Hugging Face's logo

Seed42Lab
/
trabank-vl

Safetensors
Model card Files
xet
Community
wonderwind271 commited on
Commit
e6b24e6
·
verified ·
1 Parent(s): 980e824

Upload 458 files

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +25 -0
  2. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/config.json +53 -0
  3. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/generation_config.json +7 -0
  4. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/model.safetensors +3 -0
  5. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/optimizer.pt +3 -0
  6. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/rng_state_0.pth +3 -0
  7. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/rng_state_1.pth +3 -0
  8. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/scheduler.pt +3 -0
  9. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/special_tokens_map.json +24 -0
  10. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/tokenizer_config.json +44 -0
  11. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/trainer_state.json +0 -0
  12. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/training_args.bin +3 -0
  13. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/vocab.json +0 -0
  14. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/config.json +53 -0
  15. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/generation_config.json +7 -0
  16. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/model.safetensors +3 -0
  17. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/optimizer.pt +3 -0
  18. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/rng_state_0.pth +3 -0
  19. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/rng_state_1.pth +3 -0
  20. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/scheduler.pt +3 -0
  21. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/special_tokens_map.json +24 -0
  22. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/tokenizer_config.json +44 -0
  23. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/trainer_state.json +3 -0
  24. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/training_args.bin +3 -0
  25. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/vocab.json +0 -0
  26. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/config.json +53 -0
  27. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/generation_config.json +7 -0
  28. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/model.safetensors +3 -0
  29. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/optimizer.pt +3 -0
  30. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/rng_state_0.pth +3 -0
  31. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/rng_state_1.pth +3 -0
  32. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/scheduler.pt +3 -0
  33. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/special_tokens_map.json +24 -0
  34. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/tokenizer_config.json +44 -0
  35. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/trainer_state.json +3 -0
  36. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/training_args.bin +3 -0
  37. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/vocab.json +0 -0
  38. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/config.json +53 -0
  39. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/generation_config.json +7 -0
  40. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/model.safetensors +3 -0
  41. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/optimizer.pt +3 -0
  42. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/rng_state_0.pth +3 -0
  43. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/rng_state_1.pth +3 -0
  44. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/scheduler.pt +3 -0
  45. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/special_tokens_map.json +24 -0
  46. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/tokenizer_config.json +44 -0
  47. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/trainer_state.json +3 -0
  48. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/training_args.bin +3 -0
  49. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/vocab.json +0 -0
  50. trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-130000/config.json +53 -0
.gitattributes CHANGED
@@ -205,3 +205,28 @@ trabank_vl_dino_pretrain_resize_pure_s142/checkpoint-300000/trainer_state.json f
205
  trabank_vl_dino_pretrain_resize_pure_s242/checkpoint-290000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
206
  trabank_vl_dino_pretrain_resize_pure_s242/checkpoint-300000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
207
  trabank_vl_dino_pretrain_resize_pure_s242/checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
205
  trabank_vl_dino_pretrain_resize_pure_s242/checkpoint-290000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
206
  trabank_vl_dino_pretrain_resize_pure_s242/checkpoint-300000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
207
  trabank_vl_dino_pretrain_resize_pure_s242/checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
208
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
209
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
210
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
211
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-130000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
212
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-140000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
213
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-150000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
214
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-160000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
215
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-170000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
216
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-180000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
217
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-190000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
218
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-200000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
219
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-210000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
220
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-220000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
221
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-230000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
222
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-240000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
223
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
224
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
225
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
226
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-280000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
227
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-290000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
228
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-300000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
229
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-70000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
230
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
231
+ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
232
+ trabank_vl_dino_pretrain_resize_pure_s342/trainer_state.json filter=lfs diff=lfs merge=lfs -text
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "LlavaGPTForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 1,
10
+ "detect_loss": false,
11
+ "embd_pdrop": 0.1,
12
+ "eos_token_id": 2,
13
+ "freeze_mm_mlp_adapter": false,
14
+ "hidden_act": "silu",
15
+ "image_aspect_ratio": "pad",
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 3072,
18
+ "is_decoder": true,
19
+ "layer_norm_epsilon": 1e-05,
20
+ "mm_projector_lr": null,
21
+ "mm_use_im_patch_token": false,
22
+ "mm_use_im_start_end": false,
23
+ "mm_vision_tower": "dino",
24
+ "model_type": "gpt2",
25
+ "n_embd": 768,
26
+ "n_head": 12,
27
+ "n_inner": null,
28
+ "n_layer": 12,
29
+ "n_positions": 1024,
30
+ "num_key_value_heads": 12,
31
+ "reorder_and_upcast_attn": false,
32
+ "resid_pdrop": 0.1,
33
+ "rms_norm_eps": 1e-05,
34
+ "rope_scaling": null,
35
+ "rope_theta": 10000.0,
36
+ "scale_attn_by_inverse_layer_idx": false,
37
+ "scale_attn_weights": true,
38
+ "summary_activation": null,
39
+ "summary_first_dropout": 0.1,
40
+ "summary_proj_to_labels": true,
41
+ "summary_type": "cls_index",
42
+ "summary_use_proj": true,
43
+ "tie_word_embeddings": false,
44
+ "tokenizer_class": "LlamaTokenizer",
45
+ "tokenizer_model_max_length": 2048,
46
+ "tokenizer_padding_side": "right",
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.38.0",
49
+ "tune_mm_mlp_adapter": false,
50
+ "use_cache": false,
51
+ "vision_tower_type": "dino",
52
+ "vocab_size": 25005
53
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.38.0",
6
+ "use_cache": false
7
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24fe646fda422a3e97bd762bf819a488d33eaebecc59c1e29cb938e8bf1a6b0d
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c55eec6f8f4896c68a427d9e86cadd9460814f1a8be306c7499f0c6aaca2300
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12eff270f648bdf381a78bc84a6407ef6eb328f9f543c20ef1934e9e15578aca
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d97abf69f04b6c6bb20005a179953db3d7f148ade94f39e198f40ba1a9d441c2
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f87764ef5b840371a47e1810bb5307032d92a93164912b3c66aa00d07547a08
3
+ size 1064
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<SOS>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<EOS>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<UNK>",
17
+ "unk_token": {
18
+ "content": "<UNK>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/tokenizer_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<PAD>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<UNK>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<SOS>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<EOS>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<SOS>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "eos_token": "<EOS>",
39
+ "model_max_length": 2048,
40
+ "pad_token": "<UNK>",
41
+ "padding_side": "right",
42
+ "tokenizer_class": "CustomWordTokenizer",
43
+ "unk_token": "<UNK>"
44
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5044b3ea2e123a57429a33f55e5dd06c1ec4ed4a5c7012981804777f3108ffb0
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-10000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "LlavaGPTForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 1,
10
+ "detect_loss": false,
11
+ "embd_pdrop": 0.1,
12
+ "eos_token_id": 2,
13
+ "freeze_mm_mlp_adapter": false,
14
+ "hidden_act": "silu",
15
+ "image_aspect_ratio": "pad",
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 3072,
18
+ "is_decoder": true,
19
+ "layer_norm_epsilon": 1e-05,
20
+ "mm_projector_lr": null,
21
+ "mm_use_im_patch_token": false,
22
+ "mm_use_im_start_end": false,
23
+ "mm_vision_tower": "dino",
24
+ "model_type": "gpt2",
25
+ "n_embd": 768,
26
+ "n_head": 12,
27
+ "n_inner": null,
28
+ "n_layer": 12,
29
+ "n_positions": 1024,
30
+ "num_key_value_heads": 12,
31
+ "reorder_and_upcast_attn": false,
32
+ "resid_pdrop": 0.1,
33
+ "rms_norm_eps": 1e-05,
34
+ "rope_scaling": null,
35
+ "rope_theta": 10000.0,
36
+ "scale_attn_by_inverse_layer_idx": false,
37
+ "scale_attn_weights": true,
38
+ "summary_activation": null,
39
+ "summary_first_dropout": 0.1,
40
+ "summary_proj_to_labels": true,
41
+ "summary_type": "cls_index",
42
+ "summary_use_proj": true,
43
+ "tie_word_embeddings": false,
44
+ "tokenizer_class": "LlamaTokenizer",
45
+ "tokenizer_model_max_length": 2048,
46
+ "tokenizer_padding_side": "right",
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.38.0",
49
+ "tune_mm_mlp_adapter": false,
50
+ "use_cache": false,
51
+ "vision_tower_type": "dino",
52
+ "vocab_size": 25005
53
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.38.0",
6
+ "use_cache": false
7
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9e4e7a47d59f09e06e134e5ffe5e4083ec838559fc7bc6b39c7195399b32a23
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27096c7346f91c66874a9f75f607ed7577b788dd5420e5f49a1f29ea7850e855
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b852505d708b79d3c9a7265ed187e9eef8543137cfbeb2662a541c4b120163a0
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e2a9be3ad0c1678bf929b17486c225db24e6d4b39958ec98aa21378e4ac330f
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:500e6eec4249d40d0afce425030cf0b73f377bd28beb47fa674dd78e3a5d7ff2
3
+ size 1064
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<SOS>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<EOS>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<UNK>",
17
+ "unk_token": {
18
+ "content": "<UNK>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/tokenizer_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<PAD>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<UNK>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<SOS>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<EOS>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<SOS>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "eos_token": "<EOS>",
39
+ "model_max_length": 2048,
40
+ "pad_token": "<UNK>",
41
+ "padding_side": "right",
42
+ "tokenizer_class": "CustomWordTokenizer",
43
+ "unk_token": "<UNK>"
44
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d238dfb31b388e125a568f5395415f8cce23ace58f26acd0ad79414b98d1ca6
3
+ size 16065600
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5044b3ea2e123a57429a33f55e5dd06c1ec4ed4a5c7012981804777f3108ffb0
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-100000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "LlavaGPTForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 1,
10
+ "detect_loss": false,
11
+ "embd_pdrop": 0.1,
12
+ "eos_token_id": 2,
13
+ "freeze_mm_mlp_adapter": false,
14
+ "hidden_act": "silu",
15
+ "image_aspect_ratio": "pad",
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 3072,
18
+ "is_decoder": true,
19
+ "layer_norm_epsilon": 1e-05,
20
+ "mm_projector_lr": null,
21
+ "mm_use_im_patch_token": false,
22
+ "mm_use_im_start_end": false,
23
+ "mm_vision_tower": "dino",
24
+ "model_type": "gpt2",
25
+ "n_embd": 768,
26
+ "n_head": 12,
27
+ "n_inner": null,
28
+ "n_layer": 12,
29
+ "n_positions": 1024,
30
+ "num_key_value_heads": 12,
31
+ "reorder_and_upcast_attn": false,
32
+ "resid_pdrop": 0.1,
33
+ "rms_norm_eps": 1e-05,
34
+ "rope_scaling": null,
35
+ "rope_theta": 10000.0,
36
+ "scale_attn_by_inverse_layer_idx": false,
37
+ "scale_attn_weights": true,
38
+ "summary_activation": null,
39
+ "summary_first_dropout": 0.1,
40
+ "summary_proj_to_labels": true,
41
+ "summary_type": "cls_index",
42
+ "summary_use_proj": true,
43
+ "tie_word_embeddings": false,
44
+ "tokenizer_class": "LlamaTokenizer",
45
+ "tokenizer_model_max_length": 2048,
46
+ "tokenizer_padding_side": "right",
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.38.0",
49
+ "tune_mm_mlp_adapter": false,
50
+ "use_cache": false,
51
+ "vision_tower_type": "dino",
52
+ "vocab_size": 25005
53
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.38.0",
6
+ "use_cache": false
7
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5231ef32269966bddc23aa15ea114c153fb8b62243d7b1900c435b6c03cae14
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b4dc605a5e630c2aaec0db651e5f3289306ab0560531b3705d1dda31e65a6e1
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7869340588d384637955e5009cc005b4d9d5429660b7eb7428d13d783b3fa39a
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6112b704b9e2956a7eac92287ac9b91c1f8be5976cd5b7c3af1e7e20c9046125
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d88dec93b828de6efe816f92835c7e1e61469d3d29bd40488cf508b43dcf5b89
3
+ size 1064
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<SOS>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<EOS>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<UNK>",
17
+ "unk_token": {
18
+ "content": "<UNK>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/tokenizer_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<PAD>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<UNK>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<SOS>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<EOS>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<SOS>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "eos_token": "<EOS>",
39
+ "model_max_length": 2048,
40
+ "pad_token": "<UNK>",
41
+ "padding_side": "right",
42
+ "tokenizer_class": "CustomWordTokenizer",
43
+ "unk_token": "<UNK>"
44
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5938e2fec392cc815e6c22e59b768f688f13b897578206223e0c3fa0801ba806
3
+ size 17684775
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5044b3ea2e123a57429a33f55e5dd06c1ec4ed4a5c7012981804777f3108ffb0
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-110000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "LlavaGPTForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 1,
10
+ "detect_loss": false,
11
+ "embd_pdrop": 0.1,
12
+ "eos_token_id": 2,
13
+ "freeze_mm_mlp_adapter": false,
14
+ "hidden_act": "silu",
15
+ "image_aspect_ratio": "pad",
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 3072,
18
+ "is_decoder": true,
19
+ "layer_norm_epsilon": 1e-05,
20
+ "mm_projector_lr": null,
21
+ "mm_use_im_patch_token": false,
22
+ "mm_use_im_start_end": false,
23
+ "mm_vision_tower": "dino",
24
+ "model_type": "gpt2",
25
+ "n_embd": 768,
26
+ "n_head": 12,
27
+ "n_inner": null,
28
+ "n_layer": 12,
29
+ "n_positions": 1024,
30
+ "num_key_value_heads": 12,
31
+ "reorder_and_upcast_attn": false,
32
+ "resid_pdrop": 0.1,
33
+ "rms_norm_eps": 1e-05,
34
+ "rope_scaling": null,
35
+ "rope_theta": 10000.0,
36
+ "scale_attn_by_inverse_layer_idx": false,
37
+ "scale_attn_weights": true,
38
+ "summary_activation": null,
39
+ "summary_first_dropout": 0.1,
40
+ "summary_proj_to_labels": true,
41
+ "summary_type": "cls_index",
42
+ "summary_use_proj": true,
43
+ "tie_word_embeddings": false,
44
+ "tokenizer_class": "LlamaTokenizer",
45
+ "tokenizer_model_max_length": 2048,
46
+ "tokenizer_padding_side": "right",
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.38.0",
49
+ "tune_mm_mlp_adapter": false,
50
+ "use_cache": false,
51
+ "vision_tower_type": "dino",
52
+ "vocab_size": 25005
53
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.38.0",
6
+ "use_cache": false
7
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:045d61329002c74f39e3ff3d6cbc26aac31afdc674be8e1180439ef8b86eef38
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f25c66503147de92c78375028cfd1479b8280c28f1807457368141835816245
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:663b2669c31029c174a0e47e3507030706fdbfb7504d17b9249eb36672172f74
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ccdf31e372ebb50d250a8d88b631d029f9ba37d384f70ef58b4f8f48715efcf9
3
+ size 14576
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e01787c5c52762beae7a8a45a0eef397581357036f15551400946173f757aba6
3
+ size 1064
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<SOS>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<EOS>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<UNK>",
17
+ "unk_token": {
18
+ "content": "<UNK>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/tokenizer_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<PAD>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<UNK>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<SOS>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<EOS>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "bos_token": "<SOS>",
37
+ "clean_up_tokenization_spaces": true,
38
+ "eos_token": "<EOS>",
39
+ "model_max_length": 2048,
40
+ "pad_token": "<UNK>",
41
+ "padding_side": "right",
42
+ "tokenizer_class": "CustomWordTokenizer",
43
+ "unk_token": "<UNK>"
44
+ }
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c8d6f80d0c5a06a1f7b2c8ff3086819c8c536ebc70c8772c03bba486f3813ad
3
+ size 19305597
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5044b3ea2e123a57429a33f55e5dd06c1ec4ed4a5c7012981804777f3108ffb0
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-120000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-130000/config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "LlavaGPTForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 1,
10
+ "detect_loss": false,
11
+ "embd_pdrop": 0.1,
12
+ "eos_token_id": 2,
13
+ "freeze_mm_mlp_adapter": false,
14
+ "hidden_act": "silu",
15
+ "image_aspect_ratio": "pad",
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 3072,
18
+ "is_decoder": true,
19
+ "layer_norm_epsilon": 1e-05,
20
+ "mm_projector_lr": null,
21
+ "mm_use_im_patch_token": false,
22
+ "mm_use_im_start_end": false,
23
+ "mm_vision_tower": "dino",
24
+ "model_type": "gpt2",
25
+ "n_embd": 768,
26
+ "n_head": 12,
27
+ "n_inner": null,
28
+ "n_layer": 12,
29
+ "n_positions": 1024,
30
+ "num_key_value_heads": 12,
31
+ "reorder_and_upcast_attn": false,
32
+ "resid_pdrop": 0.1,
33
+ "rms_norm_eps": 1e-05,
34
+ "rope_scaling": null,
35
+ "rope_theta": 10000.0,
36
+ "scale_attn_by_inverse_layer_idx": false,
37
+ "scale_attn_weights": true,
38
+ "summary_activation": null,
39
+ "summary_first_dropout": 0.1,
40
+ "summary_proj_to_labels": true,
41
+ "summary_type": "cls_index",
42
+ "summary_use_proj": true,
43
+ "tie_word_embeddings": false,
44
+ "tokenizer_class": "LlamaTokenizer",
45
+ "tokenizer_model_max_length": 2048,
46
+ "tokenizer_padding_side": "right",
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.38.0",
49
+ "tune_mm_mlp_adapter": false,
50
+ "use_cache": false,
51
+ "vision_tower_type": "dino",
52
+ "vocab_size": 25005
53
+ }