upload the 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 +1,61 @@
+---
+library_name: transformers
+license: other
+base_model: henryhe0123/pc-agent-72b
+tags:
+- llama-factory
+- full
+- generated_from_trainer
+model-index:
+- name: Qwen2.5-VL-72B-sft-40
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# Qwen2.5-VL-72B-sft-40
+
+This model is a fine-tuned version of [/inspire/hdd/global_user/liupengfei-24025/yhhe/model/Qwen2.5-VL-72B-Instruct](https://huggingface.co//inspire/hdd/global_user/liupengfei-24025/yhhe/model/Qwen2.5-VL-72B-Instruct) on the pcagent40 dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 2e-06
+- train_batch_size: 2
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 32
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 128
+- total_eval_batch_size: 256
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.05
+- num_epochs: 2
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 4.49.0.dev0
+- Pytorch 2.6.0+cu124
+- Datasets 3.3.2
+- Tokenizers 0.21.0

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": 1.9928400954653938,
+    "total_flos": 3166299160051712.0,
+    "train_loss": 0.5405693022828353,
+    "train_runtime": 17864.3337,
+    "train_samples_per_second": 2.995,
+    "train_steps_per_second": 0.023
+}

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}"
+}

config.json

@@ -0,0 +1,52 @@
+{
+  "_name_or_path": "/inspire/hdd/global_user/liupengfei-24025/yhhe/model/Qwen2.5-VL-72B-Instruct",
+  "architectures": [
+    "Qwen2_5_VLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 8192,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 29568,
+  "max_position_embeddings": 128000,
+  "max_window_layers": 80,
+  "model_type": "qwen2_5_vl",
+  "num_attention_heads": 64,
+  "num_hidden_layers": 80,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "hidden_size": 1280,
+    "in_chans": 3,
+    "intermediate_size": 3456,
+    "model_type": "qwen2_5_vl",
+    "out_hidden_size": 8192,
+    "spatial_patch_size": 14,
+    "tokens_per_second": 2,
+    "torch_dtype": "float32"
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,13 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "top_k": 1,
+  "top_p": 0.001,
+  "transformers_version": "4.49.0.dev0"
+}

model-00001-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89873f18da2eb3689c4fc7c8937ac526a90b25f14a8c9d794d228fb6fcccae8d
+size 4687059888

model-00002-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ee4274a122f0b2b8cf263677e27c4df0911a10c37bb875eb9d86d2b00f76344
+size 4781670320

model-00003-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d281ad03a3230de926a0cba30502bb4b4f83599f46380c5c1d783129fabe5b67
+size 4964101384

model-00004-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc60786bedb6268c9151a0abb2f0cd974d30d4b4fbbbaf2c4219e9f9a1fc58a4
+size 4781637328

model-00005-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0938365164ead23c51439b2c2c49a3c8a81865d0d196614ec10f4e140ae0b1ca
+size 4781670344

model-00006-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d0df222e48b5736487c52718dee1da94cd8eff029ba2bb105a2a94ea8777bfe
+size 4781670360

model-00007-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecbbc62226ba5e53bcda6ac88644e1d897be07daa2ec8cf0fdac374037583dce
+size 4964101416

model-00008-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:203b258dbf34c61a5145864c2abe0f2d94f73637848e2fe74732d304a1e08c76
+size 4781637360

model-00009-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c364bf91653c18df2c23a3970e609ab96286f97af6bbc48d749d86e8ca53168b
+size 4781670360

model-00010-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5301004442d9307716b55531227e9a1cf99af8e74c52bffd639faf234dda1b72
+size 4781670360

model-00011-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17fc3bcd94029359251b1c0fdc88330216c7303730520bba83c82c1628cc5c37
+size 4964101416

model-00012-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:558f9d5c332089425007cce70c95cdaa167ba3a894a3dcb793fd139c63f0437e
+size 4781637360

model-00013-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:992756df5404419cc2440b2731ab41bedcf015a033d114bb6d13e29665163633
+size 4781670360

model-00014-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a786fb40a60578166262ea413495cb102a014991f0d80f484cc64663d6fec846
+size 4781670360

model-00015-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b647789a9126586d496555feb45623a85781f93093baaec3211c208f0948117
+size 4964101416

model-00016-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89e9c8c0f0a6b653b0ef2cf2d24ca2c061a48dc3e7a0c7b400bc8599d39af405
+size 4781637360

model-00017-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07cc5af11897595b8ccbb31224ccbb154c95c416160717477c4c525aab1d3bfc
+size 4781670360

model-00018-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf63e44745e14e238ca8b7ce5a12b2ae5a4fbe6da3690741cffb5cd80ff10d7d
+size 4781670360

model-00019-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3904af6b6a0ba39e4ba73545645052858c9f537e805bc4ce54857b895112d7c
+size 4964101416

model-00020-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69263cb98eb627046dbfbdce540141945059835c490e9480eb39c6acce1df7f5
+size 4781637360

model-00021-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d291d0c6eaa424301a5a4b1176d78acc608e8eb174c72ec925204224be8ee857
+size 4781670360

model-00022-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:209ce9d296c2ecc19b8223cb27f2b639150916447cea4964afc1f6279303a2a7
+size 4781670360

model-00023-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:104f4f9ec871ed389217871bea605fd84bf478f7d2db45c1df985cbf1d82f24c
+size 4964101416

model-00024-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62162bd8001113713f88f4984bd513d4cb854049873c1144e4f6f9e2206e052b
+size 4781637360

model-00025-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b5beea41ffed149cda160f89cd029bc947ed0b4177707c3f76ec5c81b57f1fa
+size 4781670360

model-00026-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38c316e66c0603a2961a21d0b22d9c85ba7e132984d7375dc83fa24780cde6ae
+size 4781670360

model-00027-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca6ee881c748d6cb27891fb46b4342db0612a46c66320e9993885b3bcc18fa13
+size 4964101416

model-00028-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ea70ea03c970a2aa3cffb33b7446b68c9e98096f0b3a533b6cac4a70544ee09
+size 4781637360

model-00029-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c115ae87825f7357263cdf074febd66ad3c580e7c84f2ab1144e6d013c63ed59
+size 4781670360

model-00030-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10bda1b034c8194ae41ae7b05af58e4356576091203fa794b614bc031c7af626
+size 4479675656

model-00031-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37360565eccf4552c46854dcd2a68785f8dd40825e7ba6ebab89e4dae7160a77
+size 2491416704

preprocessor_config.json

@@ -0,0 +1,29 @@
+{
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Qwen2_5_VLImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_pixels": 3136,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "max_pixels": 12845056,
+    "min_pixels": 3136
+  },
+  "temporal_patch_size": 2
+}

special_tokens_map.json

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

tokenizer.json

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

tokenizer_config.json

@@ -0,0 +1,210 @@
+{
+  "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": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "processor_class": "Qwen2_5_VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 1.9928400954653938,
+    "total_flos": 3166299160051712.0,
+    "train_loss": 0.5405693022828353,
+    "train_runtime": 17864.3337,
+    "train_samples_per_second": 2.995,
+    "train_steps_per_second": 0.023
+}

trainer_log.jsonl

@@ -0,0 +1,419 @@
+{"current_steps": 1, "total_steps": 418, "loss": 1.0606, "lr": 9.523809523809523e-08, "epoch": 0.00477326968973747, "percentage": 0.24, "elapsed_time": "0:00:59", "remaining_time": "6:53:17"}
+{"current_steps": 2, "total_steps": 418, "loss": 1.0205, "lr": 1.9047619047619045e-07, "epoch": 0.00954653937947494, "percentage": 0.48, "elapsed_time": "0:01:39", "remaining_time": "5:44:44"}
+{"current_steps": 3, "total_steps": 418, "loss": 1.0092, "lr": 2.857142857142857e-07, "epoch": 0.014319809069212411, "percentage": 0.72, "elapsed_time": "0:02:19", "remaining_time": "5:21:07"}
+{"current_steps": 4, "total_steps": 418, "loss": 0.9935, "lr": 3.809523809523809e-07, "epoch": 0.01909307875894988, "percentage": 0.96, "elapsed_time": "0:02:58", "remaining_time": "5:07:47"}
+{"current_steps": 5, "total_steps": 418, "loss": 1.0184, "lr": 4.761904761904761e-07, "epoch": 0.02386634844868735, "percentage": 1.2, "elapsed_time": "0:03:36", "remaining_time": "4:58:38"}
+{"current_steps": 6, "total_steps": 418, "loss": 1.0072, "lr": 5.714285714285714e-07, "epoch": 0.028639618138424822, "percentage": 1.44, "elapsed_time": "0:04:16", "remaining_time": "4:53:52"}
+{"current_steps": 7, "total_steps": 418, "loss": 0.9987, "lr": 6.666666666666666e-07, "epoch": 0.03341288782816229, "percentage": 1.67, "elapsed_time": "0:04:56", "remaining_time": "4:50:06"}
+{"current_steps": 8, "total_steps": 418, "loss": 0.9639, "lr": 7.619047619047618e-07, "epoch": 0.03818615751789976, "percentage": 1.91, "elapsed_time": "0:05:35", "remaining_time": "4:46:40"}
+{"current_steps": 9, "total_steps": 418, "loss": 0.9497, "lr": 8.57142857142857e-07, "epoch": 0.04295942720763723, "percentage": 2.15, "elapsed_time": "0:06:14", "remaining_time": "4:43:39"}
+{"current_steps": 10, "total_steps": 418, "loss": 0.9488, "lr": 9.523809523809522e-07, "epoch": 0.0477326968973747, "percentage": 2.39, "elapsed_time": "0:06:54", "remaining_time": "4:41:52"}
+{"current_steps": 11, "total_steps": 418, "loss": 0.925, "lr": 1.0476190476190476e-06, "epoch": 0.05250596658711217, "percentage": 2.63, "elapsed_time": "0:07:35", "remaining_time": "4:40:37"}
+{"current_steps": 12, "total_steps": 418, "loss": 0.9231, "lr": 1.1428571428571428e-06, "epoch": 0.057279236276849645, "percentage": 2.87, "elapsed_time": "0:08:15", "remaining_time": "4:39:17"}
+{"current_steps": 13, "total_steps": 418, "loss": 0.9177, "lr": 1.238095238095238e-06, "epoch": 0.06205250596658711, "percentage": 3.11, "elapsed_time": "0:08:53", "remaining_time": "4:37:01"}
+{"current_steps": 14, "total_steps": 418, "loss": 0.8811, "lr": 1.3333333333333332e-06, "epoch": 0.06682577565632458, "percentage": 3.35, "elapsed_time": "0:09:30", "remaining_time": "4:34:22"}
+{"current_steps": 15, "total_steps": 418, "loss": 0.8533, "lr": 1.4285714285714286e-06, "epoch": 0.07159904534606205, "percentage": 3.59, "elapsed_time": "0:10:07", "remaining_time": "4:31:59"}
+{"current_steps": 16, "total_steps": 418, "loss": 0.8706, "lr": 1.5238095238095236e-06, "epoch": 0.07637231503579953, "percentage": 3.83, "elapsed_time": "0:10:44", "remaining_time": "4:30:03"}
+{"current_steps": 17, "total_steps": 418, "loss": 0.8208, "lr": 1.619047619047619e-06, "epoch": 0.081145584725537, "percentage": 4.07, "elapsed_time": "0:11:26", "remaining_time": "4:29:53"}
+{"current_steps": 18, "total_steps": 418, "loss": 0.8181, "lr": 1.714285714285714e-06, "epoch": 0.08591885441527446, "percentage": 4.31, "elapsed_time": "0:12:05", "remaining_time": "4:28:36"}
+{"current_steps": 19, "total_steps": 418, "loss": 0.8245, "lr": 1.8095238095238095e-06, "epoch": 0.09069212410501193, "percentage": 4.55, "elapsed_time": "0:12:45", "remaining_time": "4:27:56"}
+{"current_steps": 20, "total_steps": 418, "loss": 0.7545, "lr": 1.9047619047619045e-06, "epoch": 0.0954653937947494, "percentage": 4.78, "elapsed_time": "0:13:25", "remaining_time": "4:27:01"}
+{"current_steps": 21, "total_steps": 418, "loss": 0.776, "lr": 2e-06, "epoch": 0.10023866348448687, "percentage": 5.02, "elapsed_time": "0:14:02", "remaining_time": "4:25:21"}
+{"current_steps": 22, "total_steps": 418, "loss": 0.7445, "lr": 1.9999686897547167e-06, "epoch": 0.10501193317422435, "percentage": 5.26, "elapsed_time": "0:14:41", "remaining_time": "4:24:30"}
+{"current_steps": 23, "total_steps": 418, "loss": 0.7351, "lr": 1.9998747609795305e-06, "epoch": 0.10978520286396182, "percentage": 5.5, "elapsed_time": "0:15:21", "remaining_time": "4:23:48"}
+{"current_steps": 24, "total_steps": 418, "loss": 0.7332, "lr": 1.999718219556307e-06, "epoch": 0.11455847255369929, "percentage": 5.74, "elapsed_time": "0:15:58", "remaining_time": "4:22:21"}
+{"current_steps": 25, "total_steps": 418, "loss": 0.738, "lr": 1.999499075287747e-06, "epoch": 0.11933174224343675, "percentage": 5.98, "elapsed_time": "0:16:38", "remaining_time": "4:21:29"}
+{"current_steps": 26, "total_steps": 418, "loss": 0.7089, "lr": 1.999217341896772e-06, "epoch": 0.12410501193317422, "percentage": 6.22, "elapsed_time": "0:17:15", "remaining_time": "4:20:10"}
+{"current_steps": 27, "total_steps": 418, "loss": 0.6936, "lr": 1.998873037025665e-06, "epoch": 0.1288782816229117, "percentage": 6.46, "elapsed_time": "0:17:52", "remaining_time": "4:18:49"}
+{"current_steps": 28, "total_steps": 418, "loss": 0.6785, "lr": 1.9984661822349665e-06, "epoch": 0.13365155131264916, "percentage": 6.7, "elapsed_time": "0:18:28", "remaining_time": "4:17:24"}
+{"current_steps": 29, "total_steps": 418, "loss": 0.6978, "lr": 1.997996803002123e-06, "epoch": 0.13842482100238662, "percentage": 6.94, "elapsed_time": "0:19:06", "remaining_time": "4:16:19"}
+{"current_steps": 30, "total_steps": 418, "loss": 0.669, "lr": 1.9974649287198914e-06, "epoch": 0.1431980906921241, "percentage": 7.18, "elapsed_time": "0:19:47", "remaining_time": "4:16:01"}
+{"current_steps": 31, "total_steps": 418, "loss": 0.6674, "lr": 1.9968705926945013e-06, "epoch": 0.14797136038186157, "percentage": 7.42, "elapsed_time": "0:20:26", "remaining_time": "4:15:05"}
+{"current_steps": 32, "total_steps": 418, "loss": 0.6765, "lr": 1.9962138321435656e-06, "epoch": 0.15274463007159905, "percentage": 7.66, "elapsed_time": "0:21:04", "remaining_time": "4:14:12"}
+{"current_steps": 33, "total_steps": 418, "loss": 0.6745, "lr": 1.9954946881937524e-06, "epoch": 0.1575178997613365, "percentage": 7.89, "elapsed_time": "0:21:42", "remaining_time": "4:13:12"}
+{"current_steps": 34, "total_steps": 418, "loss": 0.6546, "lr": 1.994713205878208e-06, "epoch": 0.162291169451074, "percentage": 8.13, "elapsed_time": "0:22:19", "remaining_time": "4:12:10"}
+{"current_steps": 35, "total_steps": 418, "loss": 0.6612, "lr": 1.9938694341337393e-06, "epoch": 0.16706443914081145, "percentage": 8.37, "elapsed_time": "0:22:58", "remaining_time": "4:11:23"}
+{"current_steps": 36, "total_steps": 418, "loss": 0.6525, "lr": 1.9929634257977467e-06, "epoch": 0.1718377088305489, "percentage": 8.61, "elapsed_time": "0:23:37", "remaining_time": "4:10:41"}
+{"current_steps": 37, "total_steps": 418, "loss": 0.6465, "lr": 1.991995237604916e-06, "epoch": 0.1766109785202864, "percentage": 8.85, "elapsed_time": "0:24:15", "remaining_time": "4:09:47"}
+{"current_steps": 38, "total_steps": 418, "loss": 0.6581, "lr": 1.9909649301836674e-06, "epoch": 0.18138424821002386, "percentage": 9.09, "elapsed_time": "0:24:54", "remaining_time": "4:09:02"}
+{"current_steps": 39, "total_steps": 418, "loss": 0.6623, "lr": 1.9898725680523566e-06, "epoch": 0.18615751789976134, "percentage": 9.33, "elapsed_time": "0:25:30", "remaining_time": "4:07:57"}
+{"current_steps": 40, "total_steps": 418, "loss": 0.6527, "lr": 1.9887182196152367e-06, "epoch": 0.1909307875894988, "percentage": 9.57, "elapsed_time": "0:26:08", "remaining_time": "4:07:05"}
+{"current_steps": 41, "total_steps": 418, "loss": 0.6148, "lr": 1.9875019571581726e-06, "epoch": 0.1957040572792363, "percentage": 9.81, "elapsed_time": "0:26:47", "remaining_time": "4:06:22"}
+{"current_steps": 42, "total_steps": 418, "loss": 0.6288, "lr": 1.9862238568441165e-06, "epoch": 0.20047732696897375, "percentage": 10.05, "elapsed_time": "0:27:27", "remaining_time": "4:05:49"}
+{"current_steps": 43, "total_steps": 418, "loss": 0.6373, "lr": 1.9848839987083364e-06, "epoch": 0.2052505966587112, "percentage": 10.29, "elapsed_time": "0:28:05", "remaining_time": "4:04:55"}
+{"current_steps": 44, "total_steps": 418, "loss": 0.6401, "lr": 1.983482466653407e-06, "epoch": 0.2100238663484487, "percentage": 10.53, "elapsed_time": "0:28:43", "remaining_time": "4:04:10"}
+{"current_steps": 45, "total_steps": 418, "loss": 0.6274, "lr": 1.982019348443952e-06, "epoch": 0.21479713603818615, "percentage": 10.77, "elapsed_time": "0:29:24", "remaining_time": "4:03:46"}
+{"current_steps": 46, "total_steps": 418, "loss": 0.6694, "lr": 1.9804947357011523e-06, "epoch": 0.21957040572792363, "percentage": 11.0, "elapsed_time": "0:30:02", "remaining_time": "4:03:00"}
+{"current_steps": 47, "total_steps": 418, "loss": 0.6236, "lr": 1.978908723897005e-06, "epoch": 0.2243436754176611, "percentage": 11.24, "elapsed_time": "0:30:45", "remaining_time": "4:02:49"}
+{"current_steps": 48, "total_steps": 418, "loss": 0.6408, "lr": 1.9772614123483485e-06, "epoch": 0.22911694510739858, "percentage": 11.48, "elapsed_time": "0:31:23", "remaining_time": "4:02:02"}
+{"current_steps": 49, "total_steps": 418, "loss": 0.5919, "lr": 1.9755529042106393e-06, "epoch": 0.23389021479713604, "percentage": 11.72, "elapsed_time": "0:32:01", "remaining_time": "4:01:08"}
+{"current_steps": 50, "total_steps": 418, "loss": 0.5969, "lr": 1.973783306471495e-06, "epoch": 0.2386634844868735, "percentage": 11.96, "elapsed_time": "0:32:42", "remaining_time": "4:00:45"}
+{"current_steps": 51, "total_steps": 418, "loss": 0.5973, "lr": 1.971952729943994e-06, "epoch": 0.24343675417661098, "percentage": 12.2, "elapsed_time": "0:33:25", "remaining_time": "4:00:28"}
+{"current_steps": 52, "total_steps": 418, "loss": 0.6098, "lr": 1.9700612892597372e-06, "epoch": 0.24821002386634844, "percentage": 12.44, "elapsed_time": "0:34:04", "remaining_time": "3:59:48"}
+{"current_steps": 53, "total_steps": 418, "loss": 0.5991, "lr": 1.9681091028616676e-06, "epoch": 0.2529832935560859, "percentage": 12.68, "elapsed_time": "0:34:43", "remaining_time": "3:59:10"}
+{"current_steps": 54, "total_steps": 418, "loss": 0.612, "lr": 1.966096292996655e-06, "epoch": 0.2577565632458234, "percentage": 12.92, "elapsed_time": "0:35:25", "remaining_time": "3:58:44"}
+{"current_steps": 55, "total_steps": 418, "loss": 0.5949, "lr": 1.9640229857078413e-06, "epoch": 0.26252983293556087, "percentage": 13.16, "elapsed_time": "0:36:03", "remaining_time": "3:57:59"}
+{"current_steps": 56, "total_steps": 418, "loss": 0.6103, "lr": 1.9618893108267454e-06, "epoch": 0.26730310262529833, "percentage": 13.4, "elapsed_time": "0:36:44", "remaining_time": "3:57:32"}
+{"current_steps": 57, "total_steps": 418, "loss": 0.5788, "lr": 1.9596954019651354e-06, "epoch": 0.2720763723150358, "percentage": 13.64, "elapsed_time": "0:37:23", "remaining_time": "3:56:51"}
+{"current_steps": 58, "total_steps": 418, "loss": 0.6069, "lr": 1.95744139650666e-06, "epoch": 0.27684964200477324, "percentage": 13.88, "elapsed_time": "0:38:04", "remaining_time": "3:56:17"}
+{"current_steps": 59, "total_steps": 418, "loss": 0.5987, "lr": 1.955127435598247e-06, "epoch": 0.28162291169451076, "percentage": 14.11, "elapsed_time": "0:38:44", "remaining_time": "3:55:44"}
+{"current_steps": 60, "total_steps": 418, "loss": 0.6112, "lr": 1.9527536641412637e-06, "epoch": 0.2863961813842482, "percentage": 14.35, "elapsed_time": "0:39:25", "remaining_time": "3:55:11"}
+{"current_steps": 61, "total_steps": 418, "loss": 0.5827, "lr": 1.950320230782443e-06, "epoch": 0.2911694510739857, "percentage": 14.59, "elapsed_time": "0:40:06", "remaining_time": "3:54:42"}
+{"current_steps": 62, "total_steps": 418, "loss": 0.596, "lr": 1.9478272879045763e-06, "epoch": 0.29594272076372313, "percentage": 14.83, "elapsed_time": "0:40:47", "remaining_time": "3:54:13"}
+{"current_steps": 63, "total_steps": 418, "loss": 0.6131, "lr": 1.9452749916169685e-06, "epoch": 0.30071599045346065, "percentage": 15.07, "elapsed_time": "0:41:25", "remaining_time": "3:53:27"}
+{"current_steps": 64, "total_steps": 418, "loss": 0.6014, "lr": 1.942663501745666e-06, "epoch": 0.3054892601431981, "percentage": 15.31, "elapsed_time": "0:42:03", "remaining_time": "3:52:39"}
+{"current_steps": 65, "total_steps": 418, "loss": 0.599, "lr": 1.939992981823445e-06, "epoch": 0.31026252983293556, "percentage": 15.55, "elapsed_time": "0:42:43", "remaining_time": "3:51:59"}
+{"current_steps": 66, "total_steps": 418, "loss": 0.5606, "lr": 1.9372635990795744e-06, "epoch": 0.315035799522673, "percentage": 15.79, "elapsed_time": "0:43:25", "remaining_time": "3:51:36"}
+{"current_steps": 67, "total_steps": 418, "loss": 0.5845, "lr": 1.934475524429339e-06, "epoch": 0.3198090692124105, "percentage": 16.03, "elapsed_time": "0:44:06", "remaining_time": "3:51:03"}
+{"current_steps": 68, "total_steps": 418, "loss": 0.5938, "lr": 1.9316289324633416e-06, "epoch": 0.324582338902148, "percentage": 16.27, "elapsed_time": "0:44:45", "remaining_time": "3:50:21"}
+{"current_steps": 69, "total_steps": 418, "loss": 0.5971, "lr": 1.928724001436568e-06, "epoch": 0.32935560859188545, "percentage": 16.51, "elapsed_time": "0:45:24", "remaining_time": "3:49:38"}
+{"current_steps": 70, "total_steps": 418, "loss": 0.5896, "lr": 1.925760913257224e-06, "epoch": 0.3341288782816229, "percentage": 16.75, "elapsed_time": "0:46:01", "remaining_time": "3:48:46"}
+{"current_steps": 71, "total_steps": 418, "loss": 0.5957, "lr": 1.922739853475345e-06, "epoch": 0.33890214797136037, "percentage": 16.99, "elapsed_time": "0:46:38", "remaining_time": "3:47:58"}
+{"current_steps": 72, "total_steps": 418, "loss": 0.5782, "lr": 1.919661011271176e-06, "epoch": 0.3436754176610978, "percentage": 17.22, "elapsed_time": "0:47:18", "remaining_time": "3:47:19"}
+{"current_steps": 73, "total_steps": 418, "loss": 0.5912, "lr": 1.916524579443327e-06, "epoch": 0.34844868735083534, "percentage": 17.46, "elapsed_time": "0:47:55", "remaining_time": "3:46:31"}
+{"current_steps": 74, "total_steps": 418, "loss": 0.568, "lr": 1.9133307543966972e-06, "epoch": 0.3532219570405728, "percentage": 17.7, "elapsed_time": "0:48:39", "remaining_time": "3:46:13"}
+{"current_steps": 75, "total_steps": 418, "loss": 0.5831, "lr": 1.910079736130178e-06, "epoch": 0.35799522673031026, "percentage": 17.94, "elapsed_time": "0:49:17", "remaining_time": "3:45:23"}
+{"current_steps": 76, "total_steps": 418, "loss": 0.5752, "lr": 1.9067717282241275e-06, "epoch": 0.3627684964200477, "percentage": 18.18, "elapsed_time": "0:49:56", "remaining_time": "3:44:43"}
+{"current_steps": 77, "total_steps": 418, "loss": 0.6037, "lr": 1.9034069378276248e-06, "epoch": 0.36754176610978523, "percentage": 18.42, "elapsed_time": "0:50:33", "remaining_time": "3:43:53"}
+{"current_steps": 78, "total_steps": 418, "loss": 0.5814, "lr": 1.8999855756454943e-06, "epoch": 0.3723150357995227, "percentage": 18.66, "elapsed_time": "0:51:12", "remaining_time": "3:43:11"}
+{"current_steps": 79, "total_steps": 418, "loss": 0.5864, "lr": 1.8965078559251141e-06, "epoch": 0.37708830548926014, "percentage": 18.9, "elapsed_time": "0:51:50", "remaining_time": "3:42:26"}
+{"current_steps": 80, "total_steps": 418, "loss": 0.5872, "lr": 1.892973996443e-06, "epoch": 0.3818615751789976, "percentage": 19.14, "elapsed_time": "0:52:31", "remaining_time": "3:41:54"}
+{"current_steps": 81, "total_steps": 418, "loss": 0.5763, "lr": 1.8893842184911652e-06, "epoch": 0.38663484486873506, "percentage": 19.38, "elapsed_time": "0:53:11", "remaining_time": "3:41:19"}
+{"current_steps": 82, "total_steps": 418, "loss": 0.5663, "lr": 1.8857387468632673e-06, "epoch": 0.3914081145584726, "percentage": 19.62, "elapsed_time": "0:53:49", "remaining_time": "3:40:32"}
+{"current_steps": 83, "total_steps": 418, "loss": 0.5749, "lr": 1.8820378098405269e-06, "epoch": 0.39618138424821003, "percentage": 19.86, "elapsed_time": "0:54:28", "remaining_time": "3:39:51"}
+{"current_steps": 84, "total_steps": 418, "loss": 0.5791, "lr": 1.878281639177437e-06, "epoch": 0.4009546539379475, "percentage": 20.1, "elapsed_time": "0:55:05", "remaining_time": "3:39:03"}
+{"current_steps": 85, "total_steps": 418, "loss": 0.594, "lr": 1.874470470087246e-06, "epoch": 0.40572792362768495, "percentage": 20.33, "elapsed_time": "0:55:48", "remaining_time": "3:38:37"}
+{"current_steps": 86, "total_steps": 418, "loss": 0.5958, "lr": 1.8706045412272329e-06, "epoch": 0.4105011933174224, "percentage": 20.57, "elapsed_time": "0:56:28", "remaining_time": "3:37:59"}
+{"current_steps": 87, "total_steps": 418, "loss": 0.5831, "lr": 1.8666840946837588e-06, "epoch": 0.4152744630071599, "percentage": 20.81, "elapsed_time": "0:57:05", "remaining_time": "3:37:14"}
+{"current_steps": 88, "total_steps": 418, "loss": 0.5773, "lr": 1.8627093759571097e-06, "epoch": 0.4200477326968974, "percentage": 21.05, "elapsed_time": "0:57:42", "remaining_time": "3:36:25"}
+{"current_steps": 89, "total_steps": 418, "loss": 0.567, "lr": 1.8586806339461223e-06, "epoch": 0.42482100238663484, "percentage": 21.29, "elapsed_time": "0:58:19", "remaining_time": "3:35:36"}
+{"current_steps": 90, "total_steps": 418, "loss": 0.5859, "lr": 1.8545981209325974e-06, "epoch": 0.4295942720763723, "percentage": 21.53, "elapsed_time": "0:58:56", "remaining_time": "3:34:46"}
+{"current_steps": 91, "total_steps": 418, "loss": 0.5786, "lr": 1.850462092565503e-06, "epoch": 0.4343675417661098, "percentage": 21.77, "elapsed_time": "0:59:34", "remaining_time": "3:34:05"}
+{"current_steps": 92, "total_steps": 418, "loss": 0.5643, "lr": 1.846272807844964e-06, "epoch": 0.43914081145584727, "percentage": 22.01, "elapsed_time": "1:00:15", "remaining_time": "3:33:31"}
+{"current_steps": 93, "total_steps": 418, "loss": 0.5772, "lr": 1.8420305291060453e-06, "epoch": 0.4439140811455847, "percentage": 22.25, "elapsed_time": "1:00:57", "remaining_time": "3:33:02"}
+{"current_steps": 94, "total_steps": 418, "loss": 0.5973, "lr": 1.837735522002322e-06, "epoch": 0.4486873508353222, "percentage": 22.49, "elapsed_time": "1:01:35", "remaining_time": "3:32:18"}
+{"current_steps": 95, "total_steps": 418, "loss": 0.5683, "lr": 1.8333880554892465e-06, "epoch": 0.45346062052505964, "percentage": 22.73, "elapsed_time": "1:02:13", "remaining_time": "3:31:34"}
+{"current_steps": 96, "total_steps": 418, "loss": 0.5631, "lr": 1.828988401807304e-06, "epoch": 0.45823389021479716, "percentage": 22.97, "elapsed_time": "1:02:53", "remaining_time": "3:30:56"}
+{"current_steps": 97, "total_steps": 418, "loss": 0.5478, "lr": 1.8245368364649672e-06, "epoch": 0.4630071599045346, "percentage": 23.21, "elapsed_time": "1:03:30", "remaining_time": "3:30:11"}
+{"current_steps": 98, "total_steps": 418, "loss": 0.5814, "lr": 1.8200336382214404e-06, "epoch": 0.4677804295942721, "percentage": 23.44, "elapsed_time": "1:04:09", "remaining_time": "3:29:28"}
+{"current_steps": 99, "total_steps": 418, "loss": 0.5831, "lr": 1.815479089069208e-06, "epoch": 0.47255369928400953, "percentage": 23.68, "elapsed_time": "1:04:46", "remaining_time": "3:28:43"}
+{"current_steps": 100, "total_steps": 418, "loss": 0.5711, "lr": 1.8108734742163714e-06, "epoch": 0.477326968973747, "percentage": 23.92, "elapsed_time": "1:05:24", "remaining_time": "3:28:00"}
+{"current_steps": 101, "total_steps": 418, "loss": 0.5829, "lr": 1.8062170820687923e-06, "epoch": 0.4821002386634845, "percentage": 24.16, "elapsed_time": "1:10:53", "remaining_time": "3:42:29"}
+{"current_steps": 102, "total_steps": 418, "loss": 0.5651, "lr": 1.8015102042120314e-06, "epoch": 0.48687350835322196, "percentage": 24.4, "elapsed_time": "1:11:32", "remaining_time": "3:41:39"}
+{"current_steps": 103, "total_steps": 418, "loss": 0.578, "lr": 1.796753135393089e-06, "epoch": 0.4916467780429594, "percentage": 24.64, "elapsed_time": "1:12:12", "remaining_time": "3:40:49"}
+{"current_steps": 104, "total_steps": 418, "loss": 0.549, "lr": 1.791946173501948e-06, "epoch": 0.4964200477326969, "percentage": 24.88, "elapsed_time": "1:12:50", "remaining_time": "3:39:54"}
+{"current_steps": 105, "total_steps": 418, "loss": 0.5427, "lr": 1.7870896195529204e-06, "epoch": 0.5011933174224343, "percentage": 25.12, "elapsed_time": "1:13:29", "remaining_time": "3:39:05"}
+{"current_steps": 106, "total_steps": 418, "loss": 0.5765, "lr": 1.7821837776657967e-06, "epoch": 0.5059665871121718, "percentage": 25.36, "elapsed_time": "1:14:09", "remaining_time": "3:38:17"}
+{"current_steps": 107, "total_steps": 418, "loss": 0.5627, "lr": 1.777228955046803e-06, "epoch": 0.5107398568019093, "percentage": 25.6, "elapsed_time": "1:14:47", "remaining_time": "3:37:22"}
+{"current_steps": 108, "total_steps": 418, "loss": 0.5615, "lr": 1.7722254619693617e-06, "epoch": 0.5155131264916468, "percentage": 25.84, "elapsed_time": "1:15:31", "remaining_time": "3:36:46"}
+{"current_steps": 109, "total_steps": 418, "loss": 0.559, "lr": 1.7671736117546643e-06, "epoch": 0.5202863961813843, "percentage": 26.08, "elapsed_time": "1:16:09", "remaining_time": "3:35:54"}
+{"current_steps": 110, "total_steps": 418, "loss": 0.5675, "lr": 1.7620737207520498e-06, "epoch": 0.5250596658711217, "percentage": 26.32, "elapsed_time": "1:16:45", "remaining_time": "3:34:55"}
+{"current_steps": 111, "total_steps": 418, "loss": 0.564, "lr": 1.756926108319194e-06, "epoch": 0.5298329355608592, "percentage": 26.56, "elapsed_time": "1:17:28", "remaining_time": "3:34:17"}
+{"current_steps": 112, "total_steps": 418, "loss": 0.5697, "lr": 1.751731096802113e-06, "epoch": 0.5346062052505967, "percentage": 26.79, "elapsed_time": "1:18:05", "remaining_time": "3:33:21"}
+{"current_steps": 113, "total_steps": 418, "loss": 0.5556, "lr": 1.7464890115149759e-06, "epoch": 0.5393794749403341, "percentage": 27.03, "elapsed_time": "1:18:46", "remaining_time": "3:32:38"}
+{"current_steps": 114, "total_steps": 418, "loss": 0.5699, "lr": 1.7412001807197361e-06, "epoch": 0.5441527446300716, "percentage": 27.27, "elapsed_time": "1:19:24", "remaining_time": "3:31:45"}
+{"current_steps": 115, "total_steps": 418, "loss": 0.5535, "lr": 1.735864935605572e-06, "epoch": 0.548926014319809, "percentage": 27.51, "elapsed_time": "1:20:06", "remaining_time": "3:31:03"}
+{"current_steps": 116, "total_steps": 418, "loss": 0.5456, "lr": 1.7304836102681493e-06, "epoch": 0.5536992840095465, "percentage": 27.75, "elapsed_time": "1:20:44", "remaining_time": "3:30:12"}
+{"current_steps": 117, "total_steps": 418, "loss": 0.5724, "lr": 1.7250565416887015e-06, "epoch": 0.5584725536992841, "percentage": 27.99, "elapsed_time": "1:21:23", "remaining_time": "3:29:24"}
+{"current_steps": 118, "total_steps": 418, "loss": 0.568, "lr": 1.719584069712925e-06, "epoch": 0.5632458233890215, "percentage": 28.23, "elapsed_time": "1:22:02", "remaining_time": "3:28:35"}
+{"current_steps": 119, "total_steps": 418, "loss": 0.5501, "lr": 1.7140665370296992e-06, "epoch": 0.568019093078759, "percentage": 28.47, "elapsed_time": "1:22:43", "remaining_time": "3:27:51"}
+{"current_steps": 120, "total_steps": 418, "loss": 0.586, "lr": 1.708504289149628e-06, "epoch": 0.5727923627684964, "percentage": 28.71, "elapsed_time": "1:23:22", "remaining_time": "3:27:02"}
+{"current_steps": 121, "total_steps": 418, "loss": 0.5533, "lr": 1.702897674383402e-06, "epoch": 0.5775656324582339, "percentage": 28.95, "elapsed_time": "1:24:01", "remaining_time": "3:26:14"}
+{"current_steps": 122, "total_steps": 418, "loss": 0.5662, "lr": 1.697247043819988e-06, "epoch": 0.5823389021479713, "percentage": 29.19, "elapsed_time": "1:24:40", "remaining_time": "3:25:25"}
+{"current_steps": 123, "total_steps": 418, "loss": 0.5683, "lr": 1.6915527513046443e-06, "epoch": 0.5871121718377088, "percentage": 29.43, "elapsed_time": "1:25:17", "remaining_time": "3:24:33"}
+{"current_steps": 124, "total_steps": 418, "loss": 0.5621, "lr": 1.6858151534167616e-06, "epoch": 0.5918854415274463, "percentage": 29.67, "elapsed_time": "1:25:56", "remaining_time": "3:23:46"}
+{"current_steps": 125, "total_steps": 418, "loss": 0.5596, "lr": 1.6800346094475346e-06, "epoch": 0.5966587112171837, "percentage": 29.9, "elapsed_time": "1:26:36", "remaining_time": "3:23:00"}
+{"current_steps": 126, "total_steps": 418, "loss": 0.5531, "lr": 1.6742114813774618e-06, "epoch": 0.6014319809069213, "percentage": 30.14, "elapsed_time": "1:27:15", "remaining_time": "3:22:12"}
+{"current_steps": 127, "total_steps": 418, "loss": 0.5832, "lr": 1.6683461338536798e-06, "epoch": 0.6062052505966588, "percentage": 30.38, "elapsed_time": "1:27:52", "remaining_time": "3:21:21"}
+{"current_steps": 128, "total_steps": 418, "loss": 0.5541, "lr": 1.6624389341671278e-06, "epoch": 0.6109785202863962, "percentage": 30.62, "elapsed_time": "1:28:34", "remaining_time": "3:20:41"}
+{"current_steps": 129, "total_steps": 418, "loss": 0.5528, "lr": 1.656490252229548e-06, "epoch": 0.6157517899761337, "percentage": 30.86, "elapsed_time": "1:29:12", "remaining_time": "3:19:51"}
+{"current_steps": 130, "total_steps": 418, "loss": 0.5754, "lr": 1.6505004605503223e-06, "epoch": 0.6205250596658711, "percentage": 31.1, "elapsed_time": "1:29:55", "remaining_time": "3:19:12"}
+{"current_steps": 131, "total_steps": 418, "loss": 0.5659, "lr": 1.6444699342131428e-06, "epoch": 0.6252983293556086, "percentage": 31.34, "elapsed_time": "1:30:33", "remaining_time": "3:18:24"}
+{"current_steps": 132, "total_steps": 418, "loss": 0.5468, "lr": 1.638399050852528e-06, "epoch": 0.630071599045346, "percentage": 31.58, "elapsed_time": "1:31:14", "remaining_time": "3:17:41"}
+{"current_steps": 133, "total_steps": 418, "loss": 0.5547, "lr": 1.632288190630172e-06, "epoch": 0.6348448687350835, "percentage": 31.82, "elapsed_time": "1:31:52", "remaining_time": "3:16:51"}
+{"current_steps": 134, "total_steps": 418, "loss": 0.5475, "lr": 1.6261377362111396e-06, "epoch": 0.639618138424821, "percentage": 32.06, "elapsed_time": "1:32:30", "remaining_time": "3:16:03"}
+{"current_steps": 135, "total_steps": 418, "loss": 0.5622, "lr": 1.6199480727399032e-06, "epoch": 0.6443914081145584, "percentage": 32.3, "elapsed_time": "1:33:11", "remaining_time": "3:15:20"}
+{"current_steps": 136, "total_steps": 418, "loss": 0.5646, "lr": 1.6137195878162267e-06, "epoch": 0.649164677804296, "percentage": 32.54, "elapsed_time": "1:33:48", "remaining_time": "3:14:30"}
+{"current_steps": 137, "total_steps": 418, "loss": 0.5524, "lr": 1.607452671470891e-06, "epoch": 0.6539379474940334, "percentage": 32.78, "elapsed_time": "1:34:31", "remaining_time": "3:13:53"}
+{"current_steps": 138, "total_steps": 418, "loss": 0.5517, "lr": 1.601147716141272e-06, "epoch": 0.6587112171837709, "percentage": 33.01, "elapsed_time": "1:35:10", "remaining_time": "3:13:06"}
+{"current_steps": 139, "total_steps": 418, "loss": 0.5664, "lr": 1.5948051166467657e-06, "epoch": 0.6634844868735084, "percentage": 33.25, "elapsed_time": "1:35:49", "remaining_time": "3:12:19"}
+{"current_steps": 140, "total_steps": 418, "loss": 0.5611, "lr": 1.5884252701640634e-06, "epoch": 0.6682577565632458, "percentage": 33.49, "elapsed_time": "1:36:26", "remaining_time": "3:11:31"}
+{"current_steps": 141, "total_steps": 418, "loss": 0.5609, "lr": 1.5820085762022823e-06, "epoch": 0.6730310262529833, "percentage": 33.73, "elapsed_time": "1:37:05", "remaining_time": "3:10:43"}
+{"current_steps": 142, "total_steps": 418, "loss": 0.5586, "lr": 1.5755554365779455e-06, "epoch": 0.6778042959427207, "percentage": 33.97, "elapsed_time": "1:37:46", "remaining_time": "3:10:02"}
+{"current_steps": 143, "total_steps": 418, "loss": 0.5557, "lr": 1.5690662553898222e-06, "epoch": 0.6825775656324582, "percentage": 34.21, "elapsed_time": "1:38:24", "remaining_time": "3:09:15"}
+{"current_steps": 144, "total_steps": 418, "loss": 0.5379, "lr": 1.5625414389936218e-06, "epoch": 0.6873508353221957, "percentage": 34.45, "elapsed_time": "1:39:04", "remaining_time": "3:08:31"}
+{"current_steps": 145, "total_steps": 418, "loss": 0.5459, "lr": 1.555981395976548e-06, "epoch": 0.6921241050119332, "percentage": 34.69, "elapsed_time": "1:39:42", "remaining_time": "3:07:44"}
+{"current_steps": 146, "total_steps": 418, "loss": 0.5538, "lr": 1.5493865371317123e-06, "epoch": 0.6968973747016707, "percentage": 34.93, "elapsed_time": "1:40:23", "remaining_time": "3:07:01"}
+{"current_steps": 147, "total_steps": 418, "loss": 0.5511, "lr": 1.542757275432411e-06, "epoch": 0.7016706443914081, "percentage": 35.17, "elapsed_time": "1:41:03", "remaining_time": "3:06:17"}
+{"current_steps": 148, "total_steps": 418, "loss": 0.5395, "lr": 1.5360940260062635e-06, "epoch": 0.7064439140811456, "percentage": 35.41, "elapsed_time": "1:41:42", "remaining_time": "3:05:32"}
+{"current_steps": 149, "total_steps": 418, "loss": 0.5487, "lr": 1.5293972061092185e-06, "epoch": 0.711217183770883, "percentage": 35.65, "elapsed_time": "1:42:22", "remaining_time": "3:04:50"}
+{"current_steps": 150, "total_steps": 418, "loss": 0.5313, "lr": 1.522667235099422e-06, "epoch": 0.7159904534606205, "percentage": 35.89, "elapsed_time": "1:43:03", "remaining_time": "3:04:08"}
+{"current_steps": 151, "total_steps": 418, "loss": 0.548, "lr": 1.515904534410961e-06, "epoch": 0.720763723150358, "percentage": 36.12, "elapsed_time": "1:43:41", "remaining_time": "3:03:20"}
+{"current_steps": 152, "total_steps": 418, "loss": 0.5366, "lr": 1.5091095275274699e-06, "epoch": 0.7255369928400954, "percentage": 36.36, "elapsed_time": "1:44:19", "remaining_time": "3:02:33"}
+{"current_steps": 153, "total_steps": 418, "loss": 0.5365, "lr": 1.5022826399556133e-06, "epoch": 0.7303102625298329, "percentage": 36.6, "elapsed_time": "1:44:58", "remaining_time": "3:01:48"}
+{"current_steps": 154, "total_steps": 418, "loss": 0.5601, "lr": 1.4954242991984396e-06, "epoch": 0.7350835322195705, "percentage": 36.84, "elapsed_time": "1:45:39", "remaining_time": "3:01:07"}
+{"current_steps": 155, "total_steps": 418, "loss": 0.549, "lr": 1.4885349347286115e-06, "epoch": 0.7398568019093079, "percentage": 37.08, "elapsed_time": "1:46:16", "remaining_time": "3:00:18"}
+{"current_steps": 156, "total_steps": 418, "loss": 0.5516, "lr": 1.4816149779615126e-06, "epoch": 0.7446300715990454, "percentage": 37.32, "elapsed_time": "1:46:54", "remaining_time": "2:59:32"}
+{"current_steps": 157, "total_steps": 418, "loss": 0.5572, "lr": 1.474664862228229e-06, "epoch": 0.7494033412887828, "percentage": 37.56, "elapsed_time": "1:47:31", "remaining_time": "2:58:45"}
+{"current_steps": 158, "total_steps": 418, "loss": 0.5617, "lr": 1.467685022748419e-06, "epoch": 0.7541766109785203, "percentage": 37.8, "elapsed_time": "1:48:10", "remaining_time": "2:58:00"}
+{"current_steps": 159, "total_steps": 418, "loss": 0.5426, "lr": 1.4606758966030534e-06, "epoch": 0.7589498806682577, "percentage": 38.04, "elapsed_time": "1:48:52", "remaining_time": "2:57:21"}
+{"current_steps": 160, "total_steps": 418, "loss": 0.544, "lr": 1.4536379227070509e-06, "epoch": 0.7637231503579952, "percentage": 38.28, "elapsed_time": "1:49:37", "remaining_time": "2:56:45"}
+{"current_steps": 161, "total_steps": 418, "loss": 0.5435, "lr": 1.4465715417817888e-06, "epoch": 0.7684964200477327, "percentage": 38.52, "elapsed_time": "1:50:18", "remaining_time": "2:56:04"}
+{"current_steps": 162, "total_steps": 418, "loss": 0.5199, "lr": 1.4394771963275076e-06, "epoch": 0.7732696897374701, "percentage": 38.76, "elapsed_time": "1:50:56", "remaining_time": "2:55:19"}
+{"current_steps": 163, "total_steps": 418, "loss": 0.5479, "lr": 1.4323553305955997e-06, "epoch": 0.7780429594272077, "percentage": 39.0, "elapsed_time": "1:51:34", "remaining_time": "2:54:33"}
+{"current_steps": 164, "total_steps": 418, "loss": 0.5219, "lr": 1.4252063905607909e-06, "epoch": 0.7828162291169452, "percentage": 39.23, "elapsed_time": "1:52:12", "remaining_time": "2:53:46"}
+{"current_steps": 165, "total_steps": 418, "loss": 0.531, "lr": 1.4180308238932135e-06, "epoch": 0.7875894988066826, "percentage": 39.47, "elapsed_time": "1:52:52", "remaining_time": "2:53:04"}
+{"current_steps": 166, "total_steps": 418, "loss": 0.5481, "lr": 1.410829079930372e-06, "epoch": 0.7923627684964201, "percentage": 39.71, "elapsed_time": "1:53:31", "remaining_time": "2:52:20"}
+{"current_steps": 167, "total_steps": 418, "loss": 0.5478, "lr": 1.4036016096490064e-06, "epoch": 0.7971360381861575, "percentage": 39.95, "elapsed_time": "1:54:10", "remaining_time": "2:51:35"}
+{"current_steps": 168, "total_steps": 418, "loss": 0.535, "lr": 1.3963488656368517e-06, "epoch": 0.801909307875895, "percentage": 40.19, "elapsed_time": "1:54:48", "remaining_time": "2:50:50"}
+{"current_steps": 169, "total_steps": 418, "loss": 0.5277, "lr": 1.389071302064295e-06, "epoch": 0.8066825775656324, "percentage": 40.43, "elapsed_time": "1:55:27", "remaining_time": "2:50:07"}
+{"current_steps": 170, "total_steps": 418, "loss": 0.5553, "lr": 1.381769374655938e-06, "epoch": 0.8114558472553699, "percentage": 40.67, "elapsed_time": "1:56:05", "remaining_time": "2:49:22"}
+{"current_steps": 171, "total_steps": 418, "loss": 0.5518, "lr": 1.374443540662057e-06, "epoch": 0.8162291169451074, "percentage": 40.91, "elapsed_time": "1:56:46", "remaining_time": "2:48:40"}
+{"current_steps": 172, "total_steps": 418, "loss": 0.5294, "lr": 1.3670942588299705e-06, "epoch": 0.8210023866348448, "percentage": 41.15, "elapsed_time": "1:57:25", "remaining_time": "2:47:56"}
+{"current_steps": 173, "total_steps": 418, "loss": 0.5121, "lr": 1.3597219893753117e-06, "epoch": 0.8257756563245824, "percentage": 41.39, "elapsed_time": "1:58:00", "remaining_time": "2:47:07"}
+{"current_steps": 174, "total_steps": 418, "loss": 0.5259, "lr": 1.352327193953211e-06, "epoch": 0.8305489260143198, "percentage": 41.63, "elapsed_time": "1:58:41", "remaining_time": "2:46:26"}
+{"current_steps": 175, "total_steps": 418, "loss": 0.5601, "lr": 1.3449103356293852e-06, "epoch": 0.8353221957040573, "percentage": 41.87, "elapsed_time": "1:59:20", "remaining_time": "2:45:42"}
+{"current_steps": 176, "total_steps": 418, "loss": 0.5359, "lr": 1.337471878851141e-06, "epoch": 0.8400954653937948, "percentage": 42.11, "elapsed_time": "2:00:02", "remaining_time": "2:45:03"}
+{"current_steps": 177, "total_steps": 418, "loss": 0.5485, "lr": 1.3300122894182909e-06, "epoch": 0.8448687350835322, "percentage": 42.34, "elapsed_time": "2:00:42", "remaining_time": "2:44:21"}
+{"current_steps": 178, "total_steps": 418, "loss": 0.5562, "lr": 1.3225320344539842e-06, "epoch": 0.8496420047732697, "percentage": 42.58, "elapsed_time": "2:01:22", "remaining_time": "2:43:39"}
+{"current_steps": 179, "total_steps": 418, "loss": 0.5485, "lr": 1.315031582375457e-06, "epoch": 0.8544152744630071, "percentage": 42.82, "elapsed_time": "2:02:01", "remaining_time": "2:42:55"}
+{"current_steps": 180, "total_steps": 418, "loss": 0.5444, "lr": 1.3075114028646974e-06, "epoch": 0.8591885441527446, "percentage": 43.06, "elapsed_time": "2:02:41", "remaining_time": "2:42:12"}
+{"current_steps": 181, "total_steps": 418, "loss": 0.5481, "lr": 1.299971966839036e-06, "epoch": 0.863961813842482, "percentage": 43.3, "elapsed_time": "2:03:19", "remaining_time": "2:41:28"}
+{"current_steps": 182, "total_steps": 418, "loss": 0.5345, "lr": 1.292413746421655e-06, "epoch": 0.8687350835322196, "percentage": 43.54, "elapsed_time": "2:03:58", "remaining_time": "2:40:45"}
+{"current_steps": 183, "total_steps": 418, "loss": 0.512, "lr": 1.2848372149120246e-06, "epoch": 0.8735083532219571, "percentage": 43.78, "elapsed_time": "2:04:38", "remaining_time": "2:40:03"}
+{"current_steps": 184, "total_steps": 418, "loss": 0.55, "lr": 1.2772428467562651e-06, "epoch": 0.8782816229116945, "percentage": 44.02, "elapsed_time": "2:05:19", "remaining_time": "2:39:22"}
+{"current_steps": 185, "total_steps": 418, "loss": 0.5365, "lr": 1.2696311175174357e-06, "epoch": 0.883054892601432, "percentage": 44.26, "elapsed_time": "2:05:55", "remaining_time": "2:38:35"}
+{"current_steps": 186, "total_steps": 418, "loss": 0.5421, "lr": 1.2620025038457554e-06, "epoch": 0.8878281622911695, "percentage": 44.5, "elapsed_time": "2:06:34", "remaining_time": "2:37:52"}
+{"current_steps": 187, "total_steps": 418, "loss": 0.519, "lr": 1.254357483448755e-06, "epoch": 0.8926014319809069, "percentage": 44.74, "elapsed_time": "2:07:15", "remaining_time": "2:37:11"}
+{"current_steps": 188, "total_steps": 418, "loss": 0.5651, "lr": 1.2466965350613615e-06, "epoch": 0.8973747016706444, "percentage": 44.98, "elapsed_time": "2:07:56", "remaining_time": "2:36:30"}
+{"current_steps": 189, "total_steps": 418, "loss": 0.5603, "lr": 1.2390201384159219e-06, "epoch": 0.9021479713603818, "percentage": 45.22, "elapsed_time": "2:08:34", "remaining_time": "2:35:47"}
+{"current_steps": 190, "total_steps": 418, "loss": 0.5157, "lr": 1.231328774212159e-06, "epoch": 0.9069212410501193, "percentage": 45.45, "elapsed_time": "2:09:15", "remaining_time": "2:35:06"}
+{"current_steps": 191, "total_steps": 418, "loss": 0.5367, "lr": 1.223622924087073e-06, "epoch": 0.9116945107398569, "percentage": 45.69, "elapsed_time": "2:09:52", "remaining_time": "2:34:21"}
+{"current_steps": 192, "total_steps": 418, "loss": 0.5401, "lr": 1.215903070584779e-06, "epoch": 0.9164677804295943, "percentage": 45.93, "elapsed_time": "2:10:30", "remaining_time": "2:33:37"}
+{"current_steps": 193, "total_steps": 418, "loss": 0.5458, "lr": 1.2081696971262903e-06, "epoch": 0.9212410501193318, "percentage": 46.17, "elapsed_time": "2:11:09", "remaining_time": "2:32:54"}
+{"current_steps": 194, "total_steps": 418, "loss": 0.5398, "lr": 1.2004232879792464e-06, "epoch": 0.9260143198090692, "percentage": 46.41, "elapsed_time": "2:11:49", "remaining_time": "2:32:12"}
+{"current_steps": 195, "total_steps": 418, "loss": 0.5343, "lr": 1.1926643282275882e-06, "epoch": 0.9307875894988067, "percentage": 46.65, "elapsed_time": "2:12:29", "remaining_time": "2:31:30"}
+{"current_steps": 196, "total_steps": 418, "loss": 0.5181, "lr": 1.1848933037411825e-06, "epoch": 0.9355608591885441, "percentage": 46.89, "elapsed_time": "2:13:07", "remaining_time": "2:30:46"}
+{"current_steps": 197, "total_steps": 418, "loss": 0.5442, "lr": 1.1771107011453933e-06, "epoch": 0.9403341288782816, "percentage": 47.13, "elapsed_time": "2:13:47", "remaining_time": "2:30:05"}
+{"current_steps": 198, "total_steps": 418, "loss": 0.5467, "lr": 1.1693170077906143e-06, "epoch": 0.9451073985680191, "percentage": 47.37, "elapsed_time": "2:14:28", "remaining_time": "2:29:25"}
+{"current_steps": 199, "total_steps": 418, "loss": 0.5251, "lr": 1.1615127117217463e-06, "epoch": 0.9498806682577565, "percentage": 47.61, "elapsed_time": "2:15:07", "remaining_time": "2:28:42"}
+{"current_steps": 200, "total_steps": 418, "loss": 0.5368, "lr": 1.1536983016476373e-06, "epoch": 0.954653937947494, "percentage": 47.85, "elapsed_time": "2:15:44", "remaining_time": "2:27:57"}
+{"current_steps": 201, "total_steps": 418, "loss": 0.514, "lr": 1.1458742669104803e-06, "epoch": 0.9594272076372315, "percentage": 48.09, "elapsed_time": "2:21:09", "remaining_time": "2:32:23"}
+{"current_steps": 202, "total_steps": 418, "loss": 0.5327, "lr": 1.1380410974551682e-06, "epoch": 0.964200477326969, "percentage": 48.33, "elapsed_time": "2:21:50", "remaining_time": "2:31:40"}
+{"current_steps": 203, "total_steps": 418, "loss": 0.5152, "lr": 1.130199283798615e-06, "epoch": 0.9689737470167065, "percentage": 48.56, "elapsed_time": "2:22:27", "remaining_time": "2:30:53"}
+{"current_steps": 204, "total_steps": 418, "loss": 0.5376, "lr": 1.1223493169990391e-06, "epoch": 0.9737470167064439, "percentage": 48.8, "elapsed_time": "2:23:05", "remaining_time": "2:30:06"}
+{"current_steps": 205, "total_steps": 418, "loss": 0.5198, "lr": 1.1144916886252124e-06, "epoch": 0.9785202863961814, "percentage": 49.04, "elapsed_time": "2:23:47", "remaining_time": "2:29:24"}
+{"current_steps": 206, "total_steps": 418, "loss": 0.5358, "lr": 1.1066268907256782e-06, "epoch": 0.9832935560859188, "percentage": 49.28, "elapsed_time": "2:24:24", "remaining_time": "2:28:36"}
+{"current_steps": 207, "total_steps": 418, "loss": 0.5319, "lr": 1.098755415797939e-06, "epoch": 0.9880668257756563, "percentage": 49.52, "elapsed_time": "2:25:04", "remaining_time": "2:27:52"}
+{"current_steps": 208, "total_steps": 418, "loss": 0.5453, "lr": 1.0908777567576168e-06, "epoch": 0.9928400954653938, "percentage": 49.76, "elapsed_time": "2:25:41", "remaining_time": "2:27:06"}
+{"current_steps": 209, "total_steps": 418, "loss": 0.5398, "lr": 1.0829944069075847e-06, "epoch": 0.9976133651551312, "percentage": 50.0, "elapsed_time": "2:26:21", "remaining_time": "2:26:21"}
+{"current_steps": 210, "total_steps": 418, "loss": 0.2683, "lr": 1.0751058599070781e-06, "epoch": 1.0, "percentage": 50.24, "elapsed_time": "2:26:39", "remaining_time": "2:25:15"}
+{"current_steps": 211, "total_steps": 418, "loss": 0.4862, "lr": 1.0672126097407795e-06, "epoch": 1.0047732696897376, "percentage": 50.48, "elapsed_time": "2:27:19", "remaining_time": "2:24:31"}
+{"current_steps": 212, "total_steps": 418, "loss": 0.4886, "lr": 1.0593151506878865e-06, "epoch": 1.009546539379475, "percentage": 50.72, "elapsed_time": "2:27:55", "remaining_time": "2:23:43"}
+{"current_steps": 213, "total_steps": 418, "loss": 0.4755, "lr": 1.0514139772911597e-06, "epoch": 1.0143198090692125, "percentage": 50.96, "elapsed_time": "2:28:32", "remaining_time": "2:22:57"}
+{"current_steps": 214, "total_steps": 418, "loss": 0.4643, "lr": 1.043509584325953e-06, "epoch": 1.0190930787589498, "percentage": 51.2, "elapsed_time": "2:29:14", "remaining_time": "2:22:16"}
+{"current_steps": 215, "total_steps": 418, "loss": 0.4934, "lr": 1.0356024667692314e-06, "epoch": 1.0238663484486874, "percentage": 51.44, "elapsed_time": "2:29:53", "remaining_time": "2:21:31"}
+{"current_steps": 216, "total_steps": 418, "loss": 0.4976, "lr": 1.0276931197685753e-06, "epoch": 1.0286396181384247, "percentage": 51.67, "elapsed_time": "2:30:35", "remaining_time": "2:20:50"}
+{"current_steps": 217, "total_steps": 418, "loss": 0.4897, "lr": 1.0197820386111737e-06, "epoch": 1.0334128878281623, "percentage": 51.91, "elapsed_time": "2:31:16", "remaining_time": "2:20:07"}
+{"current_steps": 218, "total_steps": 418, "loss": 0.4632, "lr": 1.0118697186928105e-06, "epoch": 1.0381861575178997, "percentage": 52.15, "elapsed_time": "2:31:55", "remaining_time": "2:19:23"}
+{"current_steps": 219, "total_steps": 418, "loss": 0.4667, "lr": 1.0039566554868392e-06, "epoch": 1.0429594272076372, "percentage": 52.39, "elapsed_time": "2:32:36", "remaining_time": "2:18:40"}
+{"current_steps": 220, "total_steps": 418, "loss": 0.4913, "lr": 9.960433445131607e-07, "epoch": 1.0477326968973748, "percentage": 52.63, "elapsed_time": "2:33:16", "remaining_time": "2:17:56"}
+{"current_steps": 221, "total_steps": 418, "loss": 0.485, "lr": 9.881302813071896e-07, "epoch": 1.0525059665871122, "percentage": 52.87, "elapsed_time": "2:33:55", "remaining_time": "2:17:12"}
+{"current_steps": 222, "total_steps": 418, "loss": 0.4663, "lr": 9.802179613888262e-07, "epoch": 1.0572792362768497, "percentage": 53.11, "elapsed_time": "2:34:34", "remaining_time": "2:16:28"}
+{"current_steps": 223, "total_steps": 418, "loss": 0.4724, "lr": 9.723068802314246e-07, "epoch": 1.062052505966587, "percentage": 53.35, "elapsed_time": "2:35:11", "remaining_time": "2:15:42"}
+{"current_steps": 224, "total_steps": 418, "loss": 0.4777, "lr": 9.643975332307687e-07, "epoch": 1.0668257756563246, "percentage": 53.59, "elapsed_time": "2:35:50", "remaining_time": "2:14:57"}
+{"current_steps": 225, "total_steps": 418, "loss": 0.4664, "lr": 9.564904156740471e-07, "epoch": 1.071599045346062, "percentage": 53.83, "elapsed_time": "2:36:27", "remaining_time": "2:14:12"}
+{"current_steps": 226, "total_steps": 418, "loss": 0.4808, "lr": 9.485860227088405e-07, "epoch": 1.0763723150357996, "percentage": 54.07, "elapsed_time": "2:37:04", "remaining_time": "2:13:26"}
+{"current_steps": 227, "total_steps": 418, "loss": 0.4764, "lr": 9.406848493121134e-07, "epoch": 1.081145584725537, "percentage": 54.31, "elapsed_time": "2:37:47", "remaining_time": "2:12:45"}
+{"current_steps": 228, "total_steps": 418, "loss": 0.4711, "lr": 9.327873902592205e-07, "epoch": 1.0859188544152745, "percentage": 54.55, "elapsed_time": "2:38:25", "remaining_time": "2:12:01"}
+{"current_steps": 229, "total_steps": 418, "loss": 0.4753, "lr": 9.248941400929222e-07, "epoch": 1.0906921241050118, "percentage": 54.78, "elapsed_time": "2:39:04", "remaining_time": "2:11:17"}
+{"current_steps": 230, "total_steps": 418, "loss": 0.49, "lr": 9.17005593092415e-07, "epoch": 1.0954653937947494, "percentage": 55.02, "elapsed_time": "2:39:42", "remaining_time": "2:10:32"}
+{"current_steps": 231, "total_steps": 418, "loss": 0.4636, "lr": 9.09122243242383e-07, "epoch": 1.100238663484487, "percentage": 55.26, "elapsed_time": "2:40:25", "remaining_time": "2:09:51"}
+{"current_steps": 232, "total_steps": 418, "loss": 0.4638, "lr": 9.01244584202061e-07, "epoch": 1.1050119331742243, "percentage": 55.5, "elapsed_time": "2:41:08", "remaining_time": "2:09:11"}
+{"current_steps": 233, "total_steps": 418, "loss": 0.4951, "lr": 8.933731092743219e-07, "epoch": 1.1097852028639619, "percentage": 55.74, "elapsed_time": "2:41:49", "remaining_time": "2:08:29"}
+{"current_steps": 234, "total_steps": 418, "loss": 0.4715, "lr": 8.855083113747875e-07, "epoch": 1.1145584725536992, "percentage": 55.98, "elapsed_time": "2:42:28", "remaining_time": "2:07:45"}
+{"current_steps": 235, "total_steps": 418, "loss": 0.4792, "lr": 8.776506830009607e-07, "epoch": 1.1193317422434368, "percentage": 56.22, "elapsed_time": "2:43:07", "remaining_time": "2:07:01"}
+{"current_steps": 236, "total_steps": 418, "loss": 0.4734, "lr": 8.698007162013849e-07, "epoch": 1.1241050119331741, "percentage": 56.46, "elapsed_time": "2:43:46", "remaining_time": "2:06:18"}
+{"current_steps": 237, "total_steps": 418, "loss": 0.4899, "lr": 8.619589025448318e-07, "epoch": 1.1288782816229117, "percentage": 56.7, "elapsed_time": "2:44:25", "remaining_time": "2:05:34"}
+{"current_steps": 238, "total_steps": 418, "loss": 0.461, "lr": 8.541257330895197e-07, "epoch": 1.1336515513126493, "percentage": 56.94, "elapsed_time": "2:45:05", "remaining_time": "2:04:51"}
+{"current_steps": 239, "total_steps": 418, "loss": 0.4789, "lr": 8.463016983523627e-07, "epoch": 1.1384248210023866, "percentage": 57.18, "elapsed_time": "2:45:45", "remaining_time": "2:04:08"}
+{"current_steps": 240, "total_steps": 418, "loss": 0.4951, "lr": 8.384872882782541e-07, "epoch": 1.1431980906921242, "percentage": 57.42, "elapsed_time": "2:46:24", "remaining_time": "2:03:25"}
+{"current_steps": 241, "total_steps": 418, "loss": 0.4666, "lr": 8.306829922093857e-07, "epoch": 1.1479713603818615, "percentage": 57.66, "elapsed_time": "2:47:03", "remaining_time": "2:02:41"}
+{"current_steps": 242, "total_steps": 418, "loss": 0.475, "lr": 8.228892988546067e-07, "epoch": 1.152744630071599, "percentage": 57.89, "elapsed_time": "2:47:43", "remaining_time": "2:01:59"}
+{"current_steps": 243, "total_steps": 418, "loss": 0.4727, "lr": 8.15106696258818e-07, "epoch": 1.1575178997613365, "percentage": 58.13, "elapsed_time": "2:48:24", "remaining_time": "2:01:17"}
+{"current_steps": 244, "total_steps": 418, "loss": 0.4779, "lr": 8.073356717724115e-07, "epoch": 1.162291169451074, "percentage": 58.37, "elapsed_time": "2:49:04", "remaining_time": "2:00:34"}
+{"current_steps": 245, "total_steps": 418, "loss": 0.4647, "lr": 7.995767120207536e-07, "epoch": 1.1670644391408114, "percentage": 58.61, "elapsed_time": "2:49:46", "remaining_time": "1:59:52"}
+{"current_steps": 246, "total_steps": 418, "loss": 0.4712, "lr": 7.918303028737096e-07, "epoch": 1.171837708830549, "percentage": 58.85, "elapsed_time": "2:50:24", "remaining_time": "1:59:08"}
+{"current_steps": 247, "total_steps": 418, "loss": 0.4747, "lr": 7.840969294152211e-07, "epoch": 1.1766109785202863, "percentage": 59.09, "elapsed_time": "2:51:02", "remaining_time": "1:58:24"}
+{"current_steps": 248, "total_steps": 418, "loss": 0.4732, "lr": 7.763770759129269e-07, "epoch": 1.1813842482100239, "percentage": 59.33, "elapsed_time": "2:51:41", "remaining_time": "1:57:41"}
+{"current_steps": 249, "total_steps": 418, "loss": 0.4677, "lr": 7.68671225787841e-07, "epoch": 1.1861575178997614, "percentage": 59.57, "elapsed_time": "2:52:23", "remaining_time": "1:56:59"}
+{"current_steps": 250, "total_steps": 418, "loss": 0.4788, "lr": 7.609798615840785e-07, "epoch": 1.1909307875894988, "percentage": 59.81, "elapsed_time": "2:53:02", "remaining_time": "1:56:17"}
+{"current_steps": 251, "total_steps": 418, "loss": 0.456, "lr": 7.533034649386384e-07, "epoch": 1.1957040572792363, "percentage": 60.05, "elapsed_time": "2:53:43", "remaining_time": "1:55:35"}
+{"current_steps": 252, "total_steps": 418, "loss": 0.4768, "lr": 7.456425165512452e-07, "epoch": 1.2004773269689737, "percentage": 60.29, "elapsed_time": "2:54:20", "remaining_time": "1:54:50"}
+{"current_steps": 253, "total_steps": 418, "loss": 0.4864, "lr": 7.379974961542447e-07, "epoch": 1.2052505966587113, "percentage": 60.53, "elapsed_time": "2:55:00", "remaining_time": "1:54:07"}
+{"current_steps": 254, "total_steps": 418, "loss": 0.4768, "lr": 7.303688824825646e-07, "epoch": 1.2100238663484486, "percentage": 60.77, "elapsed_time": "2:55:38", "remaining_time": "1:53:24"}
+{"current_steps": 255, "total_steps": 418, "loss": 0.4676, "lr": 7.227571532437349e-07, "epoch": 1.2147971360381862, "percentage": 61.0, "elapsed_time": "2:56:15", "remaining_time": "1:52:40"}
+{"current_steps": 256, "total_steps": 418, "loss": 0.4688, "lr": 7.151627850879755e-07, "epoch": 1.2195704057279237, "percentage": 61.24, "elapsed_time": "2:56:56", "remaining_time": "1:51:58"}
+{"current_steps": 257, "total_steps": 418, "loss": 0.4545, "lr": 7.075862535783453e-07, "epoch": 1.224343675417661, "percentage": 61.48, "elapsed_time": "2:57:36", "remaining_time": "1:51:15"}
+{"current_steps": 258, "total_steps": 418, "loss": 0.4842, "lr": 7.00028033160964e-07, "epoch": 1.2291169451073987, "percentage": 61.72, "elapsed_time": "2:58:15", "remaining_time": "1:50:32"}
+{"current_steps": 259, "total_steps": 418, "loss": 0.4841, "lr": 6.924885971353026e-07, "epoch": 1.233890214797136, "percentage": 61.96, "elapsed_time": "2:58:55", "remaining_time": "1:49:50"}
+{"current_steps": 260, "total_steps": 418, "loss": 0.4485, "lr": 6.849684176245431e-07, "epoch": 1.2386634844868736, "percentage": 62.2, "elapsed_time": "2:59:35", "remaining_time": "1:49:08"}
+{"current_steps": 261, "total_steps": 418, "loss": 0.4632, "lr": 6.774679655460158e-07, "epoch": 1.243436754176611, "percentage": 62.44, "elapsed_time": "3:00:15", "remaining_time": "1:48:25"}
+{"current_steps": 262, "total_steps": 418, "loss": 0.4701, "lr": 6.699877105817092e-07, "epoch": 1.2482100238663485, "percentage": 62.68, "elapsed_time": "3:00:54", "remaining_time": "1:47:42"}
+{"current_steps": 263, "total_steps": 418, "loss": 0.4884, "lr": 6.625281211488591e-07, "epoch": 1.2529832935560858, "percentage": 62.92, "elapsed_time": "3:01:33", "remaining_time": "1:47:00"}
+{"current_steps": 264, "total_steps": 418, "loss": 0.4821, "lr": 6.55089664370615e-07, "epoch": 1.2577565632458234, "percentage": 63.16, "elapsed_time": "3:02:10", "remaining_time": "1:46:16"}
+{"current_steps": 265, "total_steps": 418, "loss": 0.4585, "lr": 6.476728060467888e-07, "epoch": 1.2625298329355608, "percentage": 63.4, "elapsed_time": "3:02:48", "remaining_time": "1:45:32"}
+{"current_steps": 266, "total_steps": 418, "loss": 0.4688, "lr": 6.402780106246884e-07, "epoch": 1.2673031026252983, "percentage": 63.64, "elapsed_time": "3:03:30", "remaining_time": "1:44:51"}
+{"current_steps": 267, "total_steps": 418, "loss": 0.4813, "lr": 6.329057411700298e-07, "epoch": 1.272076372315036, "percentage": 63.88, "elapsed_time": "3:04:11", "remaining_time": "1:44:10"}
+{"current_steps": 268, "total_steps": 418, "loss": 0.4878, "lr": 6.255564593379429e-07, "epoch": 1.2768496420047732, "percentage": 64.11, "elapsed_time": "3:04:51", "remaining_time": "1:43:27"}
+{"current_steps": 269, "total_steps": 418, "loss": 0.4629, "lr": 6.182306253440619e-07, "epoch": 1.2816229116945108, "percentage": 64.35, "elapsed_time": "3:05:28", "remaining_time": "1:42:44"}
+{"current_steps": 270, "total_steps": 418, "loss": 0.4842, "lr": 6.109286979357051e-07, "epoch": 1.2863961813842482, "percentage": 64.59, "elapsed_time": "3:06:06", "remaining_time": "1:42:00"}
+{"current_steps": 271, "total_steps": 418, "loss": 0.4588, "lr": 6.036511343631488e-07, "epoch": 1.2911694510739857, "percentage": 64.83, "elapsed_time": "3:06:45", "remaining_time": "1:41:18"}
+{"current_steps": 272, "total_steps": 418, "loss": 0.4817, "lr": 5.963983903509935e-07, "epoch": 1.295942720763723, "percentage": 65.07, "elapsed_time": "3:07:24", "remaining_time": "1:40:35"}
+{"current_steps": 273, "total_steps": 418, "loss": 0.4781, "lr": 5.89170920069628e-07, "epoch": 1.3007159904534606, "percentage": 65.31, "elapsed_time": "3:08:04", "remaining_time": "1:39:53"}
+{"current_steps": 274, "total_steps": 418, "loss": 0.46, "lr": 5.819691761067865e-07, "epoch": 1.3054892601431982, "percentage": 65.55, "elapsed_time": "3:08:41", "remaining_time": "1:39:09"}
+{"current_steps": 275, "total_steps": 418, "loss": 0.4647, "lr": 5.747936094392089e-07, "epoch": 1.3102625298329356, "percentage": 65.79, "elapsed_time": "3:09:21", "remaining_time": "1:38:27"}
+{"current_steps": 276, "total_steps": 418, "loss": 0.4639, "lr": 5.676446694044002e-07, "epoch": 1.315035799522673, "percentage": 66.03, "elapsed_time": "3:09:59", "remaining_time": "1:37:44"}
+{"current_steps": 277, "total_steps": 418, "loss": 0.4652, "lr": 5.605228036724927e-07, "epoch": 1.3198090692124105, "percentage": 66.27, "elapsed_time": "3:10:39", "remaining_time": "1:37:02"}
+{"current_steps": 278, "total_steps": 418, "loss": 0.4717, "lr": 5.534284582182114e-07, "epoch": 1.324582338902148, "percentage": 66.51, "elapsed_time": "3:11:20", "remaining_time": "1:36:21"}
+{"current_steps": 279, "total_steps": 418, "loss": 0.4536, "lr": 5.463620772929494e-07, "epoch": 1.3293556085918854, "percentage": 66.75, "elapsed_time": "3:12:01", "remaining_time": "1:35:39"}
+{"current_steps": 280, "total_steps": 418, "loss": 0.4649, "lr": 5.393241033969466e-07, "epoch": 1.334128878281623, "percentage": 66.99, "elapsed_time": "3:12:44", "remaining_time": "1:34:59"}
+{"current_steps": 281, "total_steps": 418, "loss": 0.4668, "lr": 5.323149772515812e-07, "epoch": 1.3389021479713603, "percentage": 67.22, "elapsed_time": "3:13:25", "remaining_time": "1:34:18"}
+{"current_steps": 282, "total_steps": 418, "loss": 0.4761, "lr": 5.253351377717706e-07, "epoch": 1.3436754176610979, "percentage": 67.46, "elapsed_time": "3:14:02", "remaining_time": "1:33:34"}
+{"current_steps": 283, "total_steps": 418, "loss": 0.469, "lr": 5.183850220384873e-07, "epoch": 1.3484486873508352, "percentage": 67.7, "elapsed_time": "3:14:43", "remaining_time": "1:32:53"}
+{"current_steps": 284, "total_steps": 418, "loss": 0.4802, "lr": 5.114650652713884e-07, "epoch": 1.3532219570405728, "percentage": 67.94, "elapsed_time": "3:15:23", "remaining_time": "1:32:11"}
+{"current_steps": 285, "total_steps": 418, "loss": 0.4773, "lr": 5.045757008015606e-07, "epoch": 1.3579952267303104, "percentage": 68.18, "elapsed_time": "3:16:04", "remaining_time": "1:31:30"}
+{"current_steps": 286, "total_steps": 418, "loss": 0.4605, "lr": 4.977173600443868e-07, "epoch": 1.3627684964200477, "percentage": 68.42, "elapsed_time": "3:16:43", "remaining_time": "1:30:47"}
+{"current_steps": 287, "total_steps": 418, "loss": 0.4767, "lr": 4.908904724725299e-07, "epoch": 1.3675417661097853, "percentage": 68.66, "elapsed_time": "3:17:21", "remaining_time": "1:30:05"}
+{"current_steps": 288, "total_steps": 418, "loss": 0.4682, "lr": 4.840954655890391e-07, "epoch": 1.3723150357995226, "percentage": 68.9, "elapsed_time": "3:18:00", "remaining_time": "1:29:22"}
+{"current_steps": 289, "total_steps": 418, "loss": 0.4855, "lr": 4.773327649005777e-07, "epoch": 1.3770883054892602, "percentage": 69.14, "elapsed_time": "3:18:41", "remaining_time": "1:28:41"}
+{"current_steps": 290, "total_steps": 418, "loss": 0.4761, "lr": 4.7060279389078184e-07, "epoch": 1.3818615751789975, "percentage": 69.38, "elapsed_time": "3:19:21", "remaining_time": "1:27:59"}
+{"current_steps": 291, "total_steps": 418, "loss": 0.4565, "lr": 4.6390597399373644e-07, "epoch": 1.3866348448687351, "percentage": 69.62, "elapsed_time": "3:20:02", "remaining_time": "1:27:18"}
+{"current_steps": 292, "total_steps": 418, "loss": 0.4826, "lr": 4.5724272456758907e-07, "epoch": 1.3914081145584727, "percentage": 69.86, "elapsed_time": "3:20:41", "remaining_time": "1:26:35"}
+{"current_steps": 293, "total_steps": 418, "loss": 0.4763, "lr": 4.506134628682877e-07, "epoch": 1.39618138424821, "percentage": 70.1, "elapsed_time": "3:21:18", "remaining_time": "1:25:52"}
+{"current_steps": 294, "total_steps": 418, "loss": 0.4672, "lr": 4.440186040234524e-07, "epoch": 1.4009546539379474, "percentage": 70.33, "elapsed_time": "3:21:57", "remaining_time": "1:25:10"}
+{"current_steps": 295, "total_steps": 418, "loss": 0.4656, "lr": 4.3745856100637834e-07, "epoch": 1.405727923627685, "percentage": 70.57, "elapsed_time": "3:22:35", "remaining_time": "1:24:28"}
+{"current_steps": 296, "total_steps": 418, "loss": 0.4676, "lr": 4.3093374461017785e-07, "epoch": 1.4105011933174225, "percentage": 70.81, "elapsed_time": "3:23:19", "remaining_time": "1:23:48"}
+{"current_steps": 297, "total_steps": 418, "loss": 0.4843, "lr": 4.244445634220545e-07, "epoch": 1.4152744630071599, "percentage": 71.05, "elapsed_time": "3:24:02", "remaining_time": "1:23:07"}
+{"current_steps": 298, "total_steps": 418, "loss": 0.4809, "lr": 4.1799142379771766e-07, "epoch": 1.4200477326968974, "percentage": 71.29, "elapsed_time": "3:24:41", "remaining_time": "1:22:25"}
+{"current_steps": 299, "total_steps": 418, "loss": 0.464, "lr": 4.115747298359363e-07, "epoch": 1.4248210023866348, "percentage": 71.53, "elapsed_time": "3:25:20", "remaining_time": "1:21:43"}
+{"current_steps": 300, "total_steps": 418, "loss": 0.4851, "lr": 4.0519488335323415e-07, "epoch": 1.4295942720763724, "percentage": 71.77, "elapsed_time": "3:25:59", "remaining_time": "1:21:01"}
+{"current_steps": 301, "total_steps": 418, "loss": 0.4594, "lr": 3.9885228385872806e-07, "epoch": 1.4343675417661097, "percentage": 72.01, "elapsed_time": "3:31:06", "remaining_time": "1:22:03"}
+{"current_steps": 302, "total_steps": 418, "loss": 0.4661, "lr": 3.925473285291091e-07, "epoch": 1.4391408114558473, "percentage": 72.25, "elapsed_time": "3:31:43", "remaining_time": "1:21:19"}
+{"current_steps": 303, "total_steps": 418, "loss": 0.4757, "lr": 3.862804121837733e-07, "epoch": 1.4439140811455848, "percentage": 72.49, "elapsed_time": "3:32:22", "remaining_time": "1:20:36"}
+{"current_steps": 304, "total_steps": 418, "loss": 0.4787, "lr": 3.8005192726009663e-07, "epoch": 1.4486873508353222, "percentage": 72.73, "elapsed_time": "3:32:59", "remaining_time": "1:19:52"}
+{"current_steps": 305, "total_steps": 418, "loss": 0.4668, "lr": 3.738622637888608e-07, "epoch": 1.4534606205250595, "percentage": 72.97, "elapsed_time": "3:33:36", "remaining_time": "1:19:08"}
+{"current_steps": 306, "total_steps": 418, "loss": 0.4606, "lr": 3.677118093698278e-07, "epoch": 1.458233890214797, "percentage": 73.21, "elapsed_time": "3:34:19", "remaining_time": "1:18:26"}
+{"current_steps": 307, "total_steps": 418, "loss": 0.4683, "lr": 3.61600949147472e-07, "epoch": 1.4630071599045347, "percentage": 73.44, "elapsed_time": "3:34:58", "remaining_time": "1:17:43"}
+{"current_steps": 308, "total_steps": 418, "loss": 0.4519, "lr": 3.5553006578685706e-07, "epoch": 1.467780429594272, "percentage": 73.68, "elapsed_time": "3:35:41", "remaining_time": "1:17:02"}
+{"current_steps": 309, "total_steps": 418, "loss": 0.469, "lr": 3.494995394496778e-07, "epoch": 1.4725536992840096, "percentage": 73.92, "elapsed_time": "3:36:22", "remaining_time": "1:16:19"}
+{"current_steps": 310, "total_steps": 418, "loss": 0.449, "lr": 3.435097477704517e-07, "epoch": 1.477326968973747, "percentage": 74.16, "elapsed_time": "3:37:00", "remaining_time": "1:15:36"}
+{"current_steps": 311, "total_steps": 418, "loss": 0.4745, "lr": 3.3756106583287205e-07, "epoch": 1.4821002386634845, "percentage": 74.4, "elapsed_time": "3:37:40", "remaining_time": "1:14:53"}
+{"current_steps": 312, "total_steps": 418, "loss": 0.4918, "lr": 3.316538661463204e-07, "epoch": 1.4868735083532219, "percentage": 74.64, "elapsed_time": "3:38:19", "remaining_time": "1:14:10"}
+{"current_steps": 313, "total_steps": 418, "loss": 0.4846, "lr": 3.2578851862253796e-07, "epoch": 1.4916467780429594, "percentage": 74.88, "elapsed_time": "3:38:58", "remaining_time": "1:13:27"}
+{"current_steps": 314, "total_steps": 418, "loss": 0.4604, "lr": 3.199653905524654e-07, "epoch": 1.496420047732697, "percentage": 75.12, "elapsed_time": "3:39:37", "remaining_time": "1:12:44"}
+{"current_steps": 315, "total_steps": 418, "loss": 0.4772, "lr": 3.1418484658323806e-07, "epoch": 1.5011933174224343, "percentage": 75.36, "elapsed_time": "3:40:16", "remaining_time": "1:12:01"}
+{"current_steps": 316, "total_steps": 418, "loss": 0.468, "lr": 3.0844724869535577e-07, "epoch": 1.5059665871121717, "percentage": 75.6, "elapsed_time": "3:40:57", "remaining_time": "1:11:19"}
+{"current_steps": 317, "total_steps": 418, "loss": 0.4808, "lr": 3.027529561800117e-07, "epoch": 1.5107398568019093, "percentage": 75.84, "elapsed_time": "3:41:38", "remaining_time": "1:10:36"}
+{"current_steps": 318, "total_steps": 418, "loss": 0.469, "lr": 2.971023256165983e-07, "epoch": 1.5155131264916468, "percentage": 76.08, "elapsed_time": "3:42:17", "remaining_time": "1:09:54"}
+{"current_steps": 319, "total_steps": 418, "loss": 0.4758, "lr": 2.9149571085037215e-07, "epoch": 1.5202863961813842, "percentage": 76.32, "elapsed_time": "3:42:59", "remaining_time": "1:09:12"}
+{"current_steps": 320, "total_steps": 418, "loss": 0.4662, "lr": 2.8593346297030073e-07, "epoch": 1.5250596658711217, "percentage": 76.56, "elapsed_time": "3:43:39", "remaining_time": "1:08:29"}
+{"current_steps": 321, "total_steps": 418, "loss": 0.4638, "lr": 2.804159302870751e-07, "epoch": 1.5298329355608593, "percentage": 76.79, "elapsed_time": "3:44:19", "remaining_time": "1:07:47"}
+{"current_steps": 322, "total_steps": 418, "loss": 0.4584, "lr": 2.7494345831129837e-07, "epoch": 1.5346062052505967, "percentage": 77.03, "elapsed_time": "3:44:59", "remaining_time": "1:07:04"}
+{"current_steps": 323, "total_steps": 418, "loss": 0.4757, "lr": 2.6951638973185073e-07, "epoch": 1.539379474940334, "percentage": 77.27, "elapsed_time": "3:45:37", "remaining_time": "1:06:21"}
+{"current_steps": 324, "total_steps": 418, "loss": 0.4807, "lr": 2.64135064394428e-07, "epoch": 1.5441527446300716, "percentage": 77.51, "elapsed_time": "3:46:17", "remaining_time": "1:05:39"}
+{"current_steps": 325, "total_steps": 418, "loss": 0.4605, "lr": 2.587998192802638e-07, "epoch": 1.5489260143198091, "percentage": 77.75, "elapsed_time": "3:46:55", "remaining_time": "1:04:56"}
+{"current_steps": 326, "total_steps": 418, "loss": 0.474, "lr": 2.5351098848502386e-07, "epoch": 1.5536992840095465, "percentage": 77.99, "elapsed_time": "3:47:37", "remaining_time": "1:04:14"}
+{"current_steps": 327, "total_steps": 418, "loss": 0.4715, "lr": 2.482689031978872e-07, "epoch": 1.558472553699284, "percentage": 78.23, "elapsed_time": "3:48:17", "remaining_time": "1:03:31"}
+{"current_steps": 328, "total_steps": 418, "loss": 0.4656, "lr": 2.4307389168080606e-07, "epoch": 1.5632458233890216, "percentage": 78.47, "elapsed_time": "3:48:58", "remaining_time": "1:02:49"}
+{"current_steps": 329, "total_steps": 418, "loss": 0.4922, "lr": 2.3792627924795038e-07, "epoch": 1.568019093078759, "percentage": 78.71, "elapsed_time": "3:49:39", "remaining_time": "1:02:07"}
+{"current_steps": 330, "total_steps": 418, "loss": 0.4692, "lr": 2.3282638824533529e-07, "epoch": 1.5727923627684963, "percentage": 78.95, "elapsed_time": "3:50:19", "remaining_time": "1:01:25"}
+{"current_steps": 331, "total_steps": 418, "loss": 0.4876, "lr": 2.277745380306383e-07, "epoch": 1.577565632458234, "percentage": 79.19, "elapsed_time": "3:51:01", "remaining_time": "1:00:43"}
+{"current_steps": 332, "total_steps": 418, "loss": 0.4918, "lr": 2.227710449531971e-07, "epoch": 1.5823389021479715, "percentage": 79.43, "elapsed_time": "3:51:42", "remaining_time": "1:00:01"}
+{"current_steps": 333, "total_steps": 418, "loss": 0.4641, "lr": 2.178162223342035e-07, "epoch": 1.5871121718377088, "percentage": 79.67, "elapsed_time": "3:52:20", "remaining_time": "0:59:18"}
+{"current_steps": 334, "total_steps": 418, "loss": 0.4661, "lr": 2.1291038044707965e-07, "epoch": 1.5918854415274462, "percentage": 79.9, "elapsed_time": "3:52:58", "remaining_time": "0:58:35"}
+{"current_steps": 335, "total_steps": 418, "loss": 0.4681, "lr": 2.0805382649805225e-07, "epoch": 1.5966587112171837, "percentage": 80.14, "elapsed_time": "3:53:38", "remaining_time": "0:57:53"}
+{"current_steps": 336, "total_steps": 418, "loss": 0.4672, "lr": 2.032468646069112e-07, "epoch": 1.6014319809069213, "percentage": 80.38, "elapsed_time": "3:54:21", "remaining_time": "0:57:11"}
+{"current_steps": 337, "total_steps": 418, "loss": 0.4767, "lr": 1.9848979578796865e-07, "epoch": 1.6062052505966586, "percentage": 80.62, "elapsed_time": "3:55:02", "remaining_time": "0:56:29"}
+{"current_steps": 338, "total_steps": 418, "loss": 0.4822, "lr": 1.937829179312076e-07, "epoch": 1.6109785202863962, "percentage": 80.86, "elapsed_time": "3:55:41", "remaining_time": "0:55:47"}
+{"current_steps": 339, "total_steps": 418, "loss": 0.4709, "lr": 1.8912652578362853e-07, "epoch": 1.6157517899761338, "percentage": 81.1, "elapsed_time": "3:56:18", "remaining_time": "0:55:04"}
+{"current_steps": 340, "total_steps": 418, "loss": 0.4604, "lr": 1.8452091093079215e-07, "epoch": 1.6205250596658711, "percentage": 81.34, "elapsed_time": "3:57:00", "remaining_time": "0:54:22"}
+{"current_steps": 341, "total_steps": 418, "loss": 0.4984, "lr": 1.7996636177855928e-07, "epoch": 1.6252983293556085, "percentage": 81.58, "elapsed_time": "3:57:40", "remaining_time": "0:53:40"}
+{"current_steps": 342, "total_steps": 418, "loss": 0.4823, "lr": 1.75463163535033e-07, "epoch": 1.630071599045346, "percentage": 81.82, "elapsed_time": "3:58:20", "remaining_time": "0:52:57"}
+{"current_steps": 343, "total_steps": 418, "loss": 0.4635, "lr": 1.7101159819269583e-07, "epoch": 1.6348448687350836, "percentage": 82.06, "elapsed_time": "3:59:01", "remaining_time": "0:52:15"}
+{"current_steps": 344, "total_steps": 418, "loss": 0.4765, "lr": 1.6661194451075345e-07, "epoch": 1.639618138424821, "percentage": 82.3, "elapsed_time": "3:59:40", "remaining_time": "0:51:33"}
+{"current_steps": 345, "total_steps": 418, "loss": 0.4533, "lr": 1.6226447799767772e-07, "epoch": 1.6443914081145583, "percentage": 82.54, "elapsed_time": "4:00:17", "remaining_time": "0:50:50"}
+{"current_steps": 346, "total_steps": 418, "loss": 0.4691, "lr": 1.5796947089395475e-07, "epoch": 1.649164677804296, "percentage": 82.78, "elapsed_time": "4:00:54", "remaining_time": "0:50:07"}
+{"current_steps": 347, "total_steps": 418, "loss": 0.4544, "lr": 1.5372719215503582e-07, "epoch": 1.6539379474940334, "percentage": 83.01, "elapsed_time": "4:01:33", "remaining_time": "0:49:25"}
+{"current_steps": 348, "total_steps": 418, "loss": 0.4806, "lr": 1.4953790743449702e-07, "epoch": 1.6587112171837708, "percentage": 83.25, "elapsed_time": "4:02:10", "remaining_time": "0:48:42"}
+{"current_steps": 349, "total_steps": 418, "loss": 0.4569, "lr": 1.4540187906740241e-07, "epoch": 1.6634844868735084, "percentage": 83.49, "elapsed_time": "4:02:49", "remaining_time": "0:48:00"}
+{"current_steps": 350, "total_steps": 418, "loss": 0.4897, "lr": 1.4131936605387762e-07, "epoch": 1.668257756563246, "percentage": 83.73, "elapsed_time": "4:03:28", "remaining_time": "0:47:18"}
+{"current_steps": 351, "total_steps": 418, "loss": 0.4799, "lr": 1.3729062404289017e-07, "epoch": 1.6730310262529833, "percentage": 83.97, "elapsed_time": "4:04:09", "remaining_time": "0:46:36"}
+{"current_steps": 352, "total_steps": 418, "loss": 0.4714, "lr": 1.3331590531624115e-07, "epoch": 1.6778042959427206, "percentage": 84.21, "elapsed_time": "4:04:52", "remaining_time": "0:45:54"}
+{"current_steps": 353, "total_steps": 418, "loss": 0.4679, "lr": 1.2939545877276726e-07, "epoch": 1.6825775656324582, "percentage": 84.45, "elapsed_time": "4:05:31", "remaining_time": "0:45:12"}
+{"current_steps": 354, "total_steps": 418, "loss": 0.4678, "lr": 1.25529529912754e-07, "epoch": 1.6873508353221958, "percentage": 84.69, "elapsed_time": "4:06:12", "remaining_time": "0:44:30"}
+{"current_steps": 355, "total_steps": 418, "loss": 0.4754, "lr": 1.2171836082256316e-07, "epoch": 1.692124105011933, "percentage": 84.93, "elapsed_time": "4:06:49", "remaining_time": "0:43:48"}
+{"current_steps": 356, "total_steps": 418, "loss": 0.4803, "lr": 1.1796219015947285e-07, "epoch": 1.6968973747016707, "percentage": 85.17, "elapsed_time": "4:07:32", "remaining_time": "0:43:06"}
+{"current_steps": 357, "total_steps": 418, "loss": 0.4939, "lr": 1.1426125313673285e-07, "epoch": 1.7016706443914082, "percentage": 85.41, "elapsed_time": "4:08:10", "remaining_time": "0:42:24"}
+{"current_steps": 358, "total_steps": 418, "loss": 0.4372, "lr": 1.1061578150883444e-07, "epoch": 1.7064439140811456, "percentage": 85.65, "elapsed_time": "4:08:48", "remaining_time": "0:41:41"}
+{"current_steps": 359, "total_steps": 418, "loss": 0.4792, "lr": 1.070260035570002e-07, "epoch": 1.711217183770883, "percentage": 85.89, "elapsed_time": "4:09:28", "remaining_time": "0:40:59"}
+{"current_steps": 360, "total_steps": 418, "loss": 0.4719, "lr": 1.0349214407488571e-07, "epoch": 1.7159904534606205, "percentage": 86.12, "elapsed_time": "4:10:08", "remaining_time": "0:40:17"}
+{"current_steps": 361, "total_steps": 418, "loss": 0.4724, "lr": 1.000144243545058e-07, "epoch": 1.720763723150358, "percentage": 86.36, "elapsed_time": "4:10:46", "remaining_time": "0:39:35"}
+{"current_steps": 362, "total_steps": 418, "loss": 0.4717, "lr": 9.659306217237517e-08, "epoch": 1.7255369928400954, "percentage": 86.6, "elapsed_time": "4:11:28", "remaining_time": "0:38:54"}
+{"current_steps": 363, "total_steps": 418, "loss": 0.4623, "lr": 9.322827177587212e-08, "epoch": 1.7303102625298328, "percentage": 86.84, "elapsed_time": "4:12:05", "remaining_time": "0:38:11"}
+{"current_steps": 364, "total_steps": 418, "loss": 0.4735, "lr": 8.992026386982221e-08, "epoch": 1.7350835322195706, "percentage": 87.08, "elapsed_time": "4:12:45", "remaining_time": "0:37:29"}
+{"current_steps": 365, "total_steps": 418, "loss": 0.4825, "lr": 8.66692456033029e-08, "epoch": 1.739856801909308, "percentage": 87.32, "elapsed_time": "4:13:21", "remaining_time": "0:36:47"}
+{"current_steps": 366, "total_steps": 418, "loss": 0.4699, "lr": 8.347542055667311e-08, "epoch": 1.7446300715990453, "percentage": 87.56, "elapsed_time": "4:14:00", "remaining_time": "0:36:05"}
+{"current_steps": 367, "total_steps": 418, "loss": 0.4679, "lr": 8.033898872882394e-08, "epoch": 1.7494033412887828, "percentage": 87.8, "elapsed_time": "4:14:39", "remaining_time": "0:35:23"}
+{"current_steps": 368, "total_steps": 418, "loss": 0.4421, "lr": 7.726014652465507e-08, "epoch": 1.7541766109785204, "percentage": 88.04, "elapsed_time": "4:15:15", "remaining_time": "0:34:40"}
+{"current_steps": 369, "total_steps": 418, "loss": 0.4778, "lr": 7.423908674277579e-08, "epoch": 1.7589498806682577, "percentage": 88.28, "elapsed_time": "4:15:53", "remaining_time": "0:33:58"}
+{"current_steps": 370, "total_steps": 418, "loss": 0.4727, "lr": 7.127599856343192e-08, "epoch": 1.763723150357995, "percentage": 88.52, "elapsed_time": "4:16:32", "remaining_time": "0:33:16"}
+{"current_steps": 371, "total_steps": 418, "loss": 0.4741, "lr": 6.837106753665823e-08, "epoch": 1.7684964200477327, "percentage": 88.76, "elapsed_time": "4:17:11", "remaining_time": "0:32:34"}
+{"current_steps": 372, "total_steps": 418, "loss": 0.4697, "lr": 6.552447557066109e-08, "epoch": 1.7732696897374702, "percentage": 89.0, "elapsed_time": "4:17:48", "remaining_time": "0:31:52"}
+{"current_steps": 373, "total_steps": 418, "loss": 0.4544, "lr": 6.273640092042575e-08, "epoch": 1.7780429594272076, "percentage": 89.23, "elapsed_time": "4:18:25", "remaining_time": "0:31:10"}
+{"current_steps": 374, "total_steps": 418, "loss": 0.4523, "lr": 6.000701817655474e-08, "epoch": 1.7828162291169452, "percentage": 89.47, "elapsed_time": "4:19:05", "remaining_time": "0:30:28"}
+{"current_steps": 375, "total_steps": 418, "loss": 0.4551, "lr": 5.733649825433384e-08, "epoch": 1.7875894988066827, "percentage": 89.71, "elapsed_time": "4:19:44", "remaining_time": "0:29:47"}
+{"current_steps": 376, "total_steps": 418, "loss": 0.494, "lr": 5.47250083830314e-08, "epoch": 1.79236276849642, "percentage": 89.95, "elapsed_time": "4:20:22", "remaining_time": "0:29:05"}
+{"current_steps": 377, "total_steps": 418, "loss": 0.4735, "lr": 5.217271209542384e-08, "epoch": 1.7971360381861574, "percentage": 90.19, "elapsed_time": "4:21:03", "remaining_time": "0:28:23"}
+{"current_steps": 378, "total_steps": 418, "loss": 0.4501, "lr": 4.967976921755679e-08, "epoch": 1.801909307875895, "percentage": 90.43, "elapsed_time": "4:21:41", "remaining_time": "0:27:41"}
+{"current_steps": 379, "total_steps": 418, "loss": 0.4686, "lr": 4.724633585873627e-08, "epoch": 1.8066825775656326, "percentage": 90.67, "elapsed_time": "4:22:21", "remaining_time": "0:26:59"}
+{"current_steps": 380, "total_steps": 418, "loss": 0.4771, "lr": 4.487256440175291e-08, "epoch": 1.81145584725537, "percentage": 90.91, "elapsed_time": "4:23:00", "remaining_time": "0:26:18"}
+{"current_steps": 381, "total_steps": 418, "loss": 0.4661, "lr": 4.255860349334006e-08, "epoch": 1.8162291169451072, "percentage": 91.15, "elapsed_time": "4:23:40", "remaining_time": "0:25:36"}
+{"current_steps": 382, "total_steps": 418, "loss": 0.4606, "lr": 4.030459803486464e-08, "epoch": 1.8210023866348448, "percentage": 91.39, "elapsed_time": "4:24:21", "remaining_time": "0:24:54"}
+{"current_steps": 383, "total_steps": 418, "loss": 0.4442, "lr": 3.811068917325444e-08, "epoch": 1.8257756563245824, "percentage": 91.63, "elapsed_time": "4:24:58", "remaining_time": "0:24:12"}
+{"current_steps": 384, "total_steps": 418, "loss": 0.4739, "lr": 3.5977014292158495e-08, "epoch": 1.8305489260143197, "percentage": 91.87, "elapsed_time": "4:25:36", "remaining_time": "0:23:31"}
+{"current_steps": 385, "total_steps": 418, "loss": 0.4719, "lr": 3.3903707003344774e-08, "epoch": 1.8353221957040573, "percentage": 92.11, "elapsed_time": "4:26:16", "remaining_time": "0:22:49"}
+{"current_steps": 386, "total_steps": 418, "loss": 0.4772, "lr": 3.189089713833226e-08, "epoch": 1.8400954653937949, "percentage": 92.34, "elapsed_time": "4:26:54", "remaining_time": "0:22:07"}
+{"current_steps": 387, "total_steps": 418, "loss": 0.4561, "lr": 2.9938710740262884e-08, "epoch": 1.8448687350835322, "percentage": 92.58, "elapsed_time": "4:27:35", "remaining_time": "0:21:26"}
+{"current_steps": 388, "total_steps": 418, "loss": 0.4565, "lr": 2.8047270056005934e-08, "epoch": 1.8496420047732696, "percentage": 92.82, "elapsed_time": "4:28:15", "remaining_time": "0:20:44"}
+{"current_steps": 389, "total_steps": 418, "loss": 0.4648, "lr": 2.6216693528505195e-08, "epoch": 1.8544152744630071, "percentage": 93.06, "elapsed_time": "4:28:54", "remaining_time": "0:20:02"}
+{"current_steps": 390, "total_steps": 418, "loss": 0.4711, "lr": 2.4447095789360884e-08, "epoch": 1.8591885441527447, "percentage": 93.3, "elapsed_time": "4:29:34", "remaining_time": "0:19:21"}
+{"current_steps": 391, "total_steps": 418, "loss": 0.4577, "lr": 2.2738587651651487e-08, "epoch": 1.863961813842482, "percentage": 93.54, "elapsed_time": "4:30:14", "remaining_time": "0:18:39"}
+{"current_steps": 392, "total_steps": 418, "loss": 0.4679, "lr": 2.109127610299466e-08, "epoch": 1.8687350835322196, "percentage": 93.78, "elapsed_time": "4:30:55", "remaining_time": "0:17:58"}
+{"current_steps": 393, "total_steps": 418, "loss": 0.4559, "lr": 1.950526429884769e-08, "epoch": 1.8735083532219572, "percentage": 94.02, "elapsed_time": "4:31:33", "remaining_time": "0:17:16"}
+{"current_steps": 394, "total_steps": 418, "loss": 0.4732, "lr": 1.7980651556048e-08, "epoch": 1.8782816229116945, "percentage": 94.26, "elapsed_time": "4:32:11", "remaining_time": "0:16:34"}
+{"current_steps": 395, "total_steps": 418, "loss": 0.4758, "lr": 1.6517533346593226e-08, "epoch": 1.8830548926014319, "percentage": 94.5, "elapsed_time": "4:32:49", "remaining_time": "0:15:53"}
+{"current_steps": 396, "total_steps": 418, "loss": 0.4932, "lr": 1.5116001291663462e-08, "epoch": 1.8878281622911695, "percentage": 94.74, "elapsed_time": "4:33:28", "remaining_time": "0:15:11"}
+{"current_steps": 397, "total_steps": 418, "loss": 0.4558, "lr": 1.3776143155883491e-08, "epoch": 1.892601431980907, "percentage": 94.98, "elapsed_time": "4:34:07", "remaining_time": "0:14:30"}
+{"current_steps": 398, "total_steps": 418, "loss": 0.4595, "lr": 1.2498042841827317e-08, "epoch": 1.8973747016706444, "percentage": 95.22, "elapsed_time": "4:34:47", "remaining_time": "0:13:48"}
+{"current_steps": 399, "total_steps": 418, "loss": 0.4625, "lr": 1.128178038476324e-08, "epoch": 1.9021479713603817, "percentage": 95.45, "elapsed_time": "4:35:26", "remaining_time": "0:13:06"}
+{"current_steps": 400, "total_steps": 418, "loss": 0.4671, "lr": 1.0127431947643316e-08, "epoch": 1.9069212410501193, "percentage": 95.69, "elapsed_time": "4:36:08", "remaining_time": "0:12:25"}
+{"current_steps": 401, "total_steps": 418, "loss": 0.464, "lr": 9.035069816332619e-09, "epoch": 1.9116945107398569, "percentage": 95.93, "elapsed_time": "4:41:27", "remaining_time": "0:11:55"}
+{"current_steps": 402, "total_steps": 418, "loss": 0.4537, "lr": 8.004762395083963e-09, "epoch": 1.9164677804295942, "percentage": 96.17, "elapsed_time": "4:42:04", "remaining_time": "0:11:13"}
+{"current_steps": 403, "total_steps": 418, "loss": 0.4914, "lr": 7.036574202253343e-09, "epoch": 1.9212410501193318, "percentage": 96.41, "elapsed_time": "4:42:44", "remaining_time": "0:10:31"}
+{"current_steps": 404, "total_steps": 418, "loss": 0.4727, "lr": 6.130565866260484e-09, "epoch": 1.9260143198090693, "percentage": 96.65, "elapsed_time": "4:43:23", "remaining_time": "0:09:49"}
+{"current_steps": 405, "total_steps": 418, "loss": 0.4767, "lr": 5.286794121791782e-09, "epoch": 1.9307875894988067, "percentage": 96.89, "elapsed_time": "4:44:04", "remaining_time": "0:09:07"}
+{"current_steps": 406, "total_steps": 418, "loss": 0.4501, "lr": 4.5053118062478025e-09, "epoch": 1.935560859188544, "percentage": 97.13, "elapsed_time": "4:44:43", "remaining_time": "0:08:24"}
+{"current_steps": 407, "total_steps": 418, "loss": 0.4747, "lr": 3.786167856434375e-09, "epoch": 1.9403341288782816, "percentage": 97.37, "elapsed_time": "4:45:21", "remaining_time": "0:07:42"}
+{"current_steps": 408, "total_steps": 418, "loss": 0.4605, "lr": 3.1294073054987102e-09, "epoch": 1.9451073985680192, "percentage": 97.61, "elapsed_time": "4:46:02", "remaining_time": "0:07:00"}
+{"current_steps": 409, "total_steps": 418, "loss": 0.4528, "lr": 2.5350712801084363e-09, "epoch": 1.9498806682577565, "percentage": 97.85, "elapsed_time": "4:46:42", "remaining_time": "0:06:18"}
+{"current_steps": 410, "total_steps": 418, "loss": 0.4585, "lr": 2.003196997877099e-09, "epoch": 1.9546539379474939, "percentage": 98.09, "elapsed_time": "4:47:20", "remaining_time": "0:05:36"}
+{"current_steps": 411, "total_steps": 418, "loss": 0.4591, "lr": 1.5338177650332517e-09, "epoch": 1.9594272076372317, "percentage": 98.33, "elapsed_time": "4:47:57", "remaining_time": "0:04:54"}
+{"current_steps": 412, "total_steps": 418, "loss": 0.4589, "lr": 1.1269629743346777e-09, "epoch": 1.964200477326969, "percentage": 98.56, "elapsed_time": "4:48:37", "remaining_time": "0:04:12"}
+{"current_steps": 413, "total_steps": 418, "loss": 0.4601, "lr": 7.826581032279734e-10, "epoch": 1.9689737470167064, "percentage": 98.8, "elapsed_time": "4:49:18", "remaining_time": "0:03:30"}
+{"current_steps": 414, "total_steps": 418, "loss": 0.4731, "lr": 5.00924712252937e-10, "epoch": 1.973747016706444, "percentage": 99.04, "elapsed_time": "4:50:01", "remaining_time": "0:02:48"}
+{"current_steps": 415, "total_steps": 418, "loss": 0.4657, "lr": 2.8178044369286945e-10, "epoch": 1.9785202863961815, "percentage": 99.28, "elapsed_time": "4:50:40", "remaining_time": "0:02:06"}
+{"current_steps": 416, "total_steps": 418, "loss": 0.452, "lr": 1.2523902046934763e-10, "epoch": 1.9832935560859188, "percentage": 99.52, "elapsed_time": "4:51:15", "remaining_time": "0:01:24"}
+{"current_steps": 417, "total_steps": 418, "loss": 0.4737, "lr": 3.131024528302273e-11, "epoch": 1.9880668257756562, "percentage": 99.76, "elapsed_time": "4:51:53", "remaining_time": "0:00:41"}
+{"current_steps": 418, "total_steps": 418, "loss": 0.4817, "lr": 0.0, "epoch": 1.9928400954653938, "percentage": 100.0, "elapsed_time": "4:52:31", "remaining_time": "0:00:00"}
+{"current_steps": 418, "total_steps": 418, "epoch": 1.9928400954653938, "percentage": 100.0, "elapsed_time": "4:57:39", "remaining_time": "0:00:00"}

trainer_state.json

@@ -0,0 +1,2968 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9928400954653938,
+  "eval_steps": 500,
+  "global_step": 418,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.00477326968973747,
+      "grad_norm": 6.016661782287111,
+      "learning_rate": 9.523809523809523e-08,
+      "loss": 1.0606,
+      "step": 1
+    },
+    {
+      "epoch": 0.00954653937947494,
+      "grad_norm": 6.0953583965166676,
+      "learning_rate": 1.9047619047619045e-07,
+      "loss": 1.0205,
+      "step": 2
+    },
+    {
+      "epoch": 0.014319809069212411,
+      "grad_norm": 5.545005089565625,
+      "learning_rate": 2.857142857142857e-07,
+      "loss": 1.0092,
+      "step": 3
+    },
+    {
+      "epoch": 0.01909307875894988,
+      "grad_norm": 6.013893802513984,
+      "learning_rate": 3.809523809523809e-07,
+      "loss": 0.9935,
+      "step": 4
+    },
+    {
+      "epoch": 0.02386634844868735,
+      "grad_norm": 5.376025361134291,
+      "learning_rate": 4.761904761904761e-07,
+      "loss": 1.0184,
+      "step": 5
+    },
+    {
+      "epoch": 0.028639618138424822,
+      "grad_norm": 5.360630219441705,
+      "learning_rate": 5.714285714285714e-07,
+      "loss": 1.0072,
+      "step": 6
+    },
+    {
+      "epoch": 0.03341288782816229,
+      "grad_norm": 5.135874756495844,
+      "learning_rate": 6.666666666666666e-07,
+      "loss": 0.9987,
+      "step": 7
+    },
+    {
+      "epoch": 0.03818615751789976,
+      "grad_norm": 4.183937042855387,
+      "learning_rate": 7.619047619047618e-07,
+      "loss": 0.9639,
+      "step": 8
+    },
+    {
+      "epoch": 0.04295942720763723,
+      "grad_norm": 4.099248159117762,
+      "learning_rate": 8.57142857142857e-07,
+      "loss": 0.9497,
+      "step": 9
+    },
+    {
+      "epoch": 0.0477326968973747,
+      "grad_norm": 4.049680347020253,
+      "learning_rate": 9.523809523809522e-07,
+      "loss": 0.9488,
+      "step": 10
+    },
+    {
+      "epoch": 0.05250596658711217,
+      "grad_norm": 3.3413136627880506,
+      "learning_rate": 1.0476190476190476e-06,
+      "loss": 0.925,
+      "step": 11
+    },
+    {
+      "epoch": 0.057279236276849645,
+      "grad_norm": 3.0774268853711955,
+      "learning_rate": 1.1428571428571428e-06,
+      "loss": 0.9231,
+      "step": 12
+    },
+    {
+      "epoch": 0.06205250596658711,
+      "grad_norm": 2.911361629869161,
+      "learning_rate": 1.238095238095238e-06,
+      "loss": 0.9177,
+      "step": 13
+    },
+    {
+      "epoch": 0.06682577565632458,
+      "grad_norm": 2.428528787361087,
+      "learning_rate": 1.3333333333333332e-06,
+      "loss": 0.8811,
+      "step": 14
+    },
+    {
+      "epoch": 0.07159904534606205,
+      "grad_norm": 1.8195746682720535,
+      "learning_rate": 1.4285714285714286e-06,
+      "loss": 0.8533,
+      "step": 15
+    },
+    {
+      "epoch": 0.07637231503579953,
+      "grad_norm": 1.947836891057091,
+      "learning_rate": 1.5238095238095236e-06,
+      "loss": 0.8706,
+      "step": 16
+    },
+    {
+      "epoch": 0.081145584725537,
+      "grad_norm": 1.8462478481431221,
+      "learning_rate": 1.619047619047619e-06,
+      "loss": 0.8208,
+      "step": 17
+    },
+    {
+      "epoch": 0.08591885441527446,
+      "grad_norm": 2.137106522697402,
+      "learning_rate": 1.714285714285714e-06,
+      "loss": 0.8181,
+      "step": 18
+    },
+    {
+      "epoch": 0.09069212410501193,
+      "grad_norm": 1.9201958176581,
+      "learning_rate": 1.8095238095238095e-06,
+      "loss": 0.8245,
+      "step": 19
+    },
+    {
+      "epoch": 0.0954653937947494,
+      "grad_norm": 1.523269971944646,
+      "learning_rate": 1.9047619047619045e-06,
+      "loss": 0.7545,
+      "step": 20
+    },
+    {
+      "epoch": 0.10023866348448687,
+      "grad_norm": 1.8071688523978784,
+      "learning_rate": 2e-06,
+      "loss": 0.776,
+      "step": 21
+    },
+    {
+      "epoch": 0.10501193317422435,
+      "grad_norm": 1.5489776099138406,
+      "learning_rate": 1.9999686897547167e-06,
+      "loss": 0.7445,
+      "step": 22
+    },
+    {
+      "epoch": 0.10978520286396182,
+      "grad_norm": 1.483755076083143,
+      "learning_rate": 1.9998747609795305e-06,
+      "loss": 0.7351,
+      "step": 23
+    },
+    {
+      "epoch": 0.11455847255369929,
+      "grad_norm": 1.4689369885238521,
+      "learning_rate": 1.999718219556307e-06,
+      "loss": 0.7332,
+      "step": 24
+    },
+    {
+      "epoch": 0.11933174224343675,
+      "grad_norm": 1.4297615546988631,
+      "learning_rate": 1.999499075287747e-06,
+      "loss": 0.738,
+      "step": 25
+    },
+    {
+      "epoch": 0.12410501193317422,
+      "grad_norm": 1.1356908706873299,
+      "learning_rate": 1.999217341896772e-06,
+      "loss": 0.7089,
+      "step": 26
+    },
+    {
+      "epoch": 0.1288782816229117,
+      "grad_norm": 1.2895783555833555,
+      "learning_rate": 1.998873037025665e-06,
+      "loss": 0.6936,
+      "step": 27
+    },
+    {
+      "epoch": 0.13365155131264916,
+      "grad_norm": 1.1235754706376115,
+      "learning_rate": 1.9984661822349665e-06,
+      "loss": 0.6785,
+      "step": 28
+    },
+    {
+      "epoch": 0.13842482100238662,
+      "grad_norm": 1.0872843150821465,
+      "learning_rate": 1.997996803002123e-06,
+      "loss": 0.6978,
+      "step": 29
+    },
+    {
+      "epoch": 0.1431980906921241,
+      "grad_norm": 1.0898740583426263,
+      "learning_rate": 1.9974649287198914e-06,
+      "loss": 0.669,
+      "step": 30
+    },
+    {
+      "epoch": 0.14797136038186157,
+      "grad_norm": 1.050118078989169,
+      "learning_rate": 1.9968705926945013e-06,
+      "loss": 0.6674,
+      "step": 31
+    },
+    {
+      "epoch": 0.15274463007159905,
+      "grad_norm": 0.8999107812930152,
+      "learning_rate": 1.9962138321435656e-06,
+      "loss": 0.6765,
+      "step": 32
+    },
+    {
+      "epoch": 0.1575178997613365,
+      "grad_norm": 0.9612552915492341,
+      "learning_rate": 1.9954946881937524e-06,
+      "loss": 0.6745,
+      "step": 33
+    },
+    {
+      "epoch": 0.162291169451074,
+      "grad_norm": 0.8921943607730816,
+      "learning_rate": 1.994713205878208e-06,
+      "loss": 0.6546,
+      "step": 34
+    },
+    {
+      "epoch": 0.16706443914081145,
+      "grad_norm": 1.1285406074360596,
+      "learning_rate": 1.9938694341337393e-06,
+      "loss": 0.6612,
+      "step": 35
+    },
+    {
+      "epoch": 0.1718377088305489,
+      "grad_norm": 0.9414713249176138,
+      "learning_rate": 1.9929634257977467e-06,
+      "loss": 0.6525,
+      "step": 36
+    },
+    {
+      "epoch": 0.1766109785202864,
+      "grad_norm": 0.8006257830637218,
+      "learning_rate": 1.991995237604916e-06,
+      "loss": 0.6465,
+      "step": 37
+    },
+    {
+      "epoch": 0.18138424821002386,
+      "grad_norm": 0.9526156911755369,
+      "learning_rate": 1.9909649301836674e-06,
+      "loss": 0.6581,
+      "step": 38
+    },
+    {
+      "epoch": 0.18615751789976134,
+      "grad_norm": 0.9928551887252647,
+      "learning_rate": 1.9898725680523566e-06,
+      "loss": 0.6623,
+      "step": 39
+    },
+    {
+      "epoch": 0.1909307875894988,
+      "grad_norm": 0.9455779960744521,
+      "learning_rate": 1.9887182196152367e-06,
+      "loss": 0.6527,
+      "step": 40
+    },
+    {
+      "epoch": 0.1957040572792363,
+      "grad_norm": 0.8607850079912805,
+      "learning_rate": 1.9875019571581726e-06,
+      "loss": 0.6148,
+      "step": 41
+    },
+    {
+      "epoch": 0.20047732696897375,
+      "grad_norm": 1.050584563468554,
+      "learning_rate": 1.9862238568441165e-06,
+      "loss": 0.6288,
+      "step": 42
+    },
+    {
+      "epoch": 0.2052505966587112,
+      "grad_norm": 0.9453773840363461,
+      "learning_rate": 1.9848839987083364e-06,
+      "loss": 0.6373,
+      "step": 43
+    },
+    {
+      "epoch": 0.2100238663484487,
+      "grad_norm": 0.8585148243018186,
+      "learning_rate": 1.983482466653407e-06,
+      "loss": 0.6401,
+      "step": 44
+    },
+    {
+      "epoch": 0.21479713603818615,
+      "grad_norm": 0.9814711262628284,
+      "learning_rate": 1.982019348443952e-06,
+      "loss": 0.6274,
+      "step": 45
+    },
+    {
+      "epoch": 0.21957040572792363,
+      "grad_norm": 0.9528618373675226,
+      "learning_rate": 1.9804947357011523e-06,
+      "loss": 0.6694,
+      "step": 46
+    },
+    {
+      "epoch": 0.2243436754176611,
+      "grad_norm": 0.7974828002980384,
+      "learning_rate": 1.978908723897005e-06,
+      "loss": 0.6236,
+      "step": 47
+    },
+    {
+      "epoch": 0.22911694510739858,
+      "grad_norm": 0.8409814486609728,
+      "learning_rate": 1.9772614123483485e-06,
+      "loss": 0.6408,
+      "step": 48
+    },
+    {
+      "epoch": 0.23389021479713604,
+      "grad_norm": 0.9308103385624037,
+      "learning_rate": 1.9755529042106393e-06,
+      "loss": 0.5919,
+      "step": 49
+    },
+    {
+      "epoch": 0.2386634844868735,
+      "grad_norm": 0.8733866970343211,
+      "learning_rate": 1.973783306471495e-06,
+      "loss": 0.5969,
+      "step": 50
+    },
+    {
+      "epoch": 0.24343675417661098,
+      "grad_norm": 0.785222837947662,
+      "learning_rate": 1.971952729943994e-06,
+      "loss": 0.5973,
+      "step": 51
+    },
+    {
+      "epoch": 0.24821002386634844,
+      "grad_norm": 0.7734659209134986,
+      "learning_rate": 1.9700612892597372e-06,
+      "loss": 0.6098,
+      "step": 52
+    },
+    {
+      "epoch": 0.2529832935560859,
+      "grad_norm": 0.7186573988976016,
+      "learning_rate": 1.9681091028616676e-06,
+      "loss": 0.5991,
+      "step": 53
+    },
+    {
+      "epoch": 0.2577565632458234,
+      "grad_norm": 0.7687713083483249,
+      "learning_rate": 1.966096292996655e-06,
+      "loss": 0.612,
+      "step": 54
+    },
+    {
+      "epoch": 0.26252983293556087,
+      "grad_norm": 0.8621825025712473,
+      "learning_rate": 1.9640229857078413e-06,
+      "loss": 0.5949,
+      "step": 55
+    },
+    {
+      "epoch": 0.26730310262529833,
+      "grad_norm": 0.8257565967386995,
+      "learning_rate": 1.9618893108267454e-06,
+      "loss": 0.6103,
+      "step": 56
+    },
+    {
+      "epoch": 0.2720763723150358,
+      "grad_norm": 0.8446529899139308,
+      "learning_rate": 1.9596954019651354e-06,
+      "loss": 0.5788,
+      "step": 57
+    },
+    {
+      "epoch": 0.27684964200477324,
+      "grad_norm": 0.7679525180581322,
+      "learning_rate": 1.95744139650666e-06,
+      "loss": 0.6069,
+      "step": 58
+    },
+    {
+      "epoch": 0.28162291169451076,
+      "grad_norm": 2.820852049381465,
+      "learning_rate": 1.955127435598247e-06,
+      "loss": 0.5987,
+      "step": 59
+    },
+    {
+      "epoch": 0.2863961813842482,
+      "grad_norm": 0.814858484667224,
+      "learning_rate": 1.9527536641412637e-06,
+      "loss": 0.6112,
+      "step": 60
+    },
+    {
+      "epoch": 0.2911694510739857,
+      "grad_norm": 0.8763795811271151,
+      "learning_rate": 1.950320230782443e-06,
+      "loss": 0.5827,
+      "step": 61
+    },
+    {
+      "epoch": 0.29594272076372313,
+      "grad_norm": 0.7856677046080051,
+      "learning_rate": 1.9478272879045763e-06,
+      "loss": 0.596,
+      "step": 62
+    },
+    {
+      "epoch": 0.30071599045346065,
+      "grad_norm": 0.922334054326887,
+      "learning_rate": 1.9452749916169685e-06,
+      "loss": 0.6131,
+      "step": 63
+    },
+    {
+      "epoch": 0.3054892601431981,
+      "grad_norm": 0.9217839433257945,
+      "learning_rate": 1.942663501745666e-06,
+      "loss": 0.6014,
+      "step": 64
+    },
+    {
+      "epoch": 0.31026252983293556,
+      "grad_norm": 0.7832581576619595,
+      "learning_rate": 1.939992981823445e-06,
+      "loss": 0.599,
+      "step": 65
+    },
+    {
+      "epoch": 0.315035799522673,
+      "grad_norm": 0.8859458814902181,
+      "learning_rate": 1.9372635990795744e-06,
+      "loss": 0.5606,
+      "step": 66
+    },
+    {
+      "epoch": 0.3198090692124105,
+      "grad_norm": 0.8263334442045513,
+      "learning_rate": 1.934475524429339e-06,
+      "loss": 0.5845,
+      "step": 67
+    },
+    {
+      "epoch": 0.324582338902148,
+      "grad_norm": 0.7750993438508201,
+      "learning_rate": 1.9316289324633416e-06,
+      "loss": 0.5938,
+      "step": 68
+    },
+    {
+      "epoch": 0.32935560859188545,
+      "grad_norm": 0.8557785609879223,
+      "learning_rate": 1.928724001436568e-06,
+      "loss": 0.5971,
+      "step": 69
+    },
+    {
+      "epoch": 0.3341288782816229,
+      "grad_norm": 0.8226562290018226,
+      "learning_rate": 1.925760913257224e-06,
+      "loss": 0.5896,
+      "step": 70
+    },
+    {
+      "epoch": 0.33890214797136037,
+      "grad_norm": 0.7875386984949746,
+      "learning_rate": 1.922739853475345e-06,
+      "loss": 0.5957,
+      "step": 71
+    },
+    {
+      "epoch": 0.3436754176610978,
+      "grad_norm": 0.7466977444466465,
+      "learning_rate": 1.919661011271176e-06,
+      "loss": 0.5782,
+      "step": 72
+    },
+    {
+      "epoch": 0.34844868735083534,
+      "grad_norm": 0.8961994451430955,
+      "learning_rate": 1.916524579443327e-06,
+      "loss": 0.5912,
+      "step": 73
+    },
+    {
+      "epoch": 0.3532219570405728,
+      "grad_norm": 0.80271020509165,
+      "learning_rate": 1.9133307543966972e-06,
+      "loss": 0.568,
+      "step": 74
+    },
+    {
+      "epoch": 0.35799522673031026,
+      "grad_norm": 9.459451148135054,
+      "learning_rate": 1.910079736130178e-06,
+      "loss": 0.5831,
+      "step": 75
+    },
+    {
+      "epoch": 0.3627684964200477,
+      "grad_norm": 0.8049759872673024,
+      "learning_rate": 1.9067717282241275e-06,
+      "loss": 0.5752,
+      "step": 76
+    },
+    {
+      "epoch": 0.36754176610978523,
+      "grad_norm": 0.9365222380955207,
+      "learning_rate": 1.9034069378276248e-06,
+      "loss": 0.6037,
+      "step": 77
+    },
+    {
+      "epoch": 0.3723150357995227,
+      "grad_norm": 0.7813871400804118,
+      "learning_rate": 1.8999855756454943e-06,
+      "loss": 0.5814,
+      "step": 78
+    },
+    {
+      "epoch": 0.37708830548926014,
+      "grad_norm": 0.8403752789759832,
+      "learning_rate": 1.8965078559251141e-06,
+      "loss": 0.5864,
+      "step": 79
+    },
+    {
+      "epoch": 0.3818615751789976,
+      "grad_norm": 0.8051967980548511,
+      "learning_rate": 1.892973996443e-06,
+      "loss": 0.5872,
+      "step": 80
+    },
+    {
+      "epoch": 0.38663484486873506,
+      "grad_norm": 0.8042594188373205,
+      "learning_rate": 1.8893842184911652e-06,
+      "loss": 0.5763,
+      "step": 81
+    },
+    {
+      "epoch": 0.3914081145584726,
+      "grad_norm": 0.9020174309993688,
+      "learning_rate": 1.8857387468632673e-06,
+      "loss": 0.5663,
+      "step": 82
+    },
+    {
+      "epoch": 0.39618138424821003,
+      "grad_norm": 0.7886287092080712,
+      "learning_rate": 1.8820378098405269e-06,
+      "loss": 0.5749,
+      "step": 83
+    },
+    {
+      "epoch": 0.4009546539379475,
+      "grad_norm": 0.7891386094058271,
+      "learning_rate": 1.878281639177437e-06,
+      "loss": 0.5791,
+      "step": 84
+    },
+    {
+      "epoch": 0.40572792362768495,
+      "grad_norm": 0.8638559742903111,
+      "learning_rate": 1.874470470087246e-06,
+      "loss": 0.594,
+      "step": 85
+    },
+    {
+      "epoch": 0.4105011933174224,
+      "grad_norm": 0.8722054176885525,
+      "learning_rate": 1.8706045412272329e-06,
+      "loss": 0.5958,
+      "step": 86
+    },
+    {
+      "epoch": 0.4152744630071599,
+      "grad_norm": 0.8861516356836725,
+      "learning_rate": 1.8666840946837588e-06,
+      "loss": 0.5831,
+      "step": 87
+    },
+    {
+      "epoch": 0.4200477326968974,
+      "grad_norm": 1.1646833402992178,
+      "learning_rate": 1.8627093759571097e-06,
+      "loss": 0.5773,
+      "step": 88
+    },
+    {
+      "epoch": 0.42482100238663484,
+      "grad_norm": 1.015546055180046,
+      "learning_rate": 1.8586806339461223e-06,
+      "loss": 0.567,
+      "step": 89
+    },
+    {
+      "epoch": 0.4295942720763723,
+      "grad_norm": 0.9466071623549958,
+      "learning_rate": 1.8545981209325974e-06,
+      "loss": 0.5859,
+      "step": 90
+    },
+    {
+      "epoch": 0.4343675417661098,
+      "grad_norm": 0.7761872762176855,
+      "learning_rate": 1.850462092565503e-06,
+      "loss": 0.5786,
+      "step": 91
+    },
+    {
+      "epoch": 0.43914081145584727,
+      "grad_norm": 0.7212954328261074,
+      "learning_rate": 1.846272807844964e-06,
+      "loss": 0.5643,
+      "step": 92
+    },
+    {
+      "epoch": 0.4439140811455847,
+      "grad_norm": 0.9652990021129971,
+      "learning_rate": 1.8420305291060453e-06,
+      "loss": 0.5772,
+      "step": 93
+    },
+    {
+      "epoch": 0.4486873508353222,
+      "grad_norm": 3.476043998914064,
+      "learning_rate": 1.837735522002322e-06,
+      "loss": 0.5973,
+      "step": 94
+    },
+    {
+      "epoch": 0.45346062052505964,
+      "grad_norm": 0.9648957060855661,
+      "learning_rate": 1.8333880554892465e-06,
+      "loss": 0.5683,
+      "step": 95
+    },
+    {
+      "epoch": 0.45823389021479716,
+      "grad_norm": 0.8226895202723103,
+      "learning_rate": 1.828988401807304e-06,
+      "loss": 0.5631,
+      "step": 96
+    },
+    {
+      "epoch": 0.4630071599045346,
+      "grad_norm": 0.8353418687299229,
+      "learning_rate": 1.8245368364649672e-06,
+      "loss": 0.5478,
+      "step": 97
+    },
+    {
+      "epoch": 0.4677804295942721,
+      "grad_norm": 0.7861845701165756,
+      "learning_rate": 1.8200336382214404e-06,
+      "loss": 0.5814,
+      "step": 98
+    },
+    {
+      "epoch": 0.47255369928400953,
+      "grad_norm": 0.7869818557092823,
+      "learning_rate": 1.815479089069208e-06,
+      "loss": 0.5831,
+      "step": 99
+    },
+    {
+      "epoch": 0.477326968973747,
+      "grad_norm": 1.0793699054838668,
+      "learning_rate": 1.8108734742163714e-06,
+      "loss": 0.5711,
+      "step": 100
+    },
+    {
+      "epoch": 0.4821002386634845,
+      "grad_norm": 0.9191351283369057,
+      "learning_rate": 1.8062170820687923e-06,
+      "loss": 0.5829,
+      "step": 101
+    },
+    {
+      "epoch": 0.48687350835322196,
+      "grad_norm": 0.8555793060148964,
+      "learning_rate": 1.8015102042120314e-06,
+      "loss": 0.5651,
+      "step": 102
+    },
+    {
+      "epoch": 0.4916467780429594,
+      "grad_norm": 0.8381062392654873,
+      "learning_rate": 1.796753135393089e-06,
+      "loss": 0.578,
+      "step": 103
+    },
+    {
+      "epoch": 0.4964200477326969,
+      "grad_norm": 0.9192300787533598,
+      "learning_rate": 1.791946173501948e-06,
+      "loss": 0.549,
+      "step": 104
+    },
+    {
+      "epoch": 0.5011933174224343,
+      "grad_norm": 0.8307533286502056,
+      "learning_rate": 1.7870896195529204e-06,
+      "loss": 0.5427,
+      "step": 105
+    },
+    {
+      "epoch": 0.5059665871121718,
+      "grad_norm": 0.7905696548307439,
+      "learning_rate": 1.7821837776657967e-06,
+      "loss": 0.5765,
+      "step": 106
+    },
+    {
+      "epoch": 0.5107398568019093,
+      "grad_norm": 0.8311340345264336,
+      "learning_rate": 1.777228955046803e-06,
+      "loss": 0.5627,
+      "step": 107
+    },
+    {
+      "epoch": 0.5155131264916468,
+      "grad_norm": 1.1408460136923761,
+      "learning_rate": 1.7722254619693617e-06,
+      "loss": 0.5615,
+      "step": 108
+    },
+    {
+      "epoch": 0.5202863961813843,
+      "grad_norm": 0.9215940982960842,
+      "learning_rate": 1.7671736117546643e-06,
+      "loss": 0.559,
+      "step": 109
+    },
+    {
+      "epoch": 0.5250596658711217,
+      "grad_norm": 0.9073194364535173,
+      "learning_rate": 1.7620737207520498e-06,
+      "loss": 0.5675,
+      "step": 110
+    },
+    {
+      "epoch": 0.5298329355608592,
+      "grad_norm": 0.9064733521778133,
+      "learning_rate": 1.756926108319194e-06,
+      "loss": 0.564,
+      "step": 111
+    },
+    {
+      "epoch": 0.5346062052505967,
+      "grad_norm": 0.8006367733355821,
+      "learning_rate": 1.751731096802113e-06,
+      "loss": 0.5697,
+      "step": 112
+    },
+    {
+      "epoch": 0.5393794749403341,
+      "grad_norm": 0.7703477827683232,
+      "learning_rate": 1.7464890115149759e-06,
+      "loss": 0.5556,
+      "step": 113
+    },
+    {
+      "epoch": 0.5441527446300716,
+      "grad_norm": 0.7808625090724881,
+      "learning_rate": 1.7412001807197361e-06,
+      "loss": 0.5699,
+      "step": 114
+    },
+    {
+      "epoch": 0.548926014319809,
+      "grad_norm": 0.7891354086520267,
+      "learning_rate": 1.735864935605572e-06,
+      "loss": 0.5535,
+      "step": 115
+    },
+    {
+      "epoch": 0.5536992840095465,
+      "grad_norm": 0.8559410057738829,
+      "learning_rate": 1.7304836102681493e-06,
+      "loss": 0.5456,
+      "step": 116
+    },
+    {
+      "epoch": 0.5584725536992841,
+      "grad_norm": 1.0113045114994854,
+      "learning_rate": 1.7250565416887015e-06,
+      "loss": 0.5724,
+      "step": 117
+    },
+    {
+      "epoch": 0.5632458233890215,
+      "grad_norm": 0.8876991951748312,
+      "learning_rate": 1.719584069712925e-06,
+      "loss": 0.568,
+      "step": 118
+    },
+    {
+      "epoch": 0.568019093078759,
+      "grad_norm": 0.8642199309829095,
+      "learning_rate": 1.7140665370296992e-06,
+      "loss": 0.5501,
+      "step": 119
+    },
+    {
+      "epoch": 0.5727923627684964,
+      "grad_norm": 0.7976943947559357,
+      "learning_rate": 1.708504289149628e-06,
+      "loss": 0.586,
+      "step": 120
+    },
+    {
+      "epoch": 0.5775656324582339,
+      "grad_norm": 0.8256312101115841,
+      "learning_rate": 1.702897674383402e-06,
+      "loss": 0.5533,
+      "step": 121
+    },
+    {
+      "epoch": 0.5823389021479713,
+      "grad_norm": 1.0090990785205396,
+      "learning_rate": 1.697247043819988e-06,
+      "loss": 0.5662,
+      "step": 122
+    },
+    {
+      "epoch": 0.5871121718377088,
+      "grad_norm": 0.9155456337094188,
+      "learning_rate": 1.6915527513046443e-06,
+      "loss": 0.5683,
+      "step": 123
+    },
+    {
+      "epoch": 0.5918854415274463,
+      "grad_norm": 0.8131468025811117,
+      "learning_rate": 1.6858151534167616e-06,
+      "loss": 0.5621,
+      "step": 124
+    },
+    {
+      "epoch": 0.5966587112171837,
+      "grad_norm": 0.8064567687343521,
+      "learning_rate": 1.6800346094475346e-06,
+      "loss": 0.5596,
+      "step": 125
+    },
+    {
+      "epoch": 0.6014319809069213,
+      "grad_norm": 0.7492395201342102,
+      "learning_rate": 1.6742114813774618e-06,
+      "loss": 0.5531,
+      "step": 126
+    },
+    {
+      "epoch": 0.6062052505966588,
+      "grad_norm": 0.7647965464540142,
+      "learning_rate": 1.6683461338536798e-06,
+      "loss": 0.5832,
+      "step": 127
+    },
+    {
+      "epoch": 0.6109785202863962,
+      "grad_norm": 0.7808066517921948,
+      "learning_rate": 1.6624389341671278e-06,
+      "loss": 0.5541,
+      "step": 128
+    },
+    {
+      "epoch": 0.6157517899761337,
+      "grad_norm": 0.8430152851631113,
+      "learning_rate": 1.656490252229548e-06,
+      "loss": 0.5528,
+      "step": 129
+    },
+    {
+      "epoch": 0.6205250596658711,
+      "grad_norm": 0.799740321239669,
+      "learning_rate": 1.6505004605503223e-06,
+      "loss": 0.5754,
+      "step": 130
+    },
+    {
+      "epoch": 0.6252983293556086,
+      "grad_norm": 0.8524369396059758,
+      "learning_rate": 1.6444699342131428e-06,
+      "loss": 0.5659,
+      "step": 131
+    },
+    {
+      "epoch": 0.630071599045346,
+      "grad_norm": 0.8594592125322017,
+      "learning_rate": 1.638399050852528e-06,
+      "loss": 0.5468,
+      "step": 132
+    },
+    {
+      "epoch": 0.6348448687350835,
+      "grad_norm": 0.8710890648276657,
+      "learning_rate": 1.632288190630172e-06,
+      "loss": 0.5547,
+      "step": 133
+    },
+    {
+      "epoch": 0.639618138424821,
+      "grad_norm": 1.3695399621239903,
+      "learning_rate": 1.6261377362111396e-06,
+      "loss": 0.5475,
+      "step": 134
+    },
+    {
+      "epoch": 0.6443914081145584,
+      "grad_norm": 0.9119912953537386,
+      "learning_rate": 1.6199480727399032e-06,
+      "loss": 0.5622,
+      "step": 135
+    },
+    {
+      "epoch": 0.649164677804296,
+      "grad_norm": 0.8174877663301265,
+      "learning_rate": 1.6137195878162267e-06,
+      "loss": 0.5646,
+      "step": 136
+    },
+    {
+      "epoch": 0.6539379474940334,
+      "grad_norm": 0.9968710402813645,
+      "learning_rate": 1.607452671470891e-06,
+      "loss": 0.5524,
+      "step": 137
+    },
+    {
+      "epoch": 0.6587112171837709,
+      "grad_norm": 0.7838173267581942,
+      "learning_rate": 1.601147716141272e-06,
+      "loss": 0.5517,
+      "step": 138
+    },
+    {
+      "epoch": 0.6634844868735084,
+      "grad_norm": 0.8600041378892647,
+      "learning_rate": 1.5948051166467657e-06,
+      "loss": 0.5664,
+      "step": 139
+    },
+    {
+      "epoch": 0.6682577565632458,
+      "grad_norm": 0.7393813982622772,
+      "learning_rate": 1.5884252701640634e-06,
+      "loss": 0.5611,
+      "step": 140
+    },
+    {
+      "epoch": 0.6730310262529833,
+      "grad_norm": 0.8312116599801993,
+      "learning_rate": 1.5820085762022823e-06,
+      "loss": 0.5609,
+      "step": 141
+    },
+    {
+      "epoch": 0.6778042959427207,
+      "grad_norm": 0.782610924284724,
+      "learning_rate": 1.5755554365779455e-06,
+      "loss": 0.5586,
+      "step": 142
+    },
+    {
+      "epoch": 0.6825775656324582,
+      "grad_norm": 0.7869375949652244,
+      "learning_rate": 1.5690662553898222e-06,
+      "loss": 0.5557,
+      "step": 143
+    },
+    {
+      "epoch": 0.6873508353221957,
+      "grad_norm": 0.7871275055021261,
+      "learning_rate": 1.5625414389936218e-06,
+      "loss": 0.5379,
+      "step": 144
+    },
+    {
+      "epoch": 0.6921241050119332,
+      "grad_norm": 0.7978567113817064,
+      "learning_rate": 1.555981395976548e-06,
+      "loss": 0.5459,
+      "step": 145
+    },
+    {
+      "epoch": 0.6968973747016707,
+      "grad_norm": 0.8678454065910531,
+      "learning_rate": 1.5493865371317123e-06,
+      "loss": 0.5538,
+      "step": 146
+    },
+    {
+      "epoch": 0.7016706443914081,
+      "grad_norm": 0.8640558568867235,
+      "learning_rate": 1.542757275432411e-06,
+      "loss": 0.5511,
+      "step": 147
+    },
+    {
+      "epoch": 0.7064439140811456,
+      "grad_norm": 0.8257539417151866,
+      "learning_rate": 1.5360940260062635e-06,
+      "loss": 0.5395,
+      "step": 148
+    },
+    {
+      "epoch": 0.711217183770883,
+      "grad_norm": 0.7735477084244853,
+      "learning_rate": 1.5293972061092185e-06,
+      "loss": 0.5487,
+      "step": 149
+    },
+    {
+      "epoch": 0.7159904534606205,
+      "grad_norm": 2.21607832896325,
+      "learning_rate": 1.522667235099422e-06,
+      "loss": 0.5313,
+      "step": 150
+    },
+    {
+      "epoch": 0.720763723150358,
+      "grad_norm": 0.8260305997634725,
+      "learning_rate": 1.515904534410961e-06,
+      "loss": 0.548,
+      "step": 151
+    },
+    {
+      "epoch": 0.7255369928400954,
+      "grad_norm": 0.9282281415854876,
+      "learning_rate": 1.5091095275274699e-06,
+      "loss": 0.5366,
+      "step": 152
+    },
+    {
+      "epoch": 0.7303102625298329,
+      "grad_norm": 0.835392664470487,
+      "learning_rate": 1.5022826399556133e-06,
+      "loss": 0.5365,
+      "step": 153
+    },
+    {
+      "epoch": 0.7350835322195705,
+      "grad_norm": 1.0014547232970634,
+      "learning_rate": 1.4954242991984396e-06,
+      "loss": 0.5601,
+      "step": 154
+    },
+    {
+      "epoch": 0.7398568019093079,
+      "grad_norm": 0.7999358357306402,
+      "learning_rate": 1.4885349347286115e-06,
+      "loss": 0.549,
+      "step": 155
+    },
+    {
+      "epoch": 0.7446300715990454,
+      "grad_norm": 0.7456244196208853,
+      "learning_rate": 1.4816149779615126e-06,
+      "loss": 0.5516,
+      "step": 156
+    },
+    {
+      "epoch": 0.7494033412887828,
+      "grad_norm": 0.7568817924270603,
+      "learning_rate": 1.474664862228229e-06,
+      "loss": 0.5572,
+      "step": 157
+    },
+    {
+      "epoch": 0.7541766109785203,
+      "grad_norm": 0.9329993871672655,
+      "learning_rate": 1.467685022748419e-06,
+      "loss": 0.5617,
+      "step": 158
+    },
+    {
+      "epoch": 0.7589498806682577,
+      "grad_norm": 0.7402702977169047,
+      "learning_rate": 1.4606758966030534e-06,
+      "loss": 0.5426,
+      "step": 159
+    },
+    {
+      "epoch": 0.7637231503579952,
+      "grad_norm": 0.7912657849322988,
+      "learning_rate": 1.4536379227070509e-06,
+      "loss": 0.544,
+      "step": 160
+    },
+    {
+      "epoch": 0.7684964200477327,
+      "grad_norm": 0.8280839624728757,
+      "learning_rate": 1.4465715417817888e-06,
+      "loss": 0.5435,
+      "step": 161
+    },
+    {
+      "epoch": 0.7732696897374701,
+      "grad_norm": 0.7376680395132865,
+      "learning_rate": 1.4394771963275076e-06,
+      "loss": 0.5199,
+      "step": 162
+    },
+    {
+      "epoch": 0.7780429594272077,
+      "grad_norm": 0.7984252215551224,
+      "learning_rate": 1.4323553305955997e-06,
+      "loss": 0.5479,
+      "step": 163
+    },
+    {
+      "epoch": 0.7828162291169452,
+      "grad_norm": 0.788726316639838,
+      "learning_rate": 1.4252063905607909e-06,
+      "loss": 0.5219,
+      "step": 164
+    },
+    {
+      "epoch": 0.7875894988066826,
+      "grad_norm": 0.7350598897520126,
+      "learning_rate": 1.4180308238932135e-06,
+      "loss": 0.531,
+      "step": 165
+    },
+    {
+      "epoch": 0.7923627684964201,
+      "grad_norm": 0.7786806805958749,
+      "learning_rate": 1.410829079930372e-06,
+      "loss": 0.5481,
+      "step": 166
+    },
+    {
+      "epoch": 0.7971360381861575,
+      "grad_norm": 0.9607237271282482,
+      "learning_rate": 1.4036016096490064e-06,
+      "loss": 0.5478,
+      "step": 167
+    },
+    {
+      "epoch": 0.801909307875895,
+      "grad_norm": 0.7782148550862285,
+      "learning_rate": 1.3963488656368517e-06,
+      "loss": 0.535,
+      "step": 168
+    },
+    {
+      "epoch": 0.8066825775656324,
+      "grad_norm": 0.8100946646751193,
+      "learning_rate": 1.389071302064295e-06,
+      "loss": 0.5277,
+      "step": 169
+    },
+    {
+      "epoch": 0.8114558472553699,
+      "grad_norm": 0.7502947220609039,
+      "learning_rate": 1.381769374655938e-06,
+      "loss": 0.5553,
+      "step": 170
+    },
+    {
+      "epoch": 0.8162291169451074,
+      "grad_norm": 0.9124000354997026,
+      "learning_rate": 1.374443540662057e-06,
+      "loss": 0.5518,
+      "step": 171
+    },
+    {
+      "epoch": 0.8210023866348448,
+      "grad_norm": 0.8409623949497625,
+      "learning_rate": 1.3670942588299705e-06,
+      "loss": 0.5294,
+      "step": 172
+    },
+    {
+      "epoch": 0.8257756563245824,
+      "grad_norm": 0.8018568702519514,
+      "learning_rate": 1.3597219893753117e-06,
+      "loss": 0.5121,
+      "step": 173
+    },
+    {
+      "epoch": 0.8305489260143198,
+      "grad_norm": 0.9262097539109866,
+      "learning_rate": 1.352327193953211e-06,
+      "loss": 0.5259,
+      "step": 174
+    },
+    {
+      "epoch": 0.8353221957040573,
+      "grad_norm": 0.7289872898963717,
+      "learning_rate": 1.3449103356293852e-06,
+      "loss": 0.5601,
+      "step": 175
+    },
+    {
+      "epoch": 0.8400954653937948,
+      "grad_norm": 0.7836398407929648,
+      "learning_rate": 1.337471878851141e-06,
+      "loss": 0.5359,
+      "step": 176
+    },
+    {
+      "epoch": 0.8448687350835322,
+      "grad_norm": 0.8058359597234802,
+      "learning_rate": 1.3300122894182909e-06,
+      "loss": 0.5485,
+      "step": 177
+    },
+    {
+      "epoch": 0.8496420047732697,
+      "grad_norm": 0.9118002301436436,
+      "learning_rate": 1.3225320344539842e-06,
+      "loss": 0.5562,
+      "step": 178
+    },
+    {
+      "epoch": 0.8544152744630071,
+      "grad_norm": 0.7609979767002807,
+      "learning_rate": 1.315031582375457e-06,
+      "loss": 0.5485,
+      "step": 179
+    },
+    {
+      "epoch": 0.8591885441527446,
+      "grad_norm": 0.7105869344115592,
+      "learning_rate": 1.3075114028646974e-06,
+      "loss": 0.5444,
+      "step": 180
+    },
+    {
+      "epoch": 0.863961813842482,
+      "grad_norm": 0.8004311294692876,
+      "learning_rate": 1.299971966839036e-06,
+      "loss": 0.5481,
+      "step": 181
+    },
+    {
+      "epoch": 0.8687350835322196,
+      "grad_norm": 0.7667234252631754,
+      "learning_rate": 1.292413746421655e-06,
+      "loss": 0.5345,
+      "step": 182
+    },
+    {
+      "epoch": 0.8735083532219571,
+      "grad_norm": 0.7709523318159157,
+      "learning_rate": 1.2848372149120246e-06,
+      "loss": 0.512,
+      "step": 183
+    },
+    {
+      "epoch": 0.8782816229116945,
+      "grad_norm": 0.8742048693859581,
+      "learning_rate": 1.2772428467562651e-06,
+      "loss": 0.55,
+      "step": 184
+    },
+    {
+      "epoch": 0.883054892601432,
+      "grad_norm": 0.8768649061250284,
+      "learning_rate": 1.2696311175174357e-06,
+      "loss": 0.5365,
+      "step": 185
+    },
+    {
+      "epoch": 0.8878281622911695,
+      "grad_norm": 0.8468420712736167,
+      "learning_rate": 1.2620025038457554e-06,
+      "loss": 0.5421,
+      "step": 186
+    },
+    {
+      "epoch": 0.8926014319809069,
+      "grad_norm": 0.725877140171063,
+      "learning_rate": 1.254357483448755e-06,
+      "loss": 0.519,
+      "step": 187
+    },
+    {
+      "epoch": 0.8973747016706444,
+      "grad_norm": 0.7168188099187686,
+      "learning_rate": 1.2466965350613615e-06,
+      "loss": 0.5651,
+      "step": 188
+    },
+    {
+      "epoch": 0.9021479713603818,
+      "grad_norm": 0.8993966404570418,
+      "learning_rate": 1.2390201384159219e-06,
+      "loss": 0.5603,
+      "step": 189
+    },
+    {
+      "epoch": 0.9069212410501193,
+      "grad_norm": 0.741646072361816,
+      "learning_rate": 1.231328774212159e-06,
+      "loss": 0.5157,
+      "step": 190
+    },
+    {
+      "epoch": 0.9116945107398569,
+      "grad_norm": 0.7741706595084717,
+      "learning_rate": 1.223622924087073e-06,
+      "loss": 0.5367,
+      "step": 191
+    },
+    {
+      "epoch": 0.9164677804295943,
+      "grad_norm": 0.760645151447744,
+      "learning_rate": 1.215903070584779e-06,
+      "loss": 0.5401,
+      "step": 192
+    },
+    {
+      "epoch": 0.9212410501193318,
+      "grad_norm": 0.7462809840684769,
+      "learning_rate": 1.2081696971262903e-06,
+      "loss": 0.5458,
+      "step": 193
+    },
+    {
+      "epoch": 0.9260143198090692,
+      "grad_norm": 0.867349599337623,
+      "learning_rate": 1.2004232879792464e-06,
+      "loss": 0.5398,
+      "step": 194
+    },
+    {
+      "epoch": 0.9307875894988067,
+      "grad_norm": 0.7728255267176583,
+      "learning_rate": 1.1926643282275882e-06,
+      "loss": 0.5343,
+      "step": 195
+    },
+    {
+      "epoch": 0.9355608591885441,
+      "grad_norm": 0.7946709962404823,
+      "learning_rate": 1.1848933037411825e-06,
+      "loss": 0.5181,
+      "step": 196
+    },
+    {
+      "epoch": 0.9403341288782816,
+      "grad_norm": 0.7159173523126642,
+      "learning_rate": 1.1771107011453933e-06,
+      "loss": 0.5442,
+      "step": 197
+    },
+    {
+      "epoch": 0.9451073985680191,
+      "grad_norm": 0.8493976289870552,
+      "learning_rate": 1.1693170077906143e-06,
+      "loss": 0.5467,
+      "step": 198
+    },
+    {
+      "epoch": 0.9498806682577565,
+      "grad_norm": 0.7390118080756048,
+      "learning_rate": 1.1615127117217463e-06,
+      "loss": 0.5251,
+      "step": 199
+    },
+    {
+      "epoch": 0.954653937947494,
+      "grad_norm": 0.7595495597083671,
+      "learning_rate": 1.1536983016476373e-06,
+      "loss": 0.5368,
+      "step": 200
+    },
+    {
+      "epoch": 0.9594272076372315,
+      "grad_norm": 0.7399505119485492,
+      "learning_rate": 1.1458742669104803e-06,
+      "loss": 0.514,
+      "step": 201
+    },
+    {
+      "epoch": 0.964200477326969,
+      "grad_norm": 0.7693531287817772,
+      "learning_rate": 1.1380410974551682e-06,
+      "loss": 0.5327,
+      "step": 202
+    },
+    {
+      "epoch": 0.9689737470167065,
+      "grad_norm": 0.7361655101073081,
+      "learning_rate": 1.130199283798615e-06,
+      "loss": 0.5152,
+      "step": 203
+    },
+    {
+      "epoch": 0.9737470167064439,
+      "grad_norm": 0.8174253218643999,
+      "learning_rate": 1.1223493169990391e-06,
+      "loss": 0.5376,
+      "step": 204
+    },
+    {
+      "epoch": 0.9785202863961814,
+      "grad_norm": 0.7646163527785592,
+      "learning_rate": 1.1144916886252124e-06,
+      "loss": 0.5198,
+      "step": 205
+    },
+    {
+      "epoch": 0.9832935560859188,
+      "grad_norm": 0.7600726494815581,
+      "learning_rate": 1.1066268907256782e-06,
+      "loss": 0.5358,
+      "step": 206
+    },
+    {
+      "epoch": 0.9880668257756563,
+      "grad_norm": 0.8292480992474258,
+      "learning_rate": 1.098755415797939e-06,
+      "loss": 0.5319,
+      "step": 207
+    },
+    {
+      "epoch": 0.9928400954653938,
+      "grad_norm": 0.7584975382780693,
+      "learning_rate": 1.0908777567576168e-06,
+      "loss": 0.5453,
+      "step": 208
+    },
+    {
+      "epoch": 0.9976133651551312,
+      "grad_norm": 0.7360353406613074,
+      "learning_rate": 1.0829944069075847e-06,
+      "loss": 0.5398,
+      "step": 209
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.7360353406613074,
+      "learning_rate": 1.0751058599070781e-06,
+      "loss": 0.2683,
+      "step": 210
+    },
+    {
+      "epoch": 1.0047732696897376,
+      "grad_norm": 0.7735348980384088,
+      "learning_rate": 1.0672126097407795e-06,
+      "loss": 0.4862,
+      "step": 211
+    },
+    {
+      "epoch": 1.009546539379475,
+      "grad_norm": 0.6892850244639656,
+      "learning_rate": 1.0593151506878865e-06,
+      "loss": 0.4886,
+      "step": 212
+    },
+    {
+      "epoch": 1.0143198090692125,
+      "grad_norm": 0.7416432308937427,
+      "learning_rate": 1.0514139772911597e-06,
+      "loss": 0.4755,
+      "step": 213
+    },
+    {
+      "epoch": 1.0190930787589498,
+      "grad_norm": 0.6788376232914372,
+      "learning_rate": 1.043509584325953e-06,
+      "loss": 0.4643,
+      "step": 214
+    },
+    {
+      "epoch": 1.0238663484486874,
+      "grad_norm": 0.7328906073842687,
+      "learning_rate": 1.0356024667692314e-06,
+      "loss": 0.4934,
+      "step": 215
+    },
+    {
+      "epoch": 1.0286396181384247,
+      "grad_norm": 0.7697429459150121,
+      "learning_rate": 1.0276931197685753e-06,
+      "loss": 0.4976,
+      "step": 216
+    },
+    {
+      "epoch": 1.0334128878281623,
+      "grad_norm": 0.7939705310040335,
+      "learning_rate": 1.0197820386111737e-06,
+      "loss": 0.4897,
+      "step": 217
+    },
+    {
+      "epoch": 1.0381861575178997,
+      "grad_norm": 0.9752936792347606,
+      "learning_rate": 1.0118697186928105e-06,
+      "loss": 0.4632,
+      "step": 218
+    },
+    {
+      "epoch": 1.0429594272076372,
+      "grad_norm": 0.810300278966379,
+      "learning_rate": 1.0039566554868392e-06,
+      "loss": 0.4667,
+      "step": 219
+    },
+    {
+      "epoch": 1.0477326968973748,
+      "grad_norm": 0.7651633767231123,
+      "learning_rate": 9.960433445131607e-07,
+      "loss": 0.4913,
+      "step": 220
+    },
+    {
+      "epoch": 1.0525059665871122,
+      "grad_norm": 0.7783544485209318,
+      "learning_rate": 9.881302813071896e-07,
+      "loss": 0.485,
+      "step": 221
+    },
+    {
+      "epoch": 1.0572792362768497,
+      "grad_norm": 0.7728747490030172,
+      "learning_rate": 9.802179613888262e-07,
+      "loss": 0.4663,
+      "step": 222
+    },
+    {
+      "epoch": 1.062052505966587,
+      "grad_norm": 0.7199803548701269,
+      "learning_rate": 9.723068802314246e-07,
+      "loss": 0.4724,
+      "step": 223
+    },
+    {
+      "epoch": 1.0668257756563246,
+      "grad_norm": 0.8173682429078198,
+      "learning_rate": 9.643975332307687e-07,
+      "loss": 0.4777,
+      "step": 224
+    },
+    {
+      "epoch": 1.071599045346062,
+      "grad_norm": 0.9029276240129886,
+      "learning_rate": 9.564904156740471e-07,
+      "loss": 0.4664,
+      "step": 225
+    },
+    {
+      "epoch": 1.0763723150357996,
+      "grad_norm": 0.7595074592495551,
+      "learning_rate": 9.485860227088405e-07,
+      "loss": 0.4808,
+      "step": 226
+    },
+    {
+      "epoch": 1.081145584725537,
+      "grad_norm": 0.8019805756491788,
+      "learning_rate": 9.406848493121134e-07,
+      "loss": 0.4764,
+      "step": 227
+    },
+    {
+      "epoch": 1.0859188544152745,
+      "grad_norm": 0.7750922258239085,
+      "learning_rate": 9.327873902592205e-07,
+      "loss": 0.4711,
+      "step": 228
+    },
+    {
+      "epoch": 1.0906921241050118,
+      "grad_norm": 0.7272348247085987,
+      "learning_rate": 9.248941400929222e-07,
+      "loss": 0.4753,
+      "step": 229
+    },
+    {
+      "epoch": 1.0954653937947494,
+      "grad_norm": 0.8135968715591004,
+      "learning_rate": 9.17005593092415e-07,
+      "loss": 0.49,
+      "step": 230
+    },
+    {
+      "epoch": 1.100238663484487,
+      "grad_norm": 0.784517413630989,
+      "learning_rate": 9.09122243242383e-07,
+      "loss": 0.4636,
+      "step": 231
+    },
+    {
+      "epoch": 1.1050119331742243,
+      "grad_norm": 0.7967633635464352,
+      "learning_rate": 9.01244584202061e-07,
+      "loss": 0.4638,
+      "step": 232
+    },
+    {
+      "epoch": 1.1097852028639619,
+      "grad_norm": 0.7347125585892648,
+      "learning_rate": 8.933731092743219e-07,
+      "loss": 0.4951,
+      "step": 233
+    },
+    {
+      "epoch": 1.1145584725536992,
+      "grad_norm": 0.762506543894173,
+      "learning_rate": 8.855083113747875e-07,
+      "loss": 0.4715,
+      "step": 234
+    },
+    {
+      "epoch": 1.1193317422434368,
+      "grad_norm": 0.7657159811972606,
+      "learning_rate": 8.776506830009607e-07,
+      "loss": 0.4792,
+      "step": 235
+    },
+    {
+      "epoch": 1.1241050119331741,
+      "grad_norm": 0.7746280343348994,
+      "learning_rate": 8.698007162013849e-07,
+      "loss": 0.4734,
+      "step": 236
+    },
+    {
+      "epoch": 1.1288782816229117,
+      "grad_norm": 0.7287940319917965,
+      "learning_rate": 8.619589025448318e-07,
+      "loss": 0.4899,
+      "step": 237
+    },
+    {
+      "epoch": 1.1336515513126493,
+      "grad_norm": 0.7283506274833321,
+      "learning_rate": 8.541257330895197e-07,
+      "loss": 0.461,
+      "step": 238
+    },
+    {
+      "epoch": 1.1384248210023866,
+      "grad_norm": 1.109020964160513,
+      "learning_rate": 8.463016983523627e-07,
+      "loss": 0.4789,
+      "step": 239
+    },
+    {
+      "epoch": 1.1431980906921242,
+      "grad_norm": 0.8916069268430648,
+      "learning_rate": 8.384872882782541e-07,
+      "loss": 0.4951,
+      "step": 240
+    },
+    {
+      "epoch": 1.1479713603818615,
+      "grad_norm": 0.7832561259348029,
+      "learning_rate": 8.306829922093857e-07,
+      "loss": 0.4666,
+      "step": 241
+    },
+    {
+      "epoch": 1.152744630071599,
+      "grad_norm": 0.7246823419762234,
+      "learning_rate": 8.228892988546067e-07,
+      "loss": 0.475,
+      "step": 242
+    },
+    {
+      "epoch": 1.1575178997613365,
+      "grad_norm": 0.705366097498364,
+      "learning_rate": 8.15106696258818e-07,
+      "loss": 0.4727,
+      "step": 243
+    },
+    {
+      "epoch": 1.162291169451074,
+      "grad_norm": 0.7563603316000965,
+      "learning_rate": 8.073356717724115e-07,
+      "loss": 0.4779,
+      "step": 244
+    },
+    {
+      "epoch": 1.1670644391408114,
+      "grad_norm": 0.7463996376621957,
+      "learning_rate": 7.995767120207536e-07,
+      "loss": 0.4647,
+      "step": 245
+    },
+    {
+      "epoch": 1.171837708830549,
+      "grad_norm": 0.7117618711530662,
+      "learning_rate": 7.918303028737096e-07,
+      "loss": 0.4712,
+      "step": 246
+    },
+    {
+      "epoch": 1.1766109785202863,
+      "grad_norm": 0.7445420769436453,
+      "learning_rate": 7.840969294152211e-07,
+      "loss": 0.4747,
+      "step": 247
+    },
+    {
+      "epoch": 1.1813842482100239,
+      "grad_norm": 0.7339272409779617,
+      "learning_rate": 7.763770759129269e-07,
+      "loss": 0.4732,
+      "step": 248
+    },
+    {
+      "epoch": 1.1861575178997614,
+      "grad_norm": 0.7680499628702099,
+      "learning_rate": 7.68671225787841e-07,
+      "loss": 0.4677,
+      "step": 249
+    },
+    {
+      "epoch": 1.1909307875894988,
+      "grad_norm": 0.7289596879207738,
+      "learning_rate": 7.609798615840785e-07,
+      "loss": 0.4788,
+      "step": 250
+    },
+    {
+      "epoch": 1.1957040572792363,
+      "grad_norm": 0.7375098113291024,
+      "learning_rate": 7.533034649386384e-07,
+      "loss": 0.456,
+      "step": 251
+    },
+    {
+      "epoch": 1.2004773269689737,
+      "grad_norm": 0.7788484912408599,
+      "learning_rate": 7.456425165512452e-07,
+      "loss": 0.4768,
+      "step": 252
+    },
+    {
+      "epoch": 1.2052505966587113,
+      "grad_norm": 0.7545300469644135,
+      "learning_rate": 7.379974961542447e-07,
+      "loss": 0.4864,
+      "step": 253
+    },
+    {
+      "epoch": 1.2100238663484486,
+      "grad_norm": 0.8818787967594464,
+      "learning_rate": 7.303688824825646e-07,
+      "loss": 0.4768,
+      "step": 254
+    },
+    {
+      "epoch": 1.2147971360381862,
+      "grad_norm": 0.7762788166887581,
+      "learning_rate": 7.227571532437349e-07,
+      "loss": 0.4676,
+      "step": 255
+    },
+    {
+      "epoch": 1.2195704057279237,
+      "grad_norm": 0.674374793234199,
+      "learning_rate": 7.151627850879755e-07,
+      "loss": 0.4688,
+      "step": 256
+    },
+    {
+      "epoch": 1.224343675417661,
+      "grad_norm": 0.7391271163895584,
+      "learning_rate": 7.075862535783453e-07,
+      "loss": 0.4545,
+      "step": 257
+    },
+    {
+      "epoch": 1.2291169451073987,
+      "grad_norm": 0.7377869581736503,
+      "learning_rate": 7.00028033160964e-07,
+      "loss": 0.4842,
+      "step": 258
+    },
+    {
+      "epoch": 1.233890214797136,
+      "grad_norm": 0.7182033053068443,
+      "learning_rate": 6.924885971353026e-07,
+      "loss": 0.4841,
+      "step": 259
+    },
+    {
+      "epoch": 1.2386634844868736,
+      "grad_norm": 0.7165206421556828,
+      "learning_rate": 6.849684176245431e-07,
+      "loss": 0.4485,
+      "step": 260
+    },
+    {
+      "epoch": 1.243436754176611,
+      "grad_norm": 0.8274126483370449,
+      "learning_rate": 6.774679655460158e-07,
+      "loss": 0.4632,
+      "step": 261
+    },
+    {
+      "epoch": 1.2482100238663485,
+      "grad_norm": 0.7849668814937834,
+      "learning_rate": 6.699877105817092e-07,
+      "loss": 0.4701,
+      "step": 262
+    },
+    {
+      "epoch": 1.2529832935560858,
+      "grad_norm": 0.7246643685451561,
+      "learning_rate": 6.625281211488591e-07,
+      "loss": 0.4884,
+      "step": 263
+    },
+    {
+      "epoch": 1.2577565632458234,
+      "grad_norm": 0.7413214893244733,
+      "learning_rate": 6.55089664370615e-07,
+      "loss": 0.4821,
+      "step": 264
+    },
+    {
+      "epoch": 1.2625298329355608,
+      "grad_norm": 0.7307541408287506,
+      "learning_rate": 6.476728060467888e-07,
+      "loss": 0.4585,
+      "step": 265
+    },
+    {
+      "epoch": 1.2673031026252983,
+      "grad_norm": 0.7439228818529052,
+      "learning_rate": 6.402780106246884e-07,
+      "loss": 0.4688,
+      "step": 266
+    },
+    {
+      "epoch": 1.272076372315036,
+      "grad_norm": 0.7075632105234686,
+      "learning_rate": 6.329057411700298e-07,
+      "loss": 0.4813,
+      "step": 267
+    },
+    {
+      "epoch": 1.2768496420047732,
+      "grad_norm": 0.757650326028371,
+      "learning_rate": 6.255564593379429e-07,
+      "loss": 0.4878,
+      "step": 268
+    },
+    {
+      "epoch": 1.2816229116945108,
+      "grad_norm": 0.729712295017678,
+      "learning_rate": 6.182306253440619e-07,
+      "loss": 0.4629,
+      "step": 269
+    },
+    {
+      "epoch": 1.2863961813842482,
+      "grad_norm": 0.8230987908171445,
+      "learning_rate": 6.109286979357051e-07,
+      "loss": 0.4842,
+      "step": 270
+    },
+    {
+      "epoch": 1.2911694510739857,
+      "grad_norm": 0.7878144207218812,
+      "learning_rate": 6.036511343631488e-07,
+      "loss": 0.4588,
+      "step": 271
+    },
+    {
+      "epoch": 1.295942720763723,
+      "grad_norm": 0.7162555025211284,
+      "learning_rate": 5.963983903509935e-07,
+      "loss": 0.4817,
+      "step": 272
+    },
+    {
+      "epoch": 1.3007159904534606,
+      "grad_norm": 0.7352227500252277,
+      "learning_rate": 5.89170920069628e-07,
+      "loss": 0.4781,
+      "step": 273
+    },
+    {
+      "epoch": 1.3054892601431982,
+      "grad_norm": 0.7097358431174013,
+      "learning_rate": 5.819691761067865e-07,
+      "loss": 0.46,
+      "step": 274
+    },
+    {
+      "epoch": 1.3102625298329356,
+      "grad_norm": 1.146161188184777,
+      "learning_rate": 5.747936094392089e-07,
+      "loss": 0.4647,
+      "step": 275
+    },
+    {
+      "epoch": 1.315035799522673,
+      "grad_norm": 0.7072592435264768,
+      "learning_rate": 5.676446694044002e-07,
+      "loss": 0.4639,
+      "step": 276
+    },
+    {
+      "epoch": 1.3198090692124105,
+      "grad_norm": 0.7215149618117556,
+      "learning_rate": 5.605228036724927e-07,
+      "loss": 0.4652,
+      "step": 277
+    },
+    {
+      "epoch": 1.324582338902148,
+      "grad_norm": 0.670785774408122,
+      "learning_rate": 5.534284582182114e-07,
+      "loss": 0.4717,
+      "step": 278
+    },
+    {
+      "epoch": 1.3293556085918854,
+      "grad_norm": 0.747767864677791,
+      "learning_rate": 5.463620772929494e-07,
+      "loss": 0.4536,
+      "step": 279
+    },
+    {
+      "epoch": 1.334128878281623,
+      "grad_norm": 0.8516514509018951,
+      "learning_rate": 5.393241033969466e-07,
+      "loss": 0.4649,
+      "step": 280
+    },
+    {
+      "epoch": 1.3389021479713603,
+      "grad_norm": 0.8138001829719436,
+      "learning_rate": 5.323149772515812e-07,
+      "loss": 0.4668,
+      "step": 281
+    },
+    {
+      "epoch": 1.3436754176610979,
+      "grad_norm": 0.7576171145048753,
+      "learning_rate": 5.253351377717706e-07,
+      "loss": 0.4761,
+      "step": 282
+    },
+    {
+      "epoch": 1.3484486873508352,
+      "grad_norm": 0.8613520066962265,
+      "learning_rate": 5.183850220384873e-07,
+      "loss": 0.469,
+      "step": 283
+    },
+    {
+      "epoch": 1.3532219570405728,
+      "grad_norm": 0.766228885306893,
+      "learning_rate": 5.114650652713884e-07,
+      "loss": 0.4802,
+      "step": 284
+    },
+    {
+      "epoch": 1.3579952267303104,
+      "grad_norm": 0.7068637893292556,
+      "learning_rate": 5.045757008015606e-07,
+      "loss": 0.4773,
+      "step": 285
+    },
+    {
+      "epoch": 1.3627684964200477,
+      "grad_norm": 0.8429657657602729,
+      "learning_rate": 4.977173600443868e-07,
+      "loss": 0.4605,
+      "step": 286
+    },
+    {
+      "epoch": 1.3675417661097853,
+      "grad_norm": 0.7007932505507933,
+      "learning_rate": 4.908904724725299e-07,
+      "loss": 0.4767,
+      "step": 287
+    },
+    {
+      "epoch": 1.3723150357995226,
+      "grad_norm": 0.7671222670718428,
+      "learning_rate": 4.840954655890391e-07,
+      "loss": 0.4682,
+      "step": 288
+    },
+    {
+      "epoch": 1.3770883054892602,
+      "grad_norm": 0.694265618019185,
+      "learning_rate": 4.773327649005777e-07,
+      "loss": 0.4855,
+      "step": 289
+    },
+    {
+      "epoch": 1.3818615751789975,
+      "grad_norm": 0.7519150028535938,
+      "learning_rate": 4.7060279389078184e-07,
+      "loss": 0.4761,
+      "step": 290
+    },
+    {
+      "epoch": 1.3866348448687351,
+      "grad_norm": 0.7486630511459641,
+      "learning_rate": 4.6390597399373644e-07,
+      "loss": 0.4565,
+      "step": 291
+    },
+    {
+      "epoch": 1.3914081145584727,
+      "grad_norm": 0.7422555751664944,
+      "learning_rate": 4.5724272456758907e-07,
+      "loss": 0.4826,
+      "step": 292
+    },
+    {
+      "epoch": 1.39618138424821,
+      "grad_norm": 0.77856112043872,
+      "learning_rate": 4.506134628682877e-07,
+      "loss": 0.4763,
+      "step": 293
+    },
+    {
+      "epoch": 1.4009546539379474,
+      "grad_norm": 0.7684572854516972,
+      "learning_rate": 4.440186040234524e-07,
+      "loss": 0.4672,
+      "step": 294
+    },
+    {
+      "epoch": 1.405727923627685,
+      "grad_norm": 0.7665847058665568,
+      "learning_rate": 4.3745856100637834e-07,
+      "loss": 0.4656,
+      "step": 295
+    },
+    {
+      "epoch": 1.4105011933174225,
+      "grad_norm": 0.733469970387663,
+      "learning_rate": 4.3093374461017785e-07,
+      "loss": 0.4676,
+      "step": 296
+    },
+    {
+      "epoch": 1.4152744630071599,
+      "grad_norm": 0.8421640257156171,
+      "learning_rate": 4.244445634220545e-07,
+      "loss": 0.4843,
+      "step": 297
+    },
+    {
+      "epoch": 1.4200477326968974,
+      "grad_norm": 0.8009564109297522,
+      "learning_rate": 4.1799142379771766e-07,
+      "loss": 0.4809,
+      "step": 298
+    },
+    {
+      "epoch": 1.4248210023866348,
+      "grad_norm": 0.7033349702559853,
+      "learning_rate": 4.115747298359363e-07,
+      "loss": 0.464,
+      "step": 299
+    },
+    {
+      "epoch": 1.4295942720763724,
+      "grad_norm": 0.7437100788001662,
+      "learning_rate": 4.0519488335323415e-07,
+      "loss": 0.4851,
+      "step": 300
+    },
+    {
+      "epoch": 1.4343675417661097,
+      "grad_norm": 0.7732697984175376,
+      "learning_rate": 3.9885228385872806e-07,
+      "loss": 0.4594,
+      "step": 301
+    },
+    {
+      "epoch": 1.4391408114558473,
+      "grad_norm": 0.7940793070581448,
+      "learning_rate": 3.925473285291091e-07,
+      "loss": 0.4661,
+      "step": 302
+    },
+    {
+      "epoch": 1.4439140811455848,
+      "grad_norm": 0.7351909971969558,
+      "learning_rate": 3.862804121837733e-07,
+      "loss": 0.4757,
+      "step": 303
+    },
+    {
+      "epoch": 1.4486873508353222,
+      "grad_norm": 0.781207875542895,
+      "learning_rate": 3.8005192726009663e-07,
+      "loss": 0.4787,
+      "step": 304
+    },
+    {
+      "epoch": 1.4534606205250595,
+      "grad_norm": 0.7991516861553173,
+      "learning_rate": 3.738622637888608e-07,
+      "loss": 0.4668,
+      "step": 305
+    },
+    {
+      "epoch": 1.458233890214797,
+      "grad_norm": 0.8987252432386614,
+      "learning_rate": 3.677118093698278e-07,
+      "loss": 0.4606,
+      "step": 306
+    },
+    {
+      "epoch": 1.4630071599045347,
+      "grad_norm": 0.698103668533834,
+      "learning_rate": 3.61600949147472e-07,
+      "loss": 0.4683,
+      "step": 307
+    },
+    {
+      "epoch": 1.467780429594272,
+      "grad_norm": 0.7560261667555234,
+      "learning_rate": 3.5553006578685706e-07,
+      "loss": 0.4519,
+      "step": 308
+    },
+    {
+      "epoch": 1.4725536992840096,
+      "grad_norm": 0.7382407678980342,
+      "learning_rate": 3.494995394496778e-07,
+      "loss": 0.469,
+      "step": 309
+    },
+    {
+      "epoch": 1.477326968973747,
+      "grad_norm": 0.720898348204588,
+      "learning_rate": 3.435097477704517e-07,
+      "loss": 0.449,
+      "step": 310
+    },
+    {
+      "epoch": 1.4821002386634845,
+      "grad_norm": 0.7319822241837816,
+      "learning_rate": 3.3756106583287205e-07,
+      "loss": 0.4745,
+      "step": 311
+    },
+    {
+      "epoch": 1.4868735083532219,
+      "grad_norm": 0.7518826329514531,
+      "learning_rate": 3.316538661463204e-07,
+      "loss": 0.4918,
+      "step": 312
+    },
+    {
+      "epoch": 1.4916467780429594,
+      "grad_norm": 0.8013086574909619,
+      "learning_rate": 3.2578851862253796e-07,
+      "loss": 0.4846,
+      "step": 313
+    },
+    {
+      "epoch": 1.496420047732697,
+      "grad_norm": 0.7101861238945232,
+      "learning_rate": 3.199653905524654e-07,
+      "loss": 0.4604,
+      "step": 314
+    },
+    {
+      "epoch": 1.5011933174224343,
+      "grad_norm": 0.7204781171906866,
+      "learning_rate": 3.1418484658323806e-07,
+      "loss": 0.4772,
+      "step": 315
+    },
+    {
+      "epoch": 1.5059665871121717,
+      "grad_norm": 0.73033687450555,
+      "learning_rate": 3.0844724869535577e-07,
+      "loss": 0.468,
+      "step": 316
+    },
+    {
+      "epoch": 1.5107398568019093,
+      "grad_norm": 0.7700114197888783,
+      "learning_rate": 3.027529561800117e-07,
+      "loss": 0.4808,
+      "step": 317
+    },
+    {
+      "epoch": 1.5155131264916468,
+      "grad_norm": 0.8599415830432524,
+      "learning_rate": 2.971023256165983e-07,
+      "loss": 0.469,
+      "step": 318
+    },
+    {
+      "epoch": 1.5202863961813842,
+      "grad_norm": 0.7490557961852297,
+      "learning_rate": 2.9149571085037215e-07,
+      "loss": 0.4758,
+      "step": 319
+    },
+    {
+      "epoch": 1.5250596658711217,
+      "grad_norm": 0.6911043116400506,
+      "learning_rate": 2.8593346297030073e-07,
+      "loss": 0.4662,
+      "step": 320
+    },
+    {
+      "epoch": 1.5298329355608593,
+      "grad_norm": 0.7444306144257443,
+      "learning_rate": 2.804159302870751e-07,
+      "loss": 0.4638,
+      "step": 321
+    },
+    {
+      "epoch": 1.5346062052505967,
+      "grad_norm": 0.6930295325600317,
+      "learning_rate": 2.7494345831129837e-07,
+      "loss": 0.4584,
+      "step": 322
+    },
+    {
+      "epoch": 1.539379474940334,
+      "grad_norm": 0.7461580524158721,
+      "learning_rate": 2.6951638973185073e-07,
+      "loss": 0.4757,
+      "step": 323
+    },
+    {
+      "epoch": 1.5441527446300716,
+      "grad_norm": 0.7678530858976563,
+      "learning_rate": 2.64135064394428e-07,
+      "loss": 0.4807,
+      "step": 324
+    },
+    {
+      "epoch": 1.5489260143198091,
+      "grad_norm": 1.8144860255245707,
+      "learning_rate": 2.587998192802638e-07,
+      "loss": 0.4605,
+      "step": 325
+    },
+    {
+      "epoch": 1.5536992840095465,
+      "grad_norm": 0.7200319691236525,
+      "learning_rate": 2.5351098848502386e-07,
+      "loss": 0.474,
+      "step": 326
+    },
+    {
+      "epoch": 1.558472553699284,
+      "grad_norm": 0.7134577877268367,
+      "learning_rate": 2.482689031978872e-07,
+      "loss": 0.4715,
+      "step": 327
+    },
+    {
+      "epoch": 1.5632458233890216,
+      "grad_norm": 0.9468756981275396,
+      "learning_rate": 2.4307389168080606e-07,
+      "loss": 0.4656,
+      "step": 328
+    },
+    {
+      "epoch": 1.568019093078759,
+      "grad_norm": 0.6688722309384391,
+      "learning_rate": 2.3792627924795038e-07,
+      "loss": 0.4922,
+      "step": 329
+    },
+    {
+      "epoch": 1.5727923627684963,
+      "grad_norm": 0.7125789762828182,
+      "learning_rate": 2.3282638824533529e-07,
+      "loss": 0.4692,
+      "step": 330
+    },
+    {
+      "epoch": 1.577565632458234,
+      "grad_norm": 0.8844333458882234,
+      "learning_rate": 2.277745380306383e-07,
+      "loss": 0.4876,
+      "step": 331
+    },
+    {
+      "epoch": 1.5823389021479715,
+      "grad_norm": 0.7950308834961601,
+      "learning_rate": 2.227710449531971e-07,
+      "loss": 0.4918,
+      "step": 332
+    },
+    {
+      "epoch": 1.5871121718377088,
+      "grad_norm": 0.796382860942759,
+      "learning_rate": 2.178162223342035e-07,
+      "loss": 0.4641,
+      "step": 333
+    },
+    {
+      "epoch": 1.5918854415274462,
+      "grad_norm": 0.7285520770077796,
+      "learning_rate": 2.1291038044707965e-07,
+      "loss": 0.4661,
+      "step": 334
+    },
+    {
+      "epoch": 1.5966587112171837,
+      "grad_norm": 0.6921820001369808,
+      "learning_rate": 2.0805382649805225e-07,
+      "loss": 0.4681,
+      "step": 335
+    },
+    {
+      "epoch": 1.6014319809069213,
+      "grad_norm": 0.7552481890637776,
+      "learning_rate": 2.032468646069112e-07,
+      "loss": 0.4672,
+      "step": 336
+    },
+    {
+      "epoch": 1.6062052505966586,
+      "grad_norm": 0.7155745101307224,
+      "learning_rate": 1.9848979578796865e-07,
+      "loss": 0.4767,
+      "step": 337
+    },
+    {
+      "epoch": 1.6109785202863962,
+      "grad_norm": 0.6993076336434562,
+      "learning_rate": 1.937829179312076e-07,
+      "loss": 0.4822,
+      "step": 338
+    },
+    {
+      "epoch": 1.6157517899761338,
+      "grad_norm": 0.7530303728674003,
+      "learning_rate": 1.8912652578362853e-07,
+      "loss": 0.4709,
+      "step": 339
+    },
+    {
+      "epoch": 1.6205250596658711,
+      "grad_norm": 0.7510327363849882,
+      "learning_rate": 1.8452091093079215e-07,
+      "loss": 0.4604,
+      "step": 340
+    },
+    {
+      "epoch": 1.6252983293556085,
+      "grad_norm": 0.7282910633876013,
+      "learning_rate": 1.7996636177855928e-07,
+      "loss": 0.4984,
+      "step": 341
+    },
+    {
+      "epoch": 1.630071599045346,
+      "grad_norm": 0.7524297400825809,
+      "learning_rate": 1.75463163535033e-07,
+      "loss": 0.4823,
+      "step": 342
+    },
+    {
+      "epoch": 1.6348448687350836,
+      "grad_norm": 0.7049222733481684,
+      "learning_rate": 1.7101159819269583e-07,
+      "loss": 0.4635,
+      "step": 343
+    },
+    {
+      "epoch": 1.639618138424821,
+      "grad_norm": 1.1034453594616451,
+      "learning_rate": 1.6661194451075345e-07,
+      "loss": 0.4765,
+      "step": 344
+    },
+    {
+      "epoch": 1.6443914081145583,
+      "grad_norm": 0.83013391018154,
+      "learning_rate": 1.6226447799767772e-07,
+      "loss": 0.4533,
+      "step": 345
+    },
+    {
+      "epoch": 1.649164677804296,
+      "grad_norm": 2.858030289791699,
+      "learning_rate": 1.5796947089395475e-07,
+      "loss": 0.4691,
+      "step": 346
+    },
+    {
+      "epoch": 1.6539379474940334,
+      "grad_norm": 0.7332905568570133,
+      "learning_rate": 1.5372719215503582e-07,
+      "loss": 0.4544,
+      "step": 347
+    },
+    {
+      "epoch": 1.6587112171837708,
+      "grad_norm": 0.7481224605220782,
+      "learning_rate": 1.4953790743449702e-07,
+      "loss": 0.4806,
+      "step": 348
+    },
+    {
+      "epoch": 1.6634844868735084,
+      "grad_norm": 0.9099408876904721,
+      "learning_rate": 1.4540187906740241e-07,
+      "loss": 0.4569,
+      "step": 349
+    },
+    {
+      "epoch": 1.668257756563246,
+      "grad_norm": 0.6921320546034447,
+      "learning_rate": 1.4131936605387762e-07,
+      "loss": 0.4897,
+      "step": 350
+    },
+    {
+      "epoch": 1.6730310262529833,
+      "grad_norm": 0.7172188028374827,
+      "learning_rate": 1.3729062404289017e-07,
+      "loss": 0.4799,
+      "step": 351
+    },
+    {
+      "epoch": 1.6778042959427206,
+      "grad_norm": 0.7348308299387173,
+      "learning_rate": 1.3331590531624115e-07,
+      "loss": 0.4714,
+      "step": 352
+    },
+    {
+      "epoch": 1.6825775656324582,
+      "grad_norm": 0.7524117454719962,
+      "learning_rate": 1.2939545877276726e-07,
+      "loss": 0.4679,
+      "step": 353
+    },
+    {
+      "epoch": 1.6873508353221958,
+      "grad_norm": 0.7609980327732692,
+      "learning_rate": 1.25529529912754e-07,
+      "loss": 0.4678,
+      "step": 354
+    },
+    {
+      "epoch": 1.692124105011933,
+      "grad_norm": 0.7906234591099575,
+      "learning_rate": 1.2171836082256316e-07,
+      "loss": 0.4754,
+      "step": 355
+    },
+    {
+      "epoch": 1.6968973747016707,
+      "grad_norm": 0.7519337557814546,
+      "learning_rate": 1.1796219015947285e-07,
+      "loss": 0.4803,
+      "step": 356
+    },
+    {
+      "epoch": 1.7016706443914082,
+      "grad_norm": 0.6859134821445197,
+      "learning_rate": 1.1426125313673285e-07,
+      "loss": 0.4939,
+      "step": 357
+    },
+    {
+      "epoch": 1.7064439140811456,
+      "grad_norm": 0.8229493204752176,
+      "learning_rate": 1.1061578150883444e-07,
+      "loss": 0.4372,
+      "step": 358
+    },
+    {
+      "epoch": 1.711217183770883,
+      "grad_norm": 0.692317996696451,
+      "learning_rate": 1.070260035570002e-07,
+      "loss": 0.4792,
+      "step": 359
+    },
+    {
+      "epoch": 1.7159904534606205,
+      "grad_norm": 0.7390705342617898,
+      "learning_rate": 1.0349214407488571e-07,
+      "loss": 0.4719,
+      "step": 360
+    },
+    {
+      "epoch": 1.720763723150358,
+      "grad_norm": 0.7057263439063961,
+      "learning_rate": 1.000144243545058e-07,
+      "loss": 0.4724,
+      "step": 361
+    },
+    {
+      "epoch": 1.7255369928400954,
+      "grad_norm": 0.707795857913463,
+      "learning_rate": 9.659306217237517e-08,
+      "loss": 0.4717,
+      "step": 362
+    },
+    {
+      "epoch": 1.7303102625298328,
+      "grad_norm": 0.7912536951606031,
+      "learning_rate": 9.322827177587212e-08,
+      "loss": 0.4623,
+      "step": 363
+    },
+    {
+      "epoch": 1.7350835322195706,
+      "grad_norm": 0.746736598206851,
+      "learning_rate": 8.992026386982221e-08,
+      "loss": 0.4735,
+      "step": 364
+    },
+    {
+      "epoch": 1.739856801909308,
+      "grad_norm": 0.6948885657819285,
+      "learning_rate": 8.66692456033029e-08,
+      "loss": 0.4825,
+      "step": 365
+    },
+    {
+      "epoch": 1.7446300715990453,
+      "grad_norm": 0.7262491961744311,
+      "learning_rate": 8.347542055667311e-08,
+      "loss": 0.4699,
+      "step": 366
+    },
+    {
+      "epoch": 1.7494033412887828,
+      "grad_norm": 0.7863038143235231,
+      "learning_rate": 8.033898872882394e-08,
+      "loss": 0.4679,
+      "step": 367
+    },
+    {
+      "epoch": 1.7541766109785204,
+      "grad_norm": 0.6727626949269937,
+      "learning_rate": 7.726014652465507e-08,
+      "loss": 0.4421,
+      "step": 368
+    },
+    {
+      "epoch": 1.7589498806682577,
+      "grad_norm": 0.6867145980818331,
+      "learning_rate": 7.423908674277579e-08,
+      "loss": 0.4778,
+      "step": 369
+    },
+    {
+      "epoch": 1.763723150357995,
+      "grad_norm": 1.4213029472300538,
+      "learning_rate": 7.127599856343192e-08,
+      "loss": 0.4727,
+      "step": 370
+    },
+    {
+      "epoch": 1.7684964200477327,
+      "grad_norm": 0.692012937763345,
+      "learning_rate": 6.837106753665823e-08,
+      "loss": 0.4741,
+      "step": 371
+    },
+    {
+      "epoch": 1.7732696897374702,
+      "grad_norm": 0.7092148893859065,
+      "learning_rate": 6.552447557066109e-08,
+      "loss": 0.4697,
+      "step": 372
+    },
+    {
+      "epoch": 1.7780429594272076,
+      "grad_norm": 0.6973356829898804,
+      "learning_rate": 6.273640092042575e-08,
+      "loss": 0.4544,
+      "step": 373
+    },
+    {
+      "epoch": 1.7828162291169452,
+      "grad_norm": 1.5448551643686548,
+      "learning_rate": 6.000701817655474e-08,
+      "loss": 0.4523,
+      "step": 374
+    },
+    {
+      "epoch": 1.7875894988066827,
+      "grad_norm": 1.4827724692081619,
+      "learning_rate": 5.733649825433384e-08,
+      "loss": 0.4551,
+      "step": 375
+    },
+    {
+      "epoch": 1.79236276849642,
+      "grad_norm": 0.7790516793749164,
+      "learning_rate": 5.47250083830314e-08,
+      "loss": 0.494,
+      "step": 376
+    },
+    {
+      "epoch": 1.7971360381861574,
+      "grad_norm": 0.7365514384441436,
+      "learning_rate": 5.217271209542384e-08,
+      "loss": 0.4735,
+      "step": 377
+    },
+    {
+      "epoch": 1.801909307875895,
+      "grad_norm": 0.7707502808832377,
+      "learning_rate": 4.967976921755679e-08,
+      "loss": 0.4501,
+      "step": 378
+    },
+    {
+      "epoch": 1.8066825775656326,
+      "grad_norm": 0.7176835200739754,
+      "learning_rate": 4.724633585873627e-08,
+      "loss": 0.4686,
+      "step": 379
+    },
+    {
+      "epoch": 1.81145584725537,
+      "grad_norm": 0.6889468337016494,
+      "learning_rate": 4.487256440175291e-08,
+      "loss": 0.4771,
+      "step": 380
+    },
+    {
+      "epoch": 1.8162291169451072,
+      "grad_norm": 1.0649529564643607,
+      "learning_rate": 4.255860349334006e-08,
+      "loss": 0.4661,
+      "step": 381
+    },
+    {
+      "epoch": 1.8210023866348448,
+      "grad_norm": 1.1333041301606328,
+      "learning_rate": 4.030459803486464e-08,
+      "loss": 0.4606,
+      "step": 382
+    },
+    {
+      "epoch": 1.8257756563245824,
+      "grad_norm": 0.765268616008849,
+      "learning_rate": 3.811068917325444e-08,
+      "loss": 0.4442,
+      "step": 383
+    },
+    {
+      "epoch": 1.8305489260143197,
+      "grad_norm": 0.701547689578903,
+      "learning_rate": 3.5977014292158495e-08,
+      "loss": 0.4739,
+      "step": 384
+    },
+    {
+      "epoch": 1.8353221957040573,
+      "grad_norm": 0.7141975076446941,
+      "learning_rate": 3.3903707003344774e-08,
+      "loss": 0.4719,
+      "step": 385
+    },
+    {
+      "epoch": 1.8400954653937949,
+      "grad_norm": 0.6918753885495199,
+      "learning_rate": 3.189089713833226e-08,
+      "loss": 0.4772,
+      "step": 386
+    },
+    {
+      "epoch": 1.8448687350835322,
+      "grad_norm": 0.714964202433507,
+      "learning_rate": 2.9938710740262884e-08,
+      "loss": 0.4561,
+      "step": 387
+    },
+    {
+      "epoch": 1.8496420047732696,
+      "grad_norm": 0.7838822438811583,
+      "learning_rate": 2.8047270056005934e-08,
+      "loss": 0.4565,
+      "step": 388
+    },
+    {
+      "epoch": 1.8544152744630071,
+      "grad_norm": 0.7061577623995287,
+      "learning_rate": 2.6216693528505195e-08,
+      "loss": 0.4648,
+      "step": 389
+    },
+    {
+      "epoch": 1.8591885441527447,
+      "grad_norm": 0.9071757882196184,
+      "learning_rate": 2.4447095789360884e-08,
+      "loss": 0.4711,
+      "step": 390
+    },
+    {
+      "epoch": 1.863961813842482,
+      "grad_norm": 0.765845128347514,
+      "learning_rate": 2.2738587651651487e-08,
+      "loss": 0.4577,
+      "step": 391
+    },
+    {
+      "epoch": 1.8687350835322196,
+      "grad_norm": 0.7650600946027074,
+      "learning_rate": 2.109127610299466e-08,
+      "loss": 0.4679,
+      "step": 392
+    },
+    {
+      "epoch": 1.8735083532219572,
+      "grad_norm": 0.6957819402359949,
+      "learning_rate": 1.950526429884769e-08,
+      "loss": 0.4559,
+      "step": 393
+    },
+    {
+      "epoch": 1.8782816229116945,
+      "grad_norm": 0.7430172728751436,
+      "learning_rate": 1.7980651556048e-08,
+      "loss": 0.4732,
+      "step": 394
+    },
+    {
+      "epoch": 1.8830548926014319,
+      "grad_norm": 0.6767278663023139,
+      "learning_rate": 1.6517533346593226e-08,
+      "loss": 0.4758,
+      "step": 395
+    },
+    {
+      "epoch": 1.8878281622911695,
+      "grad_norm": 0.7619777582419104,
+      "learning_rate": 1.5116001291663462e-08,
+      "loss": 0.4932,
+      "step": 396
+    },
+    {
+      "epoch": 1.892601431980907,
+      "grad_norm": 0.790159743362526,
+      "learning_rate": 1.3776143155883491e-08,
+      "loss": 0.4558,
+      "step": 397
+    },
+    {
+      "epoch": 1.8973747016706444,
+      "grad_norm": 0.7261843559497824,
+      "learning_rate": 1.2498042841827317e-08,
+      "loss": 0.4595,
+      "step": 398
+    },
+    {
+      "epoch": 1.9021479713603817,
+      "grad_norm": 0.7017669980294373,
+      "learning_rate": 1.128178038476324e-08,
+      "loss": 0.4625,
+      "step": 399
+    },
+    {
+      "epoch": 1.9069212410501193,
+      "grad_norm": 0.6784318458229694,
+      "learning_rate": 1.0127431947643316e-08,
+      "loss": 0.4671,
+      "step": 400
+    },
+    {
+      "epoch": 1.9116945107398569,
+      "grad_norm": 0.675130035717692,
+      "learning_rate": 9.035069816332619e-09,
+      "loss": 0.464,
+      "step": 401
+    },
+    {
+      "epoch": 1.9164677804295942,
+      "grad_norm": 0.8221120490850481,
+      "learning_rate": 8.004762395083963e-09,
+      "loss": 0.4537,
+      "step": 402
+    },
+    {
+      "epoch": 1.9212410501193318,
+      "grad_norm": 0.7612136405972405,
+      "learning_rate": 7.036574202253343e-09,
+      "loss": 0.4914,
+      "step": 403
+    },
+    {
+      "epoch": 1.9260143198090693,
+      "grad_norm": 0.8291594902189451,
+      "learning_rate": 6.130565866260484e-09,
+      "loss": 0.4727,
+      "step": 404
+    },
+    {
+      "epoch": 1.9307875894988067,
+      "grad_norm": 1.0329364100399496,
+      "learning_rate": 5.286794121791782e-09,
+      "loss": 0.4767,
+      "step": 405
+    },
+    {
+      "epoch": 1.935560859188544,
+      "grad_norm": 0.8758229910700595,
+      "learning_rate": 4.5053118062478025e-09,
+      "loss": 0.4501,
+      "step": 406
+    },
+    {
+      "epoch": 1.9403341288782816,
+      "grad_norm": 0.7067697193260255,
+      "learning_rate": 3.786167856434375e-09,
+      "loss": 0.4747,
+      "step": 407
+    },
+    {
+      "epoch": 1.9451073985680192,
+      "grad_norm": 0.7459961970155857,
+      "learning_rate": 3.1294073054987102e-09,
+      "loss": 0.4605,
+      "step": 408
+    },
+    {
+      "epoch": 1.9498806682577565,
+      "grad_norm": 0.7585534385150827,
+      "learning_rate": 2.5350712801084363e-09,
+      "loss": 0.4528,
+      "step": 409
+    },
+    {
+      "epoch": 1.9546539379474939,
+      "grad_norm": 0.6767868247269999,
+      "learning_rate": 2.003196997877099e-09,
+      "loss": 0.4585,
+      "step": 410
+    },
+    {
+      "epoch": 1.9594272076372317,
+      "grad_norm": 0.7126370902337825,
+      "learning_rate": 1.5338177650332517e-09,
+      "loss": 0.4591,
+      "step": 411
+    },
+    {
+      "epoch": 1.964200477326969,
+      "grad_norm": 0.7172728813954358,
+      "learning_rate": 1.1269629743346777e-09,
+      "loss": 0.4589,
+      "step": 412
+    },
+    {
+      "epoch": 1.9689737470167064,
+      "grad_norm": 0.8158860106123756,
+      "learning_rate": 7.826581032279734e-10,
+      "loss": 0.4601,
+      "step": 413
+    },
+    {
+      "epoch": 1.973747016706444,
+      "grad_norm": 0.8261699459606863,
+      "learning_rate": 5.00924712252937e-10,
+      "loss": 0.4731,
+      "step": 414
+    },
+    {
+      "epoch": 1.9785202863961815,
+      "grad_norm": 0.7168072767819187,
+      "learning_rate": 2.8178044369286945e-10,
+      "loss": 0.4657,
+      "step": 415
+    },
+    {
+      "epoch": 1.9832935560859188,
+      "grad_norm": 0.6783006404134123,
+      "learning_rate": 1.2523902046934763e-10,
+      "loss": 0.452,
+      "step": 416
+    },
+    {
+      "epoch": 1.9880668257756562,
+      "grad_norm": 0.7080089156985594,
+      "learning_rate": 3.131024528302273e-11,
+      "loss": 0.4737,
+      "step": 417
+    },
+    {
+      "epoch": 1.9928400954653938,
+      "grad_norm": 0.7031897431837284,
+      "learning_rate": 0.0,
+      "loss": 0.4817,
+      "step": 418
+    },
+    {
+      "epoch": 1.9928400954653938,
+      "step": 418,
+      "total_flos": 3166299160051712.0,
+      "train_loss": 0.5405693022828353,
+      "train_runtime": 17864.3337,
+      "train_samples_per_second": 2.995,
+      "train_steps_per_second": 0.023
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 418,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3166299160051712.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5a3ff52c13e1e21835c234d6f155545088f6f64f5c9eb422eb24377b00b0320
+size 7672