Hugging Face's logo

Seed42Lab
/
trabank-vl

Safetensors
Model card Files
xet
Community
wonderwind271 commited on
Commit
ddfac69
·
verified ·
1 Parent(s): e6b24e6

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_s442/checkpoint-10000/config.json +53 -0
  3. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/generation_config.json +7 -0
  4. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/model.safetensors +3 -0
  5. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/optimizer.pt +3 -0
  6. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/rng_state_0.pth +3 -0
  7. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/rng_state_1.pth +3 -0
  8. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/scheduler.pt +3 -0
  9. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/special_tokens_map.json +24 -0
  10. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/tokenizer_config.json +44 -0
  11. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/trainer_state.json +0 -0
  12. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/training_args.bin +3 -0
  13. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/vocab.json +0 -0
  14. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/config.json +53 -0
  15. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/generation_config.json +7 -0
  16. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/model.safetensors +3 -0
  17. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/optimizer.pt +3 -0
  18. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/rng_state_0.pth +3 -0
  19. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/rng_state_1.pth +3 -0
  20. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/scheduler.pt +3 -0
  21. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/special_tokens_map.json +24 -0
  22. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/tokenizer_config.json +44 -0
  23. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/trainer_state.json +3 -0
  24. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/training_args.bin +3 -0
  25. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/vocab.json +0 -0
  26. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/config.json +53 -0
  27. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/generation_config.json +7 -0
  28. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/model.safetensors +3 -0
  29. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/optimizer.pt +3 -0
  30. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/rng_state_0.pth +3 -0
  31. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/rng_state_1.pth +3 -0
  32. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/scheduler.pt +3 -0
  33. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/special_tokens_map.json +24 -0
  34. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/tokenizer_config.json +44 -0
  35. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/trainer_state.json +3 -0
  36. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/training_args.bin +3 -0
  37. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/vocab.json +0 -0
  38. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/config.json +53 -0
  39. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/generation_config.json +7 -0
  40. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/model.safetensors +3 -0
  41. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/optimizer.pt +3 -0
  42. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/rng_state_0.pth +3 -0
  43. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/rng_state_1.pth +3 -0
  44. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/scheduler.pt +3 -0
  45. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/special_tokens_map.json +24 -0
  46. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/tokenizer_config.json +44 -0
  47. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/trainer_state.json +3 -0
  48. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/training_args.bin +3 -0
  49. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/vocab.json +0 -0
  50. trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-130000/config.json +53 -0
.gitattributes CHANGED
@@ -230,3 +230,28 @@ trabank_vl_dino_pretrain_resize_pure_s342/checkpoint-70000/trainer_state.json fi
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
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
233
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
234
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
235
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
236
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-130000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
237
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-140000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
238
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-150000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
239
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-160000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
240
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-170000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
241
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-180000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
242
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-190000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
243
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-200000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
244
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-210000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
245
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-220000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
246
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-230000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
247
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-240000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
248
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
249
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
250
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
251
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-280000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
252
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-290000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
253
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-300000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
254
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-70000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
255
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
256
+ trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
257
+ trabank_vl_dino_pretrain_resize_pure_s442/trainer_state.json filter=lfs diff=lfs merge=lfs -text
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/checkpoint-10000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bdb31f7c8f75023bab37fcb89a97e1e1336d7e27afbeb46907f872d4ffa90f6
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5b3c406dc3383749f8b9cc96b9137ffe85f8cbb3ea515397b83bf3b02c4ae9c
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e34e8630749593c717c9ee60ce9fbce09b1aeab3b3f6b0fbfc90cfd3429bf4d6
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb79f6c8916be0b6eef4aec443c2033b5ec064596fa539fc4f3e01c44b1883e1
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/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_s442/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_s442/checkpoint-10000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45b09ab021003fcfcebb4158705234b351d6f32e8ff6d6ede80e50d64c674087
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-10000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/checkpoint-100000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a2c82f3f76b207fd277fa833a6c54ae56e5157a4d819f4b979a6ac123d1d903
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f462bbc43110d74dabce8bfef5bbf8268b6a86101ed536bfb25e2b4e5aedcf3
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebfee9edb81a2c81704e65998166bbe75e1152487bb0a9ed6943d88ee22ba017
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c05b9d6573d30212b005f6b0ea3695ef717a14e8bb2ba20b969f4ff8f1225917
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/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_s442/checkpoint-100000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:449278d376bbbf175ae3c28113a5ca8ea892cae52d86c967025f44312d129bb2
3
+ size 16063782
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45b09ab021003fcfcebb4158705234b351d6f32e8ff6d6ede80e50d64c674087
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-100000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/checkpoint-110000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf430722ef66c8e19e730f7689d3a5ddb83f6f32ebe95ddf89a3b725dd036c0c
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c59ad49498d9cdf86ca04728e93c103f537ddbf9afe0eb0bcf5acacdb593263
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7ebf146c082740a559f6585fb08174ee60dde2e42b333136765d68ec079e611
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d760e66db487ea676c87c5d4b0fe700a63ebf453ba81443c6b4b73748daa8a40
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/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_s442/checkpoint-110000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb955a76041aec61394a352620cca4549a303376f8feda2e5428c5e946ee4166
3
+ size 17682790
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45b09ab021003fcfcebb4158705234b351d6f32e8ff6d6ede80e50d64c674087
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-110000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/checkpoint-120000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e1fd894913cdd1f375bdad7e4e7e7854faa184191a71e1568230164eeff077b9
3
+ size 670210056
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8128ad81c065075a670b7fad93222686a348159c04c951887e195548e84a9416
3
+ size 994125626
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae5b1941d8695409d858180737203c3d5495cb8d97e061492b798e4bfc522108
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47333972ffa4e97b2ce952d0c37d55133b6dbedf5295244e3c568578bfc31153
3
+ size 14512
trabank_vl_dino_pretrain_resize_pure_s442/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_s442/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_s442/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_s442/checkpoint-120000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3c33e5b94dbe55e064d77c7218dedf24805b7c6e539ea5c99c6777c21933a7d
3
+ size 19303480
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45b09ab021003fcfcebb4158705234b351d6f32e8ff6d6ede80e50d64c674087
3
+ size 5496
trabank_vl_dino_pretrain_resize_pure_s442/checkpoint-120000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_pure_s442/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
+ }