Hugging Face's logo

Nyaaneet
/
donut-base-ru-test

Image-Text-to-Text
Transformers
PyTorch
TensorBoard
vision-encoder-decoder
Model card Files
xet
 Metrics Community
1
Nyaaneet commited on
Commit
00f4b27
·
1 Parent(s): 0fc6878

Training in progress, epoch 1

Browse files
Files changed (24) hide show
  1. .gitignore +1 -0
  2. config.json +195 -0
  3. pytorch_model.bin +3 -0
  4. runs/Mar07_15-30-57_ubuntu-gpu/1678203094.326108/events.out.tfevents.1678203094.ubuntu-gpu.9880.1 +3 -0
  5. runs/Mar07_15-30-57_ubuntu-gpu/events.out.tfevents.1678203094.ubuntu-gpu.9880.0 +3 -0
  6. runs/Mar07_15-33-16_ubuntu-gpu/1678203206.4658086/events.out.tfevents.1678203206.ubuntu-gpu.9880.3 +3 -0
  7. runs/Mar07_15-33-16_ubuntu-gpu/events.out.tfevents.1678203206.ubuntu-gpu.9880.2 +3 -0
  8. runs/Mar07_15-38-25_ubuntu-gpu/1678203516.2486959/events.out.tfevents.1678203516.ubuntu-gpu.11468.1 +3 -0
  9. runs/Mar07_15-38-25_ubuntu-gpu/events.out.tfevents.1678203516.ubuntu-gpu.11468.0 +3 -0
  10. runs/Mar10_15-44-40_ubuntu-gpu/1678463084.350115/events.out.tfevents.1678463084.ubuntu-gpu.9149.1 +3 -0
  11. runs/Mar10_15-44-40_ubuntu-gpu/events.out.tfevents.1678463084.ubuntu-gpu.9149.0 +3 -0
  12. runs/Mar11_18-12-23_ubuntu-gpu/1678558347.0917025/events.out.tfevents.1678558347.ubuntu-gpu.2145.1 +3 -0
  13. runs/Mar11_18-12-23_ubuntu-gpu/events.out.tfevents.1678558347.ubuntu-gpu.2145.0 +3 -0
  14. runs/Mar11_18-27-39_ubuntu-gpu/1678559262.7039926/events.out.tfevents.1678559262.ubuntu-gpu.2384.1 +3 -0
  15. runs/Mar11_18-27-39_ubuntu-gpu/events.out.tfevents.1678559262.ubuntu-gpu.2384.0 +3 -0
  16. runs/Mar11_18-33-27_ubuntu-gpu/1678559611.663038/events.out.tfevents.1678559611.ubuntu-gpu.2505.1 +3 -0
  17. runs/Mar11_18-33-27_ubuntu-gpu/events.out.tfevents.1678559611.ubuntu-gpu.2505.0 +3 -0
  18. runs/Mar11_18-44-42_ubuntu-gpu/1678560286.1684813/events.out.tfevents.1678560286.ubuntu-gpu.2746.1 +3 -0
  19. runs/Mar11_18-44-42_ubuntu-gpu/events.out.tfevents.1678560286.ubuntu-gpu.2746.0 +3 -0
  20. runs/Mar11_18-53-37_ubuntu-gpu/1678560821.0609927/events.out.tfevents.1678560821.ubuntu-gpu.3517.1 +3 -0
  21. runs/Mar11_18-53-37_ubuntu-gpu/events.out.tfevents.1678560821.ubuntu-gpu.3517.0 +3 -0
  22. runs/Mar11_19-13-27_ubuntu-gpu/1678562010.4561105/events.out.tfevents.1678562010.ubuntu-gpu.3652.1 +3 -0
  23. runs/Mar11_19-13-27_ubuntu-gpu/events.out.tfevents.1678562010.ubuntu-gpu.3652.0 +3 -0
  24. training_args.bin +3 -0
.gitignore ADDED
@@ -0,0 +1 @@
 
 
1
+ checkpoint-*/
config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_commit_hash": "a959cf33c20e09215873e338299c900f57047c61",
3
+ "_name_or_path": "naver-clova-ix/donut-base",
4
+ "architectures": [
5
+ "VisionEncoderDecoderModel"
6
+ ],
7
+ "decoder": {
8
+ "_name_or_path": "",
9
+ "activation_dropout": 0.0,
10
+ "activation_function": "gelu",
11
+ "add_cross_attention": true,
12
+ "add_final_layer_norm": true,
13
+ "architectures": null,
14
+ "attention_dropout": 0.0,
15
+ "bad_words_ids": null,
16
+ "begin_suppress_tokens": null,
17
+ "bos_token_id": 0,
18
+ "chunk_size_feed_forward": 0,
19
+ "classifier_dropout": 0.0,
20
+ "cross_attention_hidden_size": null,
21
+ "d_model": 1024,
22
+ "decoder_attention_heads": 16,
23
+ "decoder_ffn_dim": 4096,
24
+ "decoder_layerdrop": 0.0,
25
+ "decoder_layers": 4,
26
+ "decoder_start_token_id": null,
27
+ "diversity_penalty": 0.0,
28
+ "do_sample": false,
29
+ "dropout": 0.1,
30
+ "early_stopping": false,
31
+ "encoder_attention_heads": 16,
32
+ "encoder_ffn_dim": 4096,
33
+ "encoder_layerdrop": 0.0,
34
+ "encoder_layers": 12,
35
+ "encoder_no_repeat_ngram_size": 0,
36
+ "eos_token_id": 2,
37
+ "exponential_decay_length_penalty": null,
38
+ "finetuning_task": null,
39
+ "forced_bos_token_id": null,
40
+ "forced_eos_token_id": 2,
41
+ "id2label": {
42
+ "0": "LABEL_0",
43
+ "1": "LABEL_1"
44
+ },
45
+ "init_std": 0.02,
46
+ "is_decoder": true,
47
+ "is_encoder_decoder": false,
48
+ "label2id": {
49
+ "LABEL_0": 0,
50
+ "LABEL_1": 1
51
+ },
52
+ "length_penalty": 1.0,
53
+ "max_length": 512,
54
+ "max_position_embeddings": 1536,
55
+ "min_length": 0,
56
+ "model_type": "mbart",
57
+ "no_repeat_ngram_size": 0,
58
+ "num_beam_groups": 1,
59
+ "num_beams": 1,
60
+ "num_hidden_layers": 12,
61
+ "num_return_sequences": 1,
62
+ "output_attentions": false,
63
+ "output_hidden_states": false,
64
+ "output_scores": false,
65
+ "pad_token_id": 1,
66
+ "prefix": null,
67
+ "problem_type": null,
68
+ "pruned_heads": {},
69
+ "remove_invalid_values": false,
70
+ "repetition_penalty": 1.0,
71
+ "return_dict": true,
72
+ "return_dict_in_generate": false,
73
+ "scale_embedding": true,
74
+ "sep_token_id": null,
75
+ "suppress_tokens": null,
76
+ "task_specific_params": null,
77
+ "temperature": 1.0,
78
+ "tf_legacy_loss": false,
79
+ "tie_encoder_decoder": false,
80
+ "tie_word_embeddings": true,
81
+ "tokenizer_class": null,
82
+ "top_k": 50,
83
+ "top_p": 1.0,
84
+ "torch_dtype": null,
85
+ "torchscript": false,
86
+ "transformers_version": "4.27.0.dev0",
87
+ "typical_p": 1.0,
88
+ "use_bfloat16": false,
89
+ "use_cache": true,
90
+ "vocab_size": 52008
91
+ },
92
+ "decoder_start_token_id": 0,
93
+ "encoder": {
94
+ "_name_or_path": "",
95
+ "add_cross_attention": false,
96
+ "architectures": null,
97
+ "attention_probs_dropout_prob": 0.0,
98
+ "bad_words_ids": null,
99
+ "begin_suppress_tokens": null,
100
+ "bos_token_id": null,
101
+ "chunk_size_feed_forward": 0,
102
+ "cross_attention_hidden_size": null,
103
+ "decoder_start_token_id": null,
104
+ "depths": [
105
+ 2,
106
+ 2,
107
+ 14,
108
+ 2
109
+ ],
110
+ "diversity_penalty": 0.0,
111
+ "do_sample": false,
112
+ "drop_path_rate": 0.1,
113
+ "early_stopping": false,
114
+ "embed_dim": 128,
115
+ "encoder_no_repeat_ngram_size": 0,
116
+ "eos_token_id": null,
117
+ "exponential_decay_length_penalty": null,
118
+ "finetuning_task": null,
119
+ "forced_bos_token_id": null,
120
+ "forced_eos_token_id": null,
121
+ "hidden_act": "gelu",
122
+ "hidden_dropout_prob": 0.0,
123
+ "hidden_size": 1024,
124
+ "id2label": {
125
+ "0": "LABEL_0",
126
+ "1": "LABEL_1"
127
+ },
128
+ "image_size": [
129
+ 960,
130
+ 720
131
+ ],
132
+ "initializer_range": 0.02,
133
+ "is_decoder": false,
134
+ "is_encoder_decoder": false,
135
+ "label2id": {
136
+ "LABEL_0": 0,
137
+ "LABEL_1": 1
138
+ },
139
+ "layer_norm_eps": 1e-05,
140
+ "length_penalty": 1.0,
141
+ "max_length": 20,
142
+ "min_length": 0,
143
+ "mlp_ratio": 4.0,
144
+ "model_type": "donut-swin",
145
+ "no_repeat_ngram_size": 0,
146
+ "num_beam_groups": 1,
147
+ "num_beams": 1,
148
+ "num_channels": 3,
149
+ "num_heads": [
150
+ 4,
151
+ 8,
152
+ 16,
153
+ 32
154
+ ],
155
+ "num_layers": 4,
156
+ "num_return_sequences": 1,
157
+ "output_attentions": false,
158
+ "output_hidden_states": false,
159
+ "output_scores": false,
160
+ "pad_token_id": null,
161
+ "patch_size": 4,
162
+ "path_norm": true,
163
+ "prefix": null,
164
+ "problem_type": null,
165
+ "pruned_heads": {},
166
+ "qkv_bias": true,
167
+ "remove_invalid_values": false,
168
+ "repetition_penalty": 1.0,
169
+ "return_dict": true,
170
+ "return_dict_in_generate": false,
171
+ "sep_token_id": null,
172
+ "suppress_tokens": null,
173
+ "task_specific_params": null,
174
+ "temperature": 1.0,
175
+ "tf_legacy_loss": false,
176
+ "tie_encoder_decoder": false,
177
+ "tie_word_embeddings": true,
178
+ "tokenizer_class": null,
179
+ "top_k": 50,
180
+ "top_p": 1.0,
181
+ "torch_dtype": null,
182
+ "torchscript": false,
183
+ "transformers_version": "4.27.0.dev0",
184
+ "typical_p": 1.0,
185
+ "use_absolute_embeddings": false,
186
+ "use_bfloat16": false,
187
+ "window_size": 10
188
+ },
189
+ "is_encoder_decoder": true,
190
+ "model_type": "vision-encoder-decoder",
191
+ "pad_token_id": 1,
192
+ "tie_word_embeddings": false,
193
+ "torch_dtype": "float32",
194
+ "transformers_version": null
195
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3230b1d483e80c4a47c32b8dc52891e69a2386f819e9166f51375dfe81203cb8
3
+ size 786581273
runs/Mar07_15-30-57_ubuntu-gpu/1678203094.326108/events.out.tfevents.1678203094.ubuntu-gpu.9880.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f4da9bcf5f1b188c7c52da2dd0deda26221752ad94d9f0af18994370cd9e27a
3
+ size 6020
runs/Mar07_15-30-57_ubuntu-gpu/events.out.tfevents.1678203094.ubuntu-gpu.9880.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7511b98b867eb305e50f89e835850b9e2a6dc96ad46f6643123a14d984af2f9
3
+ size 8450
runs/Mar07_15-33-16_ubuntu-gpu/1678203206.4658086/events.out.tfevents.1678203206.ubuntu-gpu.9880.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33e68d9f15ef78766f95f48bda08d0d0e0d32c5cba905059e6197211d1355fb5
3
+ size 6020
runs/Mar07_15-33-16_ubuntu-gpu/events.out.tfevents.1678203206.ubuntu-gpu.9880.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12f38f84b9872df9a09ea801b733bd80344cf3ca00ad42fc0dc17b734e18ba59
3
+ size 8450
runs/Mar07_15-38-25_ubuntu-gpu/1678203516.2486959/events.out.tfevents.1678203516.ubuntu-gpu.11468.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f37405162f852cb49523c6500de80f497fbef685b749ec771dae3b78e880a2d8
3
+ size 6020
runs/Mar07_15-38-25_ubuntu-gpu/events.out.tfevents.1678203516.ubuntu-gpu.11468.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42a51da41ac55831c623e1f62ee334a2f3f1921b8d4010a3e6f5d1ec15a616a8
3
+ size 12843
runs/Mar10_15-44-40_ubuntu-gpu/1678463084.350115/events.out.tfevents.1678463084.ubuntu-gpu.9149.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d7a9313923c466b018f56011e2a1513db18e0b98d34db43d4f963eee149e92b
3
+ size 6020
runs/Mar10_15-44-40_ubuntu-gpu/events.out.tfevents.1678463084.ubuntu-gpu.9149.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8430fdfb9789c50f0521202f77360b2041a31d3e0ad655c418870597ae7bf95
3
+ size 14570
runs/Mar11_18-12-23_ubuntu-gpu/1678558347.0917025/events.out.tfevents.1678558347.ubuntu-gpu.2145.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5aa383523d59ce46a8c0fd3bcdf09bf0b1c716a86e8fdc2f33f31080fc9827d
3
+ size 6020
runs/Mar11_18-12-23_ubuntu-gpu/events.out.tfevents.1678558347.ubuntu-gpu.2145.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4dc35a0fa76d5c77ef298ae6e2d62783bd23aae8ef12f4e4fa16d3ca5e15eca
3
+ size 8603
runs/Mar11_18-27-39_ubuntu-gpu/1678559262.7039926/events.out.tfevents.1678559262.ubuntu-gpu.2384.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:314abde53240123b993a19a0b19b5de184d92b653cdac6cf6cf68699a546bac1
3
+ size 6020
runs/Mar11_18-27-39_ubuntu-gpu/events.out.tfevents.1678559262.ubuntu-gpu.2384.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f75ab56524bceb4371f754f724713f887523a08c6a9bc624d1abc02e2d460f07
3
+ size 8449
runs/Mar11_18-33-27_ubuntu-gpu/1678559611.663038/events.out.tfevents.1678559611.ubuntu-gpu.2505.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2077eceb208fe72b66f3e3693f0fd4346032eb9dbe525162f288a9c350ce25e
3
+ size 6020
runs/Mar11_18-33-27_ubuntu-gpu/events.out.tfevents.1678559611.ubuntu-gpu.2505.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68ef5e003ad0838300f3688becdb95e471a86388dbd2a1cec9f8d9f906943187
3
+ size 8449
runs/Mar11_18-44-42_ubuntu-gpu/1678560286.1684813/events.out.tfevents.1678560286.ubuntu-gpu.2746.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19b2e64a6542eff13eade186294eb03a8c6ee8c73c796225ba7b1dfb69e6ebf4
3
+ size 6020
runs/Mar11_18-44-42_ubuntu-gpu/events.out.tfevents.1678560286.ubuntu-gpu.2746.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86a80df0df2a2f8131ec0ea57f2151a52d80bd149dd638302cb7abfd0afc3c77
3
+ size 8449
runs/Mar11_18-53-37_ubuntu-gpu/1678560821.0609927/events.out.tfevents.1678560821.ubuntu-gpu.3517.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b4521fff260a1434a1249fe121422fe3162daf4ed46911798ea4729db35f463
3
+ size 6020
runs/Mar11_18-53-37_ubuntu-gpu/events.out.tfevents.1678560821.ubuntu-gpu.3517.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d134b0bc02b04a369e6739ffb5cc962839ab1f41b5133a108531bd5394be622b
3
+ size 8449
runs/Mar11_19-13-27_ubuntu-gpu/1678562010.4561105/events.out.tfevents.1678562010.ubuntu-gpu.3652.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ae6966552778b8b0654ab4d0087d8afbbed4784852d2472a6f26241340c206e
3
+ size 6020
runs/Mar11_19-13-27_ubuntu-gpu/events.out.tfevents.1678562010.ubuntu-gpu.3652.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1783ec0ffbaf1808271850d7bed0c6d18919c0f640cd4ef817f7b70c834a869
3
+ size 15825
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c008829ba8ec4528478377e53d71c22d8f16b973fdcc141a95c1429916afd526
3
+ size 3771