
+      "learning_rate": 3.4164591644181233e-07,
+      "loss": 0.385,
+      "step": 170
+    },
+    {
+      "epoch": 0.8397790055248618,
+      "grad_norm": 0.46992751955986023,
+      "learning_rate": 3.217032396915265e-07,
+      "loss": 0.366,
+      "step": 171
+    },
+    {
+      "epoch": 0.8446899938612645,
+      "grad_norm": 0.4705800712108612,
+      "learning_rate": 3.02320185676957e-07,
+      "loss": 0.352,
+      "step": 172
+    },
+    {
+      "epoch": 0.8496009821976673,
+      "grad_norm": 0.4910837709903717,
+      "learning_rate": 2.8350173406749975e-07,
+      "loss": 0.3883,
+      "step": 173
+    },
+    {
+      "epoch": 0.85451197053407,
+      "grad_norm": 0.49292171001434326,
+      "learning_rate": 2.652527194814511e-07,
+      "loss": 0.3799,
+      "step": 174
+    },
+    {
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.5116161108016968,
+      "learning_rate": 2.4757783024395244e-07,
+      "loss": 0.3845,
+      "step": 175
+    },
+    {
+      "epoch": 0.8643339472068754,
+      "grad_norm": 0.4837781488895416,
+      "learning_rate": 2.3048160718252e-07,
+      "loss": 0.3717,
+      "step": 176
+    },
+    {
+      "epoch": 0.8692449355432781,
+      "grad_norm": 0.5189177989959717,
+      "learning_rate": 2.1396844246046904e-07,
+      "loss": 0.4046,
+      "step": 177
+    },
+    {
+      "epoch": 0.8741559238796808,
+      "grad_norm": 0.49742627143859863,
+      "learning_rate": 1.980425784485293e-07,
+      "loss": 0.3988,
+      "step": 178
+    },
+    {
+      "epoch": 0.8790669122160835,
+      "grad_norm": 0.4957612454891205,
+      "learning_rate": 1.8270810663494591e-07,
+      "loss": 0.373,
+      "step": 179
+    },
+    {
+      "epoch": 0.8839779005524862,
+      "grad_norm": 0.4809194505214691,
+      "learning_rate": 1.679689665743381e-07,
+      "loss": 0.3609,
+      "step": 180
+    },
+    {
+      "epoch": 0.8888888888888888,
+      "grad_norm": 0.5370915532112122,
+      "learning_rate": 1.538289448755989e-07,
+      "loss": 0.4122,
+      "step": 181
+    },
+    {
+      "epoch": 0.8937998772252916,
+      "grad_norm": 0.4931313693523407,
+      "learning_rate": 1.4029167422908107e-07,
+      "loss": 0.3752,
+      "step": 182
+    },
+    {
+      "epoch": 0.8987108655616943,
+      "grad_norm": 0.5005168914794922,
+      "learning_rate": 1.273606324733284e-07,
+      "loss": 0.3935,
+      "step": 183
+    },
+    {
+      "epoch": 0.903621853898097,
+      "grad_norm": 0.5079485177993774,
+      "learning_rate": 1.1503914170159058e-07,
+      "loss": 0.3885,
+      "step": 184
+    },
+    {
+      "epoch": 0.9085328422344997,
+      "grad_norm": 0.5038573145866394,
+      "learning_rate": 1.0333036740834857e-07,
+      "loss": 0.3889,
+      "step": 185
+    },
+    {
+      "epoch": 0.9134438305709024,
+      "grad_norm": 0.5191468000411987,
+      "learning_rate": 9.223731767607436e-08,
+      "loss": 0.4182,
+      "step": 186
+    },
+    {
+      "epoch": 0.9183548189073051,
+      "grad_norm": 0.5079477429389954,
+      "learning_rate": 8.176284240242638e-08,
+      "loss": 0.4103,
+      "step": 187
+    },
+    {
+      "epoch": 0.9232658072437078,
+      "grad_norm": 0.5107765197753906,
+      "learning_rate": 7.190963256809069e-08,
+      "loss": 0.3801,
+      "step": 188
+    },
+    {
+      "epoch": 0.9281767955801105,
+      "grad_norm": 0.5129873752593994,
+      "learning_rate": 6.268021954544095e-08,
+      "loss": 0.4097,
+      "step": 189
+    },
+    {
+      "epoch": 0.9330877839165131,
+      "grad_norm": 0.4807911515235901,
+      "learning_rate": 5.407697444821169e-08,
+      "loss": 0.392,
+      "step": 190
+    },
+    {
+      "epoch": 0.937998772252916,
+      "grad_norm": 0.5057871341705322,
+      "learning_rate": 4.61021075223364e-08,
+      "loss": 0.4033,
+      "step": 191
+    },
+    {
+      "epoch": 0.9429097605893186,
+      "grad_norm": 0.5155328512191772,
+      "learning_rate": 3.8757667578119e-08,
+      "loss": 0.3782,
+      "step": 192
+    },
+    {
+      "epoch": 0.9478207489257213,
+      "grad_norm": 0.4686456322669983,
+      "learning_rate": 3.204554146387456e-08,
+      "loss": 0.3531,
+      "step": 193
+    },
+    {
+      "epoch": 0.952731737262124,
+      "grad_norm": 0.4980442523956299,
+      "learning_rate": 2.5967453581185187e-08,
+      "loss": 0.3826,
+      "step": 194
+    },
+    {
+      "epoch": 0.9576427255985267,
+      "grad_norm": 0.49153003096580505,
+      "learning_rate": 2.052496544188487e-08,
+      "loss": 0.3778,
+      "step": 195
+    },
+    {
+      "epoch": 0.9625537139349294,
+      "grad_norm": 0.4738159775733948,
+      "learning_rate": 1.571947526689349e-08,
+      "loss": 0.3513,
+      "step": 196
+    },
+    {
+      "epoch": 0.9674647022713321,
+      "grad_norm": 0.4973224997520447,
+      "learning_rate": 1.1552217627004426e-08,
+      "loss": 0.3937,
+      "step": 197
+    },
+    {
+      "epoch": 0.9723756906077348,
+      "grad_norm": 0.46480733156204224,
+      "learning_rate": 8.024263125710751e-09,
+      "loss": 0.3788,
+      "step": 198
+    },
+    {
+      "epoch": 0.9772866789441375,
+      "grad_norm": 0.4957633912563324,
+      "learning_rate": 5.136518124159162e-09,
+      "loss": 0.4029,
+      "step": 199
+    },
+    {
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.48968443274497986,
+      "learning_rate": 2.8897245082978863e-09,
+      "loss": 0.3644,
+      "step": 200
+    },
+    {
+      "epoch": 0.9871086556169429,
+      "grad_norm": 0.4898364245891571,
+      "learning_rate": 1.284459498280266e-09,
+      "loss": 0.4098,
+      "step": 201
+    },
+    {
+      "epoch": 0.9920196439533456,
+      "grad_norm": 0.5081222653388977,
+      "learning_rate": 3.2113550017198734e-10,
+      "loss": 0.3883,
+      "step": 202
+    },
+    {
+      "epoch": 0.9969306322897483,
+      "grad_norm": 0.49424201250076294,
+      "learning_rate": 0.0,
+      "loss": 0.396,
+      "step": 203
+    },
+    {
+      "epoch": 0.9969306322897483,
+      "step": 203,
+      "total_flos": 2.835499283468255e+18,
+      "train_loss": 0.40513286713896124,
+      "train_runtime": 4111.0607,
+      "train_samples_per_second": 12.677,
+      "train_steps_per_second": 0.049
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 203,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 400,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.835499283468255e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:555acacf1a7bb6da1c0c32e504c76de454d368dc1b9359b2d6d28af7f2f98593
+size 7160