JamesK2W commited on
Commit
38346d2
·
verified ·
1 Parent(s): e90dab9

grail_v2op_fcur_frombase_b64_sft_e0_ck640

Browse files
.gitattributes CHANGED
@@ -82,3 +82,4 @@ grail_v5_fcur_op48k_b64_sft_e3_ck800/tokenizer.json filter=lfs diff=lfs merge=lf
82
  grail_v2op_fcur_long_b64_sft_e3_ck1280/tokenizer.json filter=lfs diff=lfs merge=lfs -text
83
  grail_v2op_fcur_sr_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
84
  grail_v2op_fcurlong_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
82
  grail_v2op_fcur_long_b64_sft_e3_ck1280/tokenizer.json filter=lfs diff=lfs merge=lfs -text
83
  grail_v2op_fcur_sr_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
84
  grail_v2op_fcurlong_b64_sft_e3_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
85
+ grail_v2op_fcur_frombase_b64_sft_e0_ck640/tokenizer.json filter=lfs diff=lfs merge=lfs -text
grail_v2op_fcur_frombase_b64_sft_e0_ck640/all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.007877116975187,
3
+ "eval_loss": 0.04447107017040253,
4
+ "eval_runtime": 154.0801,
5
+ "eval_samples_per_second": 13.876,
6
+ "eval_steps_per_second": 3.472,
7
+ "total_flos": 279869738713088.0,
8
+ "train_loss": 0.10761516890488565,
9
+ "train_runtime": 7452.7238,
10
+ "train_samples_per_second": 5.496,
11
+ "train_steps_per_second": 0.086
12
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/chat_template.jinja ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {%- if messages[0].content is string %}
5
+ {{- messages[0].content }}
6
+ {%- else %}
7
+ {%- for content in messages[0].content %}
8
+ {%- if 'text' in content %}
9
+ {{- content.text }}
10
+ {%- endif %}
11
+ {%- endfor %}
12
+ {%- endif %}
13
+ {{- '\n\n' }}
14
+ {%- endif %}
15
+ {{- "# 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>" }}
16
+ {%- for tool in tools %}
17
+ {{- "\n" }}
18
+ {{- tool | tojson }}
19
+ {%- endfor %}
20
+ {{- "\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" }}
21
+ {%- else %}
22
+ {%- if messages[0].role == 'system' %}
23
+ {{- '<|im_start|>system\n' }}
24
+ {%- if messages[0].content is string %}
25
+ {{- messages[0].content }}
26
+ {%- else %}
27
+ {%- for content in messages[0].content %}
28
+ {%- if 'text' in content %}
29
+ {{- content.text }}
30
+ {%- endif %}
31
+ {%- endfor %}
32
+ {%- endif %}
33
+ {{- '<|im_end|>\n' }}
34
+ {%- endif %}
35
+ {%- endif %}
36
+ {%- set image_count = namespace(value=0) %}
37
+ {%- set video_count = namespace(value=0) %}
38
+ {%- for message in messages %}
39
+ {%- if message.role == "user" %}
40
+ {{- '<|im_start|>' + message.role + '\n' }}
41
+ {%- if message.content is string %}
42
+ {{- message.content }}
43
+ {%- else %}
44
+ {%- for content in message.content %}
45
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
46
+ {%- set image_count.value = image_count.value + 1 %}
47
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
48
+ <|vision_start|><|image_pad|><|vision_end|>
49
+ {%- elif content.type == 'video' or 'video' in content %}
50
+ {%- set video_count.value = video_count.value + 1 %}
51
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
52
+ <|vision_start|><|video_pad|><|vision_end|>
53
+ {%- elif 'text' in content %}
54
+ {{- content.text }}
55
+ {%- endif %}
56
+ {%- endfor %}
57
+ {%- endif %}
58
+ {{- '<|im_end|>\n' }}
59
+ {%- elif message.role == "assistant" %}
60
+ {{- '<|im_start|>' + message.role + '\n' }}
61
+ {%- if message.content is string %}
62
+ {{- message.content }}
63
+ {%- else %}
64
+ {%- for content_item in message.content %}
65
+ {%- if 'text' in content_item %}
66
+ {{- content_item.text }}
67
+ {%- endif %}
68
+ {%- endfor %}
69
+ {%- endif %}
70
+ {%- if message.tool_calls %}
71
+ {%- for tool_call in message.tool_calls %}
72
+ {%- if (loop.first and message.content) or (not loop.first) %}
73
+ {{- '\n' }}
74
+ {%- endif %}
75
+ {%- if tool_call.function %}
76
+ {%- set tool_call = tool_call.function %}
77
+ {%- endif %}
78
+ {{- '<tool_call>\n{"name": "' }}
79
+ {{- tool_call.name }}
80
+ {{- '", "arguments": ' }}
81
+ {%- if tool_call.arguments is string %}
82
+ {{- tool_call.arguments }}
83
+ {%- else %}
84
+ {{- tool_call.arguments | tojson }}
85
+ {%- endif %}
86
+ {{- '}\n</tool_call>' }}
87
+ {%- endfor %}
88
+ {%- endif %}
89
+ {{- '<|im_end|>\n' }}
90
+ {%- elif message.role == "tool" %}
91
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
92
+ {{- '<|im_start|>user' }}
93
+ {%- endif %}
94
+ {{- '\n<tool_response>\n' }}
95
+ {%- if message.content is string %}
96
+ {{- message.content }}
97
+ {%- else %}
98
+ {%- for content in message.content %}
99
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
100
+ {%- set image_count.value = image_count.value + 1 %}
101
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
102
+ <|vision_start|><|image_pad|><|vision_end|>
103
+ {%- elif content.type == 'video' or 'video' in content %}
104
+ {%- set video_count.value = video_count.value + 1 %}
105
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
106
+ <|vision_start|><|video_pad|><|vision_end|>
107
+ {%- elif 'text' in content %}
108
+ {{- content.text }}
109
+ {%- endif %}
110
+ {%- endfor %}
111
+ {%- endif %}
112
+ {{- '\n</tool_response>' }}
113
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
114
+ {{- '<|im_end|>\n' }}
115
+ {%- endif %}
116
+ {%- endif %}
117
+ {%- endfor %}
118
+ {%- if add_generation_prompt %}
119
+ {{- '<|im_start|>assistant\n' }}
120
+ {%- endif %}
grail_v2op_fcur_frombase_b64_sft_e0_ck640/config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "bos_token_id": null,
6
+ "dtype": "bfloat16",
7
+ "eos_token_id": 151645,
8
+ "hidden_size": 4096,
9
+ "image_token_id": 151655,
10
+ "model_type": "qwen3_vl",
11
+ "pad_token_id": 151643,
12
+ "text_config": {
13
+ "attention_bias": false,
14
+ "attention_dropout": 0.0,
15
+ "bos_token_id": 151643,
16
+ "dtype": "bfloat16",
17
+ "eos_token_id": 151645,
18
+ "head_dim": 128,
19
+ "hidden_act": "silu",
20
+ "hidden_size": 4096,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 12288,
23
+ "max_position_embeddings": 262144,
24
+ "model_type": "qwen3_vl_text",
25
+ "num_attention_heads": 32,
26
+ "num_hidden_layers": 36,
27
+ "num_key_value_heads": 8,
28
+ "pad_token_id": null,
29
+ "rms_norm_eps": 1e-06,
30
+ "rope_parameters": {
31
+ "mrope_interleaved": true,
32
+ "mrope_section": [
33
+ 24,
34
+ 20,
35
+ 20
36
+ ],
37
+ "rope_theta": 5000000,
38
+ "rope_type": "default"
39
+ },
40
+ "use_cache": false,
41
+ "vocab_size": 151936
42
+ },
43
+ "tie_word_embeddings": false,
44
+ "transformers_version": "5.8.0",
45
+ "use_cache": false,
46
+ "video_token_id": 151656,
47
+ "vision_config": {
48
+ "deepstack_visual_indexes": [
49
+ 8,
50
+ 16,
51
+ 24
52
+ ],
53
+ "depth": 27,
54
+ "dtype": "bfloat16",
55
+ "hidden_act": "gelu_pytorch_tanh",
56
+ "hidden_size": 1152,
57
+ "in_channels": 3,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4304,
60
+ "model_type": "qwen3_vl_vision",
61
+ "num_heads": 16,
62
+ "num_position_embeddings": 2304,
63
+ "out_hidden_size": 4096,
64
+ "patch_size": 16,
65
+ "spatial_merge_size": 2,
66
+ "temporal_patch_size": 2
67
+ },
68
+ "vision_end_token_id": 151653,
69
+ "vision_start_token_id": 151652
70
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/eval_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.007877116975187,
3
+ "eval_loss": 0.04447107017040253,
4
+ "eval_runtime": 154.0801,
5
+ "eval_samples_per_second": 13.876,
6
+ "eval_steps_per_second": 3.472
7
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.0,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "5.8.0"
14
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bd0fa8c59964c92095d355ea6e05d4b34b279e2bd7e051c539da4a64dca1e5d
3
+ size 17534340584
grail_v2op_fcur_frombase_b64_sft_e0_ck640/processor_config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Qwen2VLImageProcessor",
13
+ "image_std": [
14
+ 0.5,
15
+ 0.5,
16
+ 0.5
17
+ ],
18
+ "merge_size": 2,
19
+ "patch_size": 16,
20
+ "resample": 3,
21
+ "rescale_factor": 0.00392156862745098,
22
+ "size": {
23
+ "longest_edge": 16777216,
24
+ "shortest_edge": 65536
25
+ },
26
+ "temporal_patch_size": 2
27
+ },
28
+ "processor_class": "Qwen3VLProcessor",
29
+ "video_processor": {
30
+ "do_convert_rgb": true,
31
+ "do_normalize": true,
32
+ "do_rescale": true,
33
+ "do_resize": true,
34
+ "do_sample_frames": true,
35
+ "fps": 2,
36
+ "image_mean": [
37
+ 0.5,
38
+ 0.5,
39
+ 0.5
40
+ ],
41
+ "image_std": [
42
+ 0.5,
43
+ 0.5,
44
+ 0.5
45
+ ],
46
+ "max_frames": 768,
47
+ "merge_size": 2,
48
+ "min_frames": 4,
49
+ "patch_size": 16,
50
+ "resample": 3,
51
+ "rescale_factor": 0.00392156862745098,
52
+ "return_metadata": false,
53
+ "size": {
54
+ "longest_edge": 25165824,
55
+ "shortest_edge": 4096
56
+ },
57
+ "temporal_patch_size": 2,
58
+ "video_processor_type": "Qwen3VLVideoProcessor"
59
+ }
60
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/sft.yaml ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # LLaMA-Factory full-parameter SFT of Qwen3-VL-8B-Instruct on one HER-final dataset.
2
+ # Four L40S (46 GB), one run at a time: DeepSpeed ZeRO-3 with optimizer offload to host
3
+ # RAM (~130 GB per run, so two runs do not fit in 372 GB); the vision tower and projector
4
+ # stay frozen (LLaMA-Factory's default for VLMs), the LLM trains.
5
+ # Filled per run by launch_sft.sh: grail_v2op_fcur_sft_e3, exps/sft/grail_v2op_fcur_sft_e3_k25515, /scratch/kwc5354/viewagent/viewagent_private/exps/sft_runs/grail_v2op_fcur_frombase_b64_sft_e0.
6
+ ### model
7
+ model_name_or_path: /scratch/kwc5354/viewagent/viewagent_private/models/Qwen3-VL-8B-Instruct
8
+ template: qwen3_vl_nothink
9
+ trust_remote_code: true
10
+ image_max_pixels: 65536 # frames are 256x256 already; never upscale
11
+ image_min_pixels: 4096
12
+
13
+ ### method
14
+ stage: sft
15
+ do_train: true
16
+ finetuning_type: full
17
+ freeze_vision_tower: true
18
+ freeze_multi_modal_projector: true
19
+ freeze_language_model: false
20
+ deepspeed: /scratch/kwc5354/viewagent/viewagent_private/LLaMA-Factory/examples/deepspeed/ds_z3_config.json
21
+
22
+ ### dataset
23
+ dataset: grail_v2op_fcur_sft_e3
24
+ dataset_dir: exps/sft/grail_v2op_fcur_sft_e3_k25515
25
+ media_dir: /scratch/kwc5354/viewagent/viewagent_private
26
+ cutoff_len: 8192
27
+ overwrite_cache: true
28
+ preprocessing_num_workers: 16
29
+ dataloader_num_workers: 4
30
+
31
+ ### output
32
+ output_dir: /scratch/kwc5354/viewagent/viewagent_private/exps/sft_runs/grail_v2op_fcur_frombase_b64_sft_e0
33
+ logging_steps: 10
34
+ save_strategy: steps
35
+ save_steps: 640
36
+ save_total_limit: 100
37
+ save_only_model: true
38
+ plot_loss: true
39
+ overwrite_output_dir: true
40
+ report_to: wandb
41
+
42
+ ### train
43
+ per_device_train_batch_size: 1
44
+ gradient_accumulation_steps: 16
45
+ max_steps: 640
46
+ learning_rate: 1.0e-5
47
+ lr_scheduler_type: cosine
48
+ warmup_ratio: 0.03
49
+ weight_decay: 0.01
50
+ bf16: true
51
+ gradient_checkpointing: true
52
+ ddp_timeout: 180000000
53
+ seed: 0
54
+ val_size: 0.05
55
+ eval_strategy: steps
56
+ eval_steps: 640
57
+ per_device_eval_batch_size: 1
58
+ run_name: grail_v2op_fcur_frombase_b64_sft_e0
grail_v2op_fcur_frombase_b64_sft_e0_ck640/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
3
+ size 11422650
grail_v2op_fcur_frombase_b64_sft_e0_ck640/tokenizer_config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": true,
24
+ "local_files_only": false,
25
+ "model_max_length": 262144,
26
+ "pad_token": "<|endoftext|>",
27
+ "padding_side": "right",
28
+ "processor_class": "Qwen3VLProcessor",
29
+ "split_special_tokens": false,
30
+ "tokenizer_class": "Qwen2Tokenizer",
31
+ "unk_token": null
32
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.007877116975187,
3
+ "total_flos": 279869738713088.0,
4
+ "train_loss": 0.10761516890488565,
5
+ "train_runtime": 7452.7238,
6
+ "train_samples_per_second": 5.496,
7
+ "train_steps_per_second": 0.086
8
+ }
grail_v2op_fcur_frombase_b64_sft_e0_ck640/trainer_state.json ADDED
@@ -0,0 +1,499 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.007877116975187,
6
+ "eval_steps": 640,
7
+ "global_step": 640,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.015754233950374164,
14
+ "grad_norm": 8.156418714840825,
15
+ "learning_rate": 4.5e-06,
16
+ "loss": 1.1438678741455077,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.03150846790074833,
21
+ "grad_norm": 1.4883800145897648,
22
+ "learning_rate": 9.5e-06,
23
+ "loss": 0.3183743000030518,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.04726270185112249,
28
+ "grad_norm": 0.6510483783294614,
29
+ "learning_rate": 9.994801642711945e-06,
30
+ "loss": 0.21429378986358644,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.06301693580149666,
35
+ "grad_norm": 1.1537532757560485,
36
+ "learning_rate": 9.976845889479778e-06,
37
+ "loss": 0.1948988676071167,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.07877116975187082,
42
+ "grad_norm": 0.8356376761625088,
43
+ "learning_rate": 9.946114644227765e-06,
44
+ "loss": 0.17652475833892822,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.09452540370224498,
49
+ "grad_norm": 0.6308886773585931,
50
+ "learning_rate": 9.90268679362172e-06,
51
+ "loss": 0.17027043104171752,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.11027963765261914,
56
+ "grad_norm": 0.7242202873006504,
57
+ "learning_rate": 9.846673816331297e-06,
58
+ "loss": 0.16150453090667724,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.1260338716029933,
63
+ "grad_norm": 0.6682379818014444,
64
+ "learning_rate": 9.778219496865864e-06,
65
+ "loss": 0.1530665397644043,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.14178810555336746,
70
+ "grad_norm": 0.7092994580109784,
71
+ "learning_rate": 9.6974995564817e-06,
72
+ "loss": 0.1523735523223877,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.15754233950374164,
77
+ "grad_norm": 0.5572459668679396,
78
+ "learning_rate": 9.604721202108084e-06,
79
+ "loss": 0.14312981367111205,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.17329657345411578,
84
+ "grad_norm": 0.4844242026435691,
85
+ "learning_rate": 9.500122594450056e-06,
86
+ "loss": 0.1419189453125,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.18905080740448996,
91
+ "grad_norm": 0.7503931361595733,
92
+ "learning_rate": 9.383972236633308e-06,
93
+ "loss": 0.13150370121002197,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.20480504135486413,
98
+ "grad_norm": 0.6726524644215296,
99
+ "learning_rate": 9.256568284960513e-06,
100
+ "loss": 0.12960975170135497,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.22055927530523828,
105
+ "grad_norm": 0.6186023904092071,
106
+ "learning_rate": 9.118237783548376e-06,
107
+ "loss": 0.125805127620697,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.23631350925561245,
112
+ "grad_norm": 0.6263223478665572,
113
+ "learning_rate": 8.969335824810075e-06,
114
+ "loss": 0.11843910217285156,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.2520677432059866,
119
+ "grad_norm": 0.6907406366849765,
120
+ "learning_rate": 8.810244637938117e-06,
121
+ "loss": 0.11948258876800537,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.26782197715636075,
126
+ "grad_norm": 0.6372192603415691,
127
+ "learning_rate": 8.641372607727503e-06,
128
+ "loss": 0.11816874742507935,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.2835762111067349,
133
+ "grad_norm": 0.5902879573942098,
134
+ "learning_rate": 8.463153226257805e-06,
135
+ "loss": 0.11109216213226318,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.2993304450571091,
140
+ "grad_norm": 0.6387805040364644,
141
+ "learning_rate": 8.276043980125213e-06,
142
+ "loss": 0.10653928518295289,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.31508467900748327,
147
+ "grad_norm": 0.6009310109611027,
148
+ "learning_rate": 8.080525176081006e-06,
149
+ "loss": 0.10383745431900024,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.33083891295785745,
154
+ "grad_norm": 0.6229873747054233,
155
+ "learning_rate": 7.877098708091017e-06,
156
+ "loss": 0.1012420892715454,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.34659314690823156,
161
+ "grad_norm": 0.5925478818107863,
162
+ "learning_rate": 7.666286768981012e-06,
163
+ "loss": 0.09266413450241089,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.36234738085860574,
168
+ "grad_norm": 0.6737449443883872,
169
+ "learning_rate": 7.448630509975199e-06,
170
+ "loss": 0.09604828357696533,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.3781016148089799,
175
+ "grad_norm": 0.6094780735482241,
176
+ "learning_rate": 7.224688651568797e-06,
177
+ "loss": 0.09804449081420899,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.3938558487593541,
182
+ "grad_norm": 0.7633849862762201,
183
+ "learning_rate": 6.995036049300539e-06,
184
+ "loss": 0.09812162518501281,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.40961008270972826,
189
+ "grad_norm": 0.6961020224881399,
190
+ "learning_rate": 6.7602622181067425e-06,
191
+ "loss": 0.08808958530426025,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.4253643166601024,
196
+ "grad_norm": 0.5544799177674665,
197
+ "learning_rate": 6.520969819044933e-06,
198
+ "loss": 0.08887865543365478,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.44111855061047656,
203
+ "grad_norm": 0.45974638039789173,
204
+ "learning_rate": 6.277773112271545e-06,
205
+ "loss": 0.08061434626579285,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.45687278456085073,
210
+ "grad_norm": 0.6299432788529932,
211
+ "learning_rate": 6.031296380244915e-06,
212
+ "loss": 0.08566396832466125,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.4726270185112249,
217
+ "grad_norm": 0.7034387646346978,
218
+ "learning_rate": 5.782172325201155e-06,
219
+ "loss": 0.08006531000137329,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.4883812524615991,
224
+ "grad_norm": 0.5256829554852511,
225
+ "learning_rate": 5.531040445016604e-06,
226
+ "loss": 0.08020959496498108,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.5041354864119733,
231
+ "grad_norm": 0.6364707644510884,
232
+ "learning_rate": 5.278545391625931e-06,
233
+ "loss": 0.08039969205856323,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.5198897203623474,
238
+ "grad_norm": 0.6313641358903704,
239
+ "learning_rate": 5.025335316209883e-06,
240
+ "loss": 0.07396299839019775,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.5356439543127215,
245
+ "grad_norm": 0.6667563159326659,
246
+ "learning_rate": 4.7720602054004865e-06,
247
+ "loss": 0.07660128474235535,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.5513981882630957,
252
+ "grad_norm": 0.5281607924110752,
253
+ "learning_rate": 4.51937021277469e-06,
254
+ "loss": 0.07061385512351989,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.5671524222134698,
259
+ "grad_norm": 0.6852985252319812,
260
+ "learning_rate": 4.267913989919457e-06,
261
+ "loss": 0.06850517988204956,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.5829066561638441,
266
+ "grad_norm": 0.5865406066456832,
267
+ "learning_rate": 4.018337021352447e-06,
268
+ "loss": 0.06569771766662598,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.5986608901142182,
273
+ "grad_norm": 0.5567520570799167,
274
+ "learning_rate": 3.7712799675725105e-06,
275
+ "loss": 0.07000141739845275,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.6144151240645923,
280
+ "grad_norm": 0.5621079504056784,
281
+ "learning_rate": 3.5273770204933814e-06,
282
+ "loss": 0.06801856756210327,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.6301693580149665,
287
+ "grad_norm": 0.6835352744884264,
288
+ "learning_rate": 3.2872542754821334e-06,
289
+ "loss": 0.0637732744216919,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.6459235919653407,
294
+ "grad_norm": 0.6548953206201368,
295
+ "learning_rate": 3.051528124181342e-06,
296
+ "loss": 0.06148017048835754,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.6616778259157149,
301
+ "grad_norm": 0.5305027429826881,
302
+ "learning_rate": 2.8208036722405895e-06,
303
+ "loss": 0.05930085182189941,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.677432059866089,
308
+ "grad_norm": 0.6385411236808015,
309
+ "learning_rate": 2.595673186018929e-06,
310
+ "loss": 0.059533333778381346,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.6931862938164631,
315
+ "grad_norm": 0.680375803645669,
316
+ "learning_rate": 2.376714572245651e-06,
317
+ "loss": 0.056560075283050536,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.7089405277668374,
322
+ "grad_norm": 0.6246774909515327,
323
+ "learning_rate": 2.1644898945419814e-06,
324
+ "loss": 0.05835422277450562,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.7246947617172115,
329
+ "grad_norm": 0.640681699181404,
330
+ "learning_rate": 1.959543930611835e-06,
331
+ "loss": 0.057062184810638426,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.7404489956675857,
336
+ "grad_norm": 0.6882365312265537,
337
+ "learning_rate": 1.7624027738052669e-06,
338
+ "loss": 0.05848686099052429,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.7562032296179598,
343
+ "grad_norm": 0.6371729865862438,
344
+ "learning_rate": 1.5735724826443788e-06,
345
+ "loss": 0.052405011653900144,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.771957463568334,
350
+ "grad_norm": 0.8295158197391014,
351
+ "learning_rate": 1.3935377817783585e-06,
352
+ "loss": 0.05299849510192871,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.7877116975187082,
357
+ "grad_norm": 0.7095413365505934,
358
+ "learning_rate": 1.2227608177022522e-06,
359
+ "loss": 0.054823458194732666,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.8034659314690823,
364
+ "grad_norm": 0.6325223354822481,
365
+ "learning_rate": 1.0616799724335387e-06,
366
+ "loss": 0.05424237251281738,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.8192201654194565,
371
+ "grad_norm": 0.6161827536228163,
372
+ "learning_rate": 9.107087381917878e-07,
373
+ "loss": 0.04625344276428223,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.8349743993698306,
378
+ "grad_norm": 0.9065677062952852,
379
+ "learning_rate": 7.702346559700502e-07,
380
+ "loss": 0.04716808199882507,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.8507286333202048,
385
+ "grad_norm": 0.709267296018985,
386
+ "learning_rate": 6.406183207226973e-07,
387
+ "loss": 0.04759809076786041,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.866482867270579,
392
+ "grad_norm": 0.8633561813492827,
393
+ "learning_rate": 5.22192455723336e-07,
394
+ "loss": 0.04727648496627808,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.8822371012209531,
399
+ "grad_norm": 0.6216014815478527,
400
+ "learning_rate": 4.152610584689443e-07,
401
+ "loss": 0.048144012689590454,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.8979913351713273,
406
+ "grad_norm": 0.7001833996680297,
407
+ "learning_rate": 3.200986203226536e-07,
408
+ "loss": 0.043982139229774474,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.9137455691217015,
413
+ "grad_norm": 0.6550697130508242,
414
+ "learning_rate": 2.3694942189835423e-07,
415
+ "loss": 0.050806629657745364,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.9294998030720756,
420
+ "grad_norm": 0.7014326267626828,
421
+ "learning_rate": 1.6602690599586747e-07,
422
+ "loss": 0.04786196947097778,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.9452540370224498,
427
+ "grad_norm": 0.6273897124587577,
428
+ "learning_rate": 1.0751312969633665e-07,
429
+ "loss": 0.04210348129272461,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.9610082709728239,
434
+ "grad_norm": 0.7407993313914509,
435
+ "learning_rate": 6.15582970243117e-08,
436
+ "loss": 0.04657931923866272,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.9767625049231982,
441
+ "grad_norm": 0.7414317736217144,
442
+ "learning_rate": 2.828037337617273e-08,
443
+ "loss": 0.04937379062175751,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.9925167388735723,
448
+ "grad_norm": 0.6843295379160457,
449
+ "learning_rate": 7.764782704649532e-09,
450
+ "loss": 0.04136794209480286,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 1.007877116975187,
455
+ "grad_norm": 0.5389898749860669,
456
+ "learning_rate": 6.418823675857777e-11,
457
+ "loss": 0.041690993309021,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 1.007877116975187,
462
+ "eval_loss": 0.04447107017040253,
463
+ "eval_runtime": 154.0973,
464
+ "eval_samples_per_second": 13.874,
465
+ "eval_steps_per_second": 3.472,
466
+ "step": 640
467
+ },
468
+ {
469
+ "epoch": 1.007877116975187,
470
+ "step": 640,
471
+ "total_flos": 279869738713088.0,
472
+ "train_loss": 0.10761516890488565,
473
+ "train_runtime": 7452.7238,
474
+ "train_samples_per_second": 5.496,
475
+ "train_steps_per_second": 0.086
476
+ }
477
+ ],
478
+ "logging_steps": 10,
479
+ "max_steps": 640,
480
+ "num_input_tokens_seen": 0,
481
+ "num_train_epochs": 2,
482
+ "save_steps": 640,
483
+ "stateful_callbacks": {
484
+ "TrainerControl": {
485
+ "args": {
486
+ "should_epoch_stop": false,
487
+ "should_evaluate": false,
488
+ "should_log": false,
489
+ "should_save": true,
490
+ "should_training_stop": true
491
+ },
492
+ "attributes": {}
493
+ }
494
+ },
495
+ "total_flos": 279869738713088.0,
496
+ "train_batch_size": 1,
497
+ "trial_name": null,
498
+ "trial_params": null
499
+ }