grail_v3c_343_b64_sft_e3_ck640

.gitattributes

@@ -75,3 +75,4 @@ her_nemi_graph_fcur_rw_b64_sft_e3_ck300/tokenizer.json filter=lfs diff=lfs merge
 grail_v1_343_b64_sft_e3_ck300/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 grail_v2op_343_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 grail_v2opv_343_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+grail_v3c_343_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text

grail_v3c_343_b64_sft_e3_ck640/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 0.9946576007770762,
+    "eval_loss": 0.13847553730010986,
+    "eval_runtime": 158.6616,
+    "eval_samples_per_second": 13.664,
+    "eval_steps_per_second": 3.416,
+    "total_flos": 274201835470848.0,
+    "train_loss": 0.19095446057617665,
+    "train_runtime": 7793.1309,
+    "train_samples_per_second": 5.256,
+    "train_steps_per_second": 0.082
+}

grail_v3c_343_b64_sft_e3_ck640/chat_template.jinja

@@ -0,0 +1,120 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' }}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- for message in messages %}
+    {%- if message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- 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 %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content_item in message.content %}
+                {%- if 'text' in content_item %}
+                    {{- content_item.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and message.content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- 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 %}
+        {%- endif %}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

grail_v3c_343_b64_sft_e3_ck640/config.json

@@ -0,0 +1,70 @@
+{
+  "architectures": [
+    "Qwen3VLForConditionalGeneration"
+  ],
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "hidden_size": 4096,
+  "image_token_id": 151655,
+  "model_type": "qwen3_vl",
+  "pad_token_id": 151643,
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "dtype": "bfloat16",
+    "eos_token_id": 151645,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "initializer_range": 0.02,
+    "intermediate_size": 12288,
+    "max_position_embeddings": 262144,
+    "model_type": "qwen3_vl_text",
+    "num_attention_heads": 32,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 8,
+    "pad_token_id": null,
+    "rms_norm_eps": 1e-06,
+    "rope_parameters": {
+      "mrope_interleaved": true,
+      "mrope_section": [
+        24,
+        20,
+        20
+      ],
+      "rope_theta": 5000000,
+      "rope_type": "default"
+    },
+    "use_cache": false,
+    "vocab_size": 151936
+  },
+  "tie_word_embeddings": false,
+  "transformers_version": "5.8.0",
+  "use_cache": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "deepstack_visual_indexes": [
+      8,
+      16,
+      24
+    ],
+    "depth": 27,
+    "dtype": "bfloat16",
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 4304,
+    "model_type": "qwen3_vl_vision",
+    "num_heads": 16,
+    "num_position_embeddings": 2304,
+    "out_hidden_size": 4096,
+    "patch_size": 16,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652
+}

grail_v3c_343_b64_sft_e3_ck640/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 0.9946576007770762,
+    "eval_loss": 0.13847553730010986,
+    "eval_runtime": 158.6616,
+    "eval_samples_per_second": 13.664,
+    "eval_steps_per_second": 3.416
+}

grail_v3c_343_b64_sft_e3_ck640/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.0,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "5.8.0"
+}

grail_v3c_343_b64_sft_e3_ck640/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6248c31ba8f2870d4f7256dfbf56f9f133b11df6f92dc196671ff659714a8732
+size 17534340584

grail_v3c_343_b64_sft_e3_ck640/processor_config.json

@@ -0,0 +1,60 @@
+{
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "Qwen2VLImageProcessor",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "merge_size": 2,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "longest_edge": 16777216,
+      "shortest_edge": 65536
+    },
+    "temporal_patch_size": 2
+  },
+  "processor_class": "Qwen3VLProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "fps": 2,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_frames": 768,
+    "merge_size": 2,
+    "min_frames": 4,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "size": {
+      "longest_edge": 25165824,
+      "shortest_edge": 4096
+    },
+    "temporal_patch_size": 2,
+    "video_processor_type": "Qwen3VLVideoProcessor"
+  }
+}

grail_v3c_343_b64_sft_e3_ck640/sft.yaml

@@ -0,0 +1,58 @@
+# LLaMA-Factory full-parameter SFT of Qwen3-VL-8B-Instruct on one HER-final dataset.
+# Four L40S (46 GB), one run at a time: DeepSpeed ZeRO-3 with optimizer offload to host
+# RAM (~130 GB per run, so two runs do not fit in 372 GB); the vision tower and projector
+# stay frozen (LLaMA-Factory's default for VLMs), the LLM trains.
+# Filled per run by launch_sft.sh: grail_v3c_343_sft_e3, exps/sft/grail_v3c_343_sft_e3_k30449, /scratch/kwc5354/viewagent/viewagent_private/exps/sft_runs/grail_v3c_343_b64_sft_e3.
+### model
+model_name_or_path: /scratch/kwc5354/viewagent/viewagent_private/exps/models/graph_nemi_sft_e2
+template: qwen3_vl_nothink
+trust_remote_code: true
+image_max_pixels: 65536          # frames are 256x256 already; never upscale
+image_min_pixels: 4096
+
+### method
+stage: sft
+do_train: true
+finetuning_type: full
+freeze_vision_tower: true
+freeze_multi_modal_projector: true
+freeze_language_model: false
+deepspeed: /scratch/kwc5354/viewagent/viewagent_private/LLaMA-Factory/examples/deepspeed/ds_z3_config.json
+
+### dataset
+dataset: grail_v3c_343_sft_e3
+dataset_dir: exps/sft/grail_v3c_343_sft_e3_k30449
+media_dir: /scratch/kwc5354/viewagent/viewagent_private
+cutoff_len: 8192
+overwrite_cache: true
+preprocessing_num_workers: 16
+dataloader_num_workers: 4
+
+### output
+output_dir: /scratch/kwc5354/viewagent/viewagent_private/exps/sft_runs/grail_v3c_343_b64_sft_e3
+logging_steps: 10
+save_strategy: steps
+save_steps: 320
+save_total_limit: 100
+save_only_model: true
+plot_loss: true
+overwrite_output_dir: true
+report_to: wandb
+
+### train
+per_device_train_batch_size: 1
+gradient_accumulation_steps: 16
+max_steps: 640
+learning_rate: 1.0e-5
+lr_scheduler_type: cosine
+warmup_ratio: 0.03
+weight_decay: 0.01
+bf16: true
+gradient_checkpointing: true
+ddp_timeout: 180000000
+seed: 0
+val_size: 0.05
+eval_strategy: steps
+eval_steps: 320
+per_device_eval_batch_size: 1
+run_name: grail_v3c_343_b64_sft_e3

grail_v3c_343_b64_sft_e3_ck640/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650

grail_v3c_343_b64_sft_e3_ck640/tokenizer_config.json

@@ -0,0 +1,32 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_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|>"
+  ],
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "processor_class": "Qwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

grail_v3c_343_b64_sft_e3_ck640/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9946576007770762,
+    "total_flos": 274201835470848.0,
+    "train_loss": 0.19095446057617665,
+    "train_runtime": 7793.1309,
+    "train_samples_per_second": 5.256,
+    "train_steps_per_second": 0.082
+}

grail_v3c_343_b64_sft_e3_ck640/trainer_state.json

@@ -0,0 +1,507 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9946576007770762,
+  "eval_steps": 320,
+  "global_step": 640,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.015541525012141816,
+      "grad_norm": 6.873813512143041,
+      "learning_rate": 4.5e-06,
+      "loss": 1.2347082138061523,
+      "step": 10
+    },
+    {
+      "epoch": 0.03108305002428363,
+      "grad_norm": 0.8089582180325291,
+      "learning_rate": 9.5e-06,
+      "loss": 0.4094590663909912,
+      "step": 20
+    },
+    {
+      "epoch": 0.04662457503642545,
+      "grad_norm": 0.5925653601882511,
+      "learning_rate": 9.994801642711945e-06,
+      "loss": 0.26471145153045655,
+      "step": 30
+    },
+    {
+      "epoch": 0.06216610004856726,
+      "grad_norm": 0.5768531619789975,
+      "learning_rate": 9.976845889479778e-06,
+      "loss": 0.245318603515625,
+      "step": 40
+    },
+    {
+      "epoch": 0.07770762506070908,
+      "grad_norm": 0.5950301295463467,
+      "learning_rate": 9.946114644227765e-06,
+      "loss": 0.23632307052612306,
+      "step": 50
+    },
+    {
+      "epoch": 0.0932491500728509,
+      "grad_norm": 0.5856238751318771,
+      "learning_rate": 9.90268679362172e-06,
+      "loss": 0.23225913047790528,
+      "step": 60
+    },
+    {
+      "epoch": 0.10879067508499271,
+      "grad_norm": 0.4992283350673458,
+      "learning_rate": 9.846673816331297e-06,
+      "loss": 0.22472524642944336,
+      "step": 70
+    },
+    {
+      "epoch": 0.12433220009713453,
+      "grad_norm": 0.4150816971416224,
+      "learning_rate": 9.778219496865864e-06,
+      "loss": 0.2197582721710205,
+      "step": 80
+    },
+    {
+      "epoch": 0.13987372510927634,
+      "grad_norm": 0.669325010684418,
+      "learning_rate": 9.6974995564817e-06,
+      "loss": 0.217682409286499,
+      "step": 90
+    },
+    {
+      "epoch": 0.15541525012141816,
+      "grad_norm": 0.5583382650961016,
+      "learning_rate": 9.604721202108084e-06,
+      "loss": 0.21828813552856446,
+      "step": 100
+    },
+    {
+      "epoch": 0.17095677513355997,
+      "grad_norm": 0.5249121276480293,
+      "learning_rate": 9.500122594450056e-06,
+      "loss": 0.20258612632751466,
+      "step": 110
+    },
+    {
+      "epoch": 0.1864983001457018,
+      "grad_norm": 0.5361763174248738,
+      "learning_rate": 9.383972236633308e-06,
+      "loss": 0.20349197387695311,
+      "step": 120
+    },
+    {
+      "epoch": 0.2020398251578436,
+      "grad_norm": 0.6739121237642136,
+      "learning_rate": 9.256568284960513e-06,
+      "loss": 0.2077998161315918,
+      "step": 130
+    },
+    {
+      "epoch": 0.21758135016998542,
+      "grad_norm": 0.6470623123002628,
+      "learning_rate": 9.118237783548376e-06,
+      "loss": 0.20095677375793458,
+      "step": 140
+    },
+    {
+      "epoch": 0.23312287518212724,
+      "grad_norm": 0.6197494941901578,
+      "learning_rate": 8.969335824810075e-06,
+      "loss": 0.20235679149627686,
+      "step": 150
+    },
+    {
+      "epoch": 0.24866440019426905,
+      "grad_norm": 0.44126621178814734,
+      "learning_rate": 8.810244637938117e-06,
+      "loss": 0.20273089408874512,
+      "step": 160
+    },
+    {
+      "epoch": 0.2642059252064109,
+      "grad_norm": 0.4364999057459828,
+      "learning_rate": 8.641372607727503e-06,
+      "loss": 0.1848066568374634,
+      "step": 170
+    },
+    {
+      "epoch": 0.2797474502185527,
+      "grad_norm": 0.5739465336014172,
+      "learning_rate": 8.463153226257805e-06,
+      "loss": 0.17958765029907225,
+      "step": 180
+    },
+    {
+      "epoch": 0.29528897523069453,
+      "grad_norm": 0.447043998016975,
+      "learning_rate": 8.276043980125213e-06,
+      "loss": 0.1867802619934082,
+      "step": 190
+    },
+    {
+      "epoch": 0.3108305002428363,
+      "grad_norm": 0.5013400159299765,
+      "learning_rate": 8.080525176081006e-06,
+      "loss": 0.17537325620651245,
+      "step": 200
+    },
+    {
+      "epoch": 0.32637202525497816,
+      "grad_norm": 0.48026659753382955,
+      "learning_rate": 7.877098708091017e-06,
+      "loss": 0.17553436756134033,
+      "step": 210
+    },
+    {
+      "epoch": 0.34191355026711995,
+      "grad_norm": 0.45099553327060826,
+      "learning_rate": 7.666286768981012e-06,
+      "loss": 0.17689459323883056,
+      "step": 220
+    },
+    {
+      "epoch": 0.3574550752792618,
+      "grad_norm": 0.5336809758293185,
+      "learning_rate": 7.448630509975199e-06,
+      "loss": 0.17833077907562256,
+      "step": 230
+    },
+    {
+      "epoch": 0.3729966002914036,
+      "grad_norm": 0.4497088799291852,
+      "learning_rate": 7.224688651568797e-06,
+      "loss": 0.1736297845840454,
+      "step": 240
+    },
+    {
+      "epoch": 0.3885381253035454,
+      "grad_norm": 0.4659985211788695,
+      "learning_rate": 6.995036049300539e-06,
+      "loss": 0.16642980575561522,
+      "step": 250
+    },
+    {
+      "epoch": 0.4040796503156872,
+      "grad_norm": 0.3765033131259701,
+      "learning_rate": 6.7602622181067425e-06,
+      "loss": 0.16380847692489625,
+      "step": 260
+    },
+    {
+      "epoch": 0.41962117532782905,
+      "grad_norm": 0.6184076455577017,
+      "learning_rate": 6.520969819044933e-06,
+      "loss": 0.16328957080841064,
+      "step": 270
+    },
+    {
+      "epoch": 0.43516270033997084,
+      "grad_norm": 0.4003754824907938,
+      "learning_rate": 6.277773112271545e-06,
+      "loss": 0.16004817485809325,
+      "step": 280
+    },
+    {
+      "epoch": 0.4507042253521127,
+      "grad_norm": 0.4237280412111839,
+      "learning_rate": 6.031296380244915e-06,
+      "loss": 0.16191138029098512,
+      "step": 290
+    },
+    {
+      "epoch": 0.4662457503642545,
+      "grad_norm": 0.5467452799279873,
+      "learning_rate": 5.782172325201155e-06,
+      "loss": 0.16488949060440064,
+      "step": 300
+    },
+    {
+      "epoch": 0.4817872753763963,
+      "grad_norm": 0.4838907196580856,
+      "learning_rate": 5.531040445016604e-06,
+      "loss": 0.1619861602783203,
+      "step": 310
+    },
+    {
+      "epoch": 0.4973288003885381,
+      "grad_norm": 0.43764271111720604,
+      "learning_rate": 5.278545391625931e-06,
+      "loss": 0.16924021244049073,
+      "step": 320
+    },
+    {
+      "epoch": 0.4973288003885381,
+      "eval_loss": 0.16171900928020477,
+      "eval_runtime": 159.0763,
+      "eval_samples_per_second": 13.629,
+      "eval_steps_per_second": 3.407,
+      "step": 320
+    },
+    {
+      "epoch": 0.5128703254006799,
+      "grad_norm": 0.4789592734942593,
+      "learning_rate": 5.025335316209883e-06,
+      "loss": 0.16862633228302001,
+      "step": 330
+    },
+    {
+      "epoch": 0.5284118504128218,
+      "grad_norm": 0.543331328026724,
+      "learning_rate": 4.7720602054004865e-06,
+      "loss": 0.17022294998168946,
+      "step": 340
+    },
+    {
+      "epoch": 0.5439533754249636,
+      "grad_norm": 0.4117518062078225,
+      "learning_rate": 4.51937021277469e-06,
+      "loss": 0.15646584033966066,
+      "step": 350
+    },
+    {
+      "epoch": 0.5594949004371054,
+      "grad_norm": 0.44400358181103733,
+      "learning_rate": 4.267913989919457e-06,
+      "loss": 0.15944695472717285,
+      "step": 360
+    },
+    {
+      "epoch": 0.5750364254492472,
+      "grad_norm": 0.5021899616179507,
+      "learning_rate": 4.018337021352447e-06,
+      "loss": 0.15816211700439453,
+      "step": 370
+    },
+    {
+      "epoch": 0.5905779504613891,
+      "grad_norm": 0.4507626689297711,
+      "learning_rate": 3.7712799675725105e-06,
+      "loss": 0.1573774814605713,
+      "step": 380
+    },
+    {
+      "epoch": 0.6061194754735308,
+      "grad_norm": 0.5448193114508497,
+      "learning_rate": 3.5273770204933814e-06,
+      "loss": 0.15301510095596313,
+      "step": 390
+    },
+    {
+      "epoch": 0.6216610004856726,
+      "grad_norm": 0.5751067813417141,
+      "learning_rate": 3.2872542754821334e-06,
+      "loss": 0.155106782913208,
+      "step": 400
+    },
+    {
+      "epoch": 0.6372025254978144,
+      "grad_norm": 0.5633254174624588,
+      "learning_rate": 3.051528124181342e-06,
+      "loss": 0.1514876365661621,
+      "step": 410
+    },
+    {
+      "epoch": 0.6527440505099563,
+      "grad_norm": 0.6415027501092633,
+      "learning_rate": 2.8208036722405895e-06,
+      "loss": 0.15055824518203736,
+      "step": 420
+    },
+    {
+      "epoch": 0.6682855755220981,
+      "grad_norm": 0.6959980525960555,
+      "learning_rate": 2.595673186018929e-06,
+      "loss": 0.15428953170776366,
+      "step": 430
+    },
+    {
+      "epoch": 0.6838271005342399,
+      "grad_norm": 0.5083436398289134,
+      "learning_rate": 2.376714572245651e-06,
+      "loss": 0.14965728521347046,
+      "step": 440
+    },
+    {
+      "epoch": 0.6993686255463817,
+      "grad_norm": 0.5368814024450721,
+      "learning_rate": 2.1644898945419814e-06,
+      "loss": 0.14955058097839355,
+      "step": 450
+    },
+    {
+      "epoch": 0.7149101505585236,
+      "grad_norm": 0.4438281234284902,
+      "learning_rate": 1.959543930611835e-06,
+      "loss": 0.14229865074157716,
+      "step": 460
+    },
+    {
+      "epoch": 0.7304516755706654,
+      "grad_norm": 0.5700233938895723,
+      "learning_rate": 1.7624027738052669e-06,
+      "loss": 0.1536567449569702,
+      "step": 470
+    },
+    {
+      "epoch": 0.7459932005828072,
+      "grad_norm": 0.6295234701556583,
+      "learning_rate": 1.5735724826443788e-06,
+      "loss": 0.1456534147262573,
+      "step": 480
+    },
+    {
+      "epoch": 0.7615347255949491,
+      "grad_norm": 0.5635171221129934,
+      "learning_rate": 1.3935377817783585e-06,
+      "loss": 0.15256741046905517,
+      "step": 490
+    },
+    {
+      "epoch": 0.7770762506070908,
+      "grad_norm": 0.7548345737179791,
+      "learning_rate": 1.2227608177022522e-06,
+      "loss": 0.14643117189407348,
+      "step": 500
+    },
+    {
+      "epoch": 0.7926177756192326,
+      "grad_norm": 0.5423243634298157,
+      "learning_rate": 1.0616799724335387e-06,
+      "loss": 0.1445760726928711,
+      "step": 510
+    },
+    {
+      "epoch": 0.8081593006313744,
+      "grad_norm": 0.7583601351103639,
+      "learning_rate": 9.107087381917878e-07,
+      "loss": 0.1481918215751648,
+      "step": 520
+    },
+    {
+      "epoch": 0.8237008256435163,
+      "grad_norm": 0.6032896652449088,
+      "learning_rate": 7.702346559700502e-07,
+      "loss": 0.1419147253036499,
+      "step": 530
+    },
+    {
+      "epoch": 0.8392423506556581,
+      "grad_norm": 0.5902353604971191,
+      "learning_rate": 6.406183207226973e-07,
+      "loss": 0.14105987548828125,
+      "step": 540
+    },
+    {
+      "epoch": 0.8547838756677999,
+      "grad_norm": 0.664744621008367,
+      "learning_rate": 5.22192455723336e-07,
+      "loss": 0.15008857250213622,
+      "step": 550
+    },
+    {
+      "epoch": 0.8703254006799417,
+      "grad_norm": 0.6157997818656803,
+      "learning_rate": 4.152610584689443e-07,
+      "loss": 0.1405224919319153,
+      "step": 560
+    },
+    {
+      "epoch": 0.8858669256920836,
+      "grad_norm": 0.6280075692660876,
+      "learning_rate": 3.200986203226536e-07,
+      "loss": 0.1380108952522278,
+      "step": 570
+    },
+    {
+      "epoch": 0.9014084507042254,
+      "grad_norm": 0.6021314343670503,
+      "learning_rate": 2.3694942189835423e-07,
+      "loss": 0.13832625150680541,
+      "step": 580
+    },
+    {
+      "epoch": 0.9169499757163672,
+      "grad_norm": 0.6025902230879387,
+      "learning_rate": 1.6602690599586747e-07,
+      "loss": 0.14463876485824584,
+      "step": 590
+    },
+    {
+      "epoch": 0.932491500728509,
+      "grad_norm": 0.6487284330633997,
+      "learning_rate": 1.0751312969633665e-07,
+      "loss": 0.14111080169677734,
+      "step": 600
+    },
+    {
+      "epoch": 0.9480330257406508,
+      "grad_norm": 0.6468904216654519,
+      "learning_rate": 6.15582970243117e-08,
+      "loss": 0.1340947151184082,
+      "step": 610
+    },
+    {
+      "epoch": 0.9635745507527926,
+      "grad_norm": 0.6496852146596152,
+      "learning_rate": 2.828037337617273e-08,
+      "loss": 0.14275397062301637,
+      "step": 620
+    },
+    {
+      "epoch": 0.9791160757649344,
+      "grad_norm": 0.6798020315742438,
+      "learning_rate": 7.764782704649532e-09,
+      "loss": 0.13512914180755614,
+      "step": 630
+    },
+    {
+      "epoch": 0.9946576007770762,
+      "grad_norm": 0.6039107316790304,
+      "learning_rate": 6.418823675857777e-11,
+      "loss": 0.1403965473175049,
+      "step": 640
+    },
+    {
+      "epoch": 0.9946576007770762,
+      "eval_loss": 0.13847553730010986,
+      "eval_runtime": 159.1511,
+      "eval_samples_per_second": 13.622,
+      "eval_steps_per_second": 3.406,
+      "step": 640
+    },
+    {
+      "epoch": 0.9946576007770762,
+      "step": 640,
+      "total_flos": 274201835470848.0,
+      "train_loss": 0.19095446057617665,
+      "train_runtime": 7793.1309,
+      "train_samples_per_second": 5.256,
+      "train_steps_per_second": 0.082
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 640,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 320,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 274201835470848.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}