Hugging Face's logo

VisionXLab
/
FIRM-Video-8B-InternVL3

Image-Text-to-Text
Transformers
TensorBoard
Safetensors
internvl
llama-factory
full
Generated from Trainer
conversational
Model card Files
xet
 Metrics Community
zpy777 commited on
Commit
e9c2648
·
verified ·
1 Parent(s): a79047d

Upload folder using huggingface_hub

Browse files
Files changed (16) hide show
  1. .gitattributes +1 -0
  2. README.md +60 -0
  3. all_results.json +8 -0
  4. chat_template.jinja +6 -0
  5. config.json +117 -0
  6. generation_config.json +9 -0
  7. model.safetensors +3 -0
  8. processor_config.json +63 -0
  9. runs/Jul07_18-17-35_train-1622597376801900288-cbzs6f2rii68-master-0/events.out.tfevents.1783419455.train-1622597376801900288-cbzs6f2rii68-master-0.8883.0 +3 -0
  10. tokenizer.json +3 -0
  11. tokenizer_config.json +34 -0
  12. train_results.json +8 -0
  13. trainer_log.jsonl +0 -0
  14. trainer_state.json +0 -0
  15. training_args.bin +3 -0
  16. training_loss.png +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: other
4
+ base_model: OpenGVLab/InternVL3-8B
5
+ tags:
6
+ - llama-factory
7
+ - full
8
+ - generated_from_trainer
9
+ model-index:
10
+ - name: sft
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # sft
18
+
19
+ This model is a fine-tuned version of [OpenGVLab/InternVL3-8B](https://huggingface.co/OpenGVLab/InternVL3-8B) on the firm_video_if, the firm_video_vq and the firm_video_wc datasets.
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 5e-06
39
+ - train_batch_size: 2
40
+ - eval_batch_size: 8
41
+ - seed: 42
42
+ - distributed_type: multi-GPU
43
+ - num_devices: 16
44
+ - total_train_batch_size: 32
45
+ - total_eval_batch_size: 128
46
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
47
+ - lr_scheduler_type: cosine
48
+ - lr_scheduler_warmup_steps: 0.05
49
+ - num_epochs: 2.0
50
+
51
+ ### Training results
52
+
53
+
54
+
55
+ ### Framework versions
56
+
57
+ - Transformers 5.6.0
58
+ - Pytorch 2.7.1+cu126
59
+ - Datasets 3.6.0
60
+ - Tokenizers 0.22.2
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.0,
3
+ "total_flos": 2.0144853494769123e+19,
4
+ "train_loss": 0.736562816166254,
5
+ "train_runtime": 18605.1295,
6
+ "train_samples_per_second": 9.464,
7
+ "train_steps_per_second": 0.296
8
+ }
chat_template.jinja ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {% for message in messages %}{{'<|im_start|>' + message['role'] + '
2
+ '}}{% if message['content'] is string %}{{ message['content'] }}{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' %}{{ '<IMG_CONTEXT>
3
+ ' }}{% elif content['type'] == 'video' %}{{ '<video>
4
+ ' }}{% elif content['type'] == 'text' %}{{ content['text'] }}{% endif %}{% endfor %}{% endif %}{{'<|im_end|>
5
+ '}}{% endfor %}{% if add_generation_prompt %}{{'<|im_start|>assistant
6
+ ' }}{% endif %}
config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "InternVLForConditionalGeneration"
4
+ ],
5
+ "bos_token_id": null,
6
+ "downsample_ratio": 0.5,
7
+ "dtype": "bfloat16",
8
+ "eos_token_id": 151645,
9
+ "hidden_size": 3584,
10
+ "image_seq_length": 256,
11
+ "image_token_id": 151667,
12
+ "model_type": "internvl",
13
+ "pad_token_id": 151643,
14
+ "projector_hidden_act": "gelu",
15
+ "text_config": {
16
+ "architectures": [
17
+ "Qwen2ForCausalLM"
18
+ ],
19
+ "attention_dropout": 0.0,
20
+ "bos_token_id": 151643,
21
+ "dtype": "bfloat16",
22
+ "eos_token_id": 151645,
23
+ "hidden_act": "silu",
24
+ "hidden_size": 3584,
25
+ "initializer_range": 0.02,
26
+ "intermediate_size": 18944,
27
+ "layer_types": [
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention",
52
+ "full_attention",
53
+ "full_attention",
54
+ "full_attention",
55
+ "full_attention"
56
+ ],
57
+ "max_position_embeddings": 32768,
58
+ "max_window_layers": 70,
59
+ "model_type": "qwen2",
60
+ "num_attention_heads": 28,
61
+ "num_hidden_layers": 28,
62
+ "num_key_value_heads": 4,
63
+ "pad_token_id": null,
64
+ "rms_norm_eps": 1e-06,
65
+ "rope_parameters": {
66
+ "factor": 2.0,
67
+ "rope_theta": 1000000.0,
68
+ "rope_type": "dynamic",
69
+ "type": "dynamic"
70
+ },
71
+ "sliding_window": null,
72
+ "tie_word_embeddings": false,
73
+ "use_cache": false,
74
+ "use_sliding_window": false,
75
+ "vocab_size": 151674
76
+ },
77
+ "tie_word_embeddings": true,
78
+ "transformers_version": "5.6.0",
79
+ "use_cache": false,
80
+ "vision_config": {
81
+ "architectures": [
82
+ "InternVisionModel"
83
+ ],
84
+ "attention_bias": true,
85
+ "attention_dropout": 0.0,
86
+ "dropout": 0.0,
87
+ "dtype": "bfloat16",
88
+ "hidden_act": "gelu",
89
+ "hidden_dropout_prob": 0.0,
90
+ "hidden_size": 1024,
91
+ "image_size": [
92
+ 448,
93
+ 448
94
+ ],
95
+ "initializer_factor": 0.1,
96
+ "initializer_range": 1e-10,
97
+ "intermediate_size": 4096,
98
+ "layer_norm_eps": 1e-06,
99
+ "layer_scale_init_value": 0.1,
100
+ "model_type": "internvl_vision",
101
+ "norm_type": "layer_norm",
102
+ "num_attention_heads": 16,
103
+ "num_channels": 3,
104
+ "num_hidden_layers": 24,
105
+ "patch_size": [
106
+ 14,
107
+ 14
108
+ ],
109
+ "projection_dropout": 0.0,
110
+ "use_absolute_position_embeddings": true,
111
+ "use_mask_token": false,
112
+ "use_mean_pooling": true,
113
+ "use_qk_norm": false
114
+ },
115
+ "vision_feature_layer": -1,
116
+ "vision_feature_select_strategy": "default"
117
+ }
generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151645
6
+ ],
7
+ "pad_token_id": 151643,
8
+ "transformers_version": "5.6.0"
9
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f725918643cbceb506048ab69efcda771f6a8b6cd688e67fe8792fd11c1ca6b
3
+ size 15888847664
processor_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "crop_to_patches": false,
4
+ "data_format": "channels_first",
5
+ "default_to_square": true,
6
+ "do_convert_rgb": true,
7
+ "do_normalize": true,
8
+ "do_rescale": true,
9
+ "do_resize": true,
10
+ "image_mean": [
11
+ 0.485,
12
+ 0.456,
13
+ 0.406
14
+ ],
15
+ "image_processor_type": "GotOcr2ImageProcessorFast",
16
+ "image_std": [
17
+ 0.229,
18
+ 0.224,
19
+ 0.225
20
+ ],
21
+ "max_patches": 12,
22
+ "min_patches": 1,
23
+ "resample": 3,
24
+ "rescale_factor": 0.00392156862745098,
25
+ "size": {
26
+ "height": 448,
27
+ "width": 448
28
+ }
29
+ },
30
+ "image_seq_length": 256,
31
+ "processor_class": "InternVLProcessor",
32
+ "video_processor": {
33
+ "crop_to_patches": false,
34
+ "data_format": "channels_first",
35
+ "default_to_square": true,
36
+ "do_convert_rgb": true,
37
+ "do_normalize": true,
38
+ "do_rescale": true,
39
+ "do_resize": true,
40
+ "do_sample_frames": false,
41
+ "image_mean": [
42
+ 0.485,
43
+ 0.456,
44
+ 0.406
45
+ ],
46
+ "image_std": [
47
+ 0.229,
48
+ 0.224,
49
+ 0.225
50
+ ],
51
+ "initial_shift": true,
52
+ "max_patches": 12,
53
+ "min_patches": 1,
54
+ "resample": 3,
55
+ "rescale_factor": 0.00392156862745098,
56
+ "return_metadata": false,
57
+ "size": {
58
+ "height": 448,
59
+ "width": 448
60
+ },
61
+ "video_processor_type": "InternVLVideoProcessor"
62
+ }
63
+ }
runs/Jul07_18-17-35_train-1622597376801900288-cbzs6f2rii68-master-0/events.out.tfevents.1783419455.train-1622597376801900288-cbzs6f2rii68-master-0.8883.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9506801b4f3e6e25c7d0072d873258aadf476a0124623d6f8bd57571109b5b0
3
+ size 123675
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd0c8be238c81a5f0450067d9565e16794aef2699df631fcf3dd0c2ac4842b06
3
+ size 11423728
tokenizer_config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "context_image_token": "<IMG_CONTEXT>",
7
+ "end_image_token": "</img>",
8
+ "eos_token": "<|im_end|>",
9
+ "errors": "replace",
10
+ "extra_special_tokens": {
11
+ "context_image_token": "<IMG_CONTEXT>",
12
+ "end_image_token": "</img>",
13
+ "start_image_token": "<img>",
14
+ "video_token": "<video>"
15
+ },
16
+ "is_local": true,
17
+ "local_files_only": false,
18
+ "model_max_length": 8192,
19
+ "model_specific_special_tokens": {
20
+ "context_image_token": "<IMG_CONTEXT>",
21
+ "end_image_token": "</img>",
22
+ "start_image_token": "<img>",
23
+ "video_token": "<video>"
24
+ },
25
+ "pad_token": "<|endoftext|>",
26
+ "padding_side": "right",
27
+ "processor_class": "InternVLProcessor",
28
+ "return_token_type_ids": false,
29
+ "split_special_tokens": false,
30
+ "start_image_token": "<img>",
31
+ "tokenizer_class": "Qwen2Tokenizer",
32
+ "unk_token": null,
33
+ "video_token": "<video>"
34
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.0,
3
+ "total_flos": 2.0144853494769123e+19,
4
+ "train_loss": 0.736562816166254,
5
+ "train_runtime": 18605.1295,
6
+ "train_samples_per_second": 9.464,
7
+ "train_steps_per_second": 0.296
8
+ }
trainer_log.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebb030360fa4a0a686022ab764bff96705451b77501130d48d1602cbd0d1a910
3
+ size 7569
training_loss.png ADDED