Hugging Face's logo

Seed42Lab
/
trabank-vl

Safetensors
Model card Files
xet
Community
wonderwind271 commited on
Commit
75ff7f0
·
verified ·
1 Parent(s): 844cbfe

VL resize seed 42

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_s42/checkpoint-10000/config.json +53 -0
  3. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/generation_config.json +7 -0
  4. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/model.safetensors +3 -0
  5. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/optimizer.pt +3 -0
  6. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/rng_state_0.pth +3 -0
  7. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/rng_state_1.pth +3 -0
  8. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/scheduler.pt +3 -0
  9. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/special_tokens_map.json +24 -0
  10. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/tokenizer_config.json +44 -0
  11. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/trainer_state.json +0 -0
  12. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/training_args.bin +3 -0
  13. trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/vocab.json +0 -0
  14. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/config.json +53 -0
  15. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/generation_config.json +7 -0
  16. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/model.safetensors +3 -0
  17. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/optimizer.pt +3 -0
  18. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/rng_state_0.pth +3 -0
  19. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/rng_state_1.pth +3 -0
  20. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/scheduler.pt +3 -0
  21. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/special_tokens_map.json +24 -0
  22. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/tokenizer_config.json +44 -0
  23. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/trainer_state.json +3 -0
  24. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/training_args.bin +3 -0
  25. trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/vocab.json +0 -0
  26. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/config.json +53 -0
  27. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/generation_config.json +7 -0
  28. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/model.safetensors +3 -0
  29. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/optimizer.pt +3 -0
  30. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/rng_state_0.pth +3 -0
  31. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/rng_state_1.pth +3 -0
  32. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/scheduler.pt +3 -0
  33. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/special_tokens_map.json +24 -0
  34. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/tokenizer_config.json +44 -0
  35. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/trainer_state.json +3 -0
  36. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/training_args.bin +3 -0
  37. trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/vocab.json +0 -0
  38. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/config.json +53 -0
  39. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/generation_config.json +7 -0
  40. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/model.safetensors +3 -0
  41. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/optimizer.pt +3 -0
  42. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/rng_state_0.pth +3 -0
  43. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/rng_state_1.pth +3 -0
  44. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/scheduler.pt +3 -0
  45. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/special_tokens_map.json +24 -0
  46. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/tokenizer_config.json +44 -0
  47. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/trainer_state.json +3 -0
  48. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/training_args.bin +3 -0
  49. trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/vocab.json +0 -0
  50. trabank_vl_dino_pretrain_resize_s42/checkpoint-130000/config.json +53 -0
.gitattributes CHANGED
@@ -108,3 +108,28 @@ trabank_vl_dino_pretrain2_s242/checkpoint-70000/trainer_state.json filter=lfs di
108
  trabank_vl_dino_pretrain2_s242/checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
109
  trabank_vl_dino_pretrain2_s242/checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
110
  trabank_vl_dino_pretrain2_s242/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
108
  trabank_vl_dino_pretrain2_s242/checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
109
  trabank_vl_dino_pretrain2_s242/checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
110
  trabank_vl_dino_pretrain2_s242/trainer_state.json filter=lfs diff=lfs merge=lfs -text
111
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
112
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
113
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
114
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-130000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
115
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-140000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
116
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-150000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
117
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-160000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
118
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-170000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
119
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-180000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
120
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-190000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
121
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-200000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
122
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-210000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
123
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-220000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
124
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-230000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
125
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-240000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
126
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
127
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
128
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
129
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-280000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
130
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-290000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
131
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-300000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
132
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-70000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
133
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
134
+ trabank_vl_dino_pretrain_resize_s42/checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
135
+ trabank_vl_dino_pretrain_resize_s42/trainer_state.json filter=lfs diff=lfs merge=lfs -text
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-10000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64389fe96636bd0ea3b988d10ec66ae16ba70fec3b4304a719f5cee3f8bd4ef3
3
+ size 670210056
trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d424e6f0a23ebbc940b1338eee508f94b168a33875838f85002038f3970aabe8
3
+ size 994126027
trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd59fb32c15c5859fe5132fecc2146ed667e65212106d16ab4f0dfcbd720fbde
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38cc46c566ccb17818502ac4fea4d39b3588af334dfeb26105511cffe151fdaa
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4aeaedc1e6e4ac9ac4f28a8691779e4a906da7b051e36d0bbe385033f066178d
3
+ size 1465
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-10000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b7183dea559efbd08c6a3837b9b95d35d9baed595cb10121560b4e29172d85b
3
+ size 5905
trabank_vl_dino_pretrain_resize_s42/checkpoint-10000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-100000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ccd35574273c6c7af2854017d598c29f913a21eff2fedda2c3fbdadea87b264
3
+ size 670210056
trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:890a8319a3aea7c5cf85d5e3b7548e027c95e0422fa32a357735ee38e02e9fec
3
+ size 994126027
trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b51ee1592379bb2b5905c7e26bf1105ef7eddc2a35fc7200d83d1febb11e64d
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0481588ead198fe4ef55ef3e49ddad6bc933d5d83151d5e2db66b1a48197cf1e
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae0a341cc06fce0d21bf7169c4964d449cefbf6ff943711fe1a78667c32ac1df
3
+ size 1465
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-100000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:123ae2b3f4c2a61369115d44cc04f8958201bbab23137748b9acf68eb1b9ff3e
3
+ size 16065838
trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b7183dea559efbd08c6a3837b9b95d35d9baed595cb10121560b4e29172d85b
3
+ size 5905
trabank_vl_dino_pretrain_resize_s42/checkpoint-100000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-110000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b58a44c771a77cbf1b11dea0a3a8fc3ab3e9c8d1e7c8e255611d10524e8fa04d
3
+ size 670210056
trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:694e468316b8e9e50b03aab71fa8f97eae850e43a7fe632904964330e2d9ad9a
3
+ size 994126027
trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf1cee71f1adda15e4dd673e526eeae962848c026a0f40374a30e611008ab0a8
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:358dd1ab2efb123f54c1557adb504b8ca7fb10c02336ffaf514e103cc737d4a4
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db6067b09a272acaf5c85ca7f6a9e4741d56ce0b4071a298c7b399949ec03a20
3
+ size 1465
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-110000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:760711d381d6b097381c5ef5f1f594906117aea714b5e19e9040aa0db322e640
3
+ size 17684785
trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b7183dea559efbd08c6a3837b9b95d35d9baed595cb10121560b4e29172d85b
3
+ size 5905
trabank_vl_dino_pretrain_resize_s42/checkpoint-110000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-120000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec42591f4ad1125e6f47b6d8cc4e9d6eada8df9b8b0ea96612041b7d952205dc
3
+ size 670210056
trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5caabbd00087516bb24798c790cfa63de02ba03c2f5d090fdd60d0776ded226
3
+ size 994126027
trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe14a5cd2bd9387a105d1a8cbd054b61fdb61b46383690a1d3551b6801e5f708
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b11d403fc901ca22f3febea4bec3bf6559c4cb05abb591883e7fbc59f643cc2
3
+ size 14917
trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba0eb79c069ba3b4ea5167de610e72a035c58c9eda7c3683881e185425cc6ffd
3
+ size 1465
trabank_vl_dino_pretrain_resize_s42/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_s42/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_s42/checkpoint-120000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:134295ee1f908a30b9fd648a83ea76478bd88b6c004f7fb62fe8fd4842f086a7
3
+ size 19305666
trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b7183dea559efbd08c6a3837b9b95d35d9baed595cb10121560b4e29172d85b
3
+ size 5905
trabank_vl_dino_pretrain_resize_s42/checkpoint-120000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
trabank_vl_dino_pretrain_resize_s42/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
+ }