Hugging Face's logo

warshakhan
/
donut-base-SynHMP

Image-Text-to-Text
Transformers
PyTorch
TensorBoard
vision-encoder-decoder
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
warshakhan commited on
Commit
91fa94f
•
1 Parent(s): 8726b48

Training in progress, epoch 1

Browse files
Files changed (14) hide show
  1. .gitignore +1 -0
  2. config.json +195 -0
  3. pytorch_model.bin +3 -0
  4. runs/Jul24_08-17-46_DESKTOP-QS6484B/events.out.tfevents.1690168694.DESKTOP-QS6484B.26996.0 +3 -0
  5. runs/Jul24_08-24-59_DESKTOP-QS6484B/events.out.tfevents.1690169115.DESKTOP-QS6484B.26996.1 +3 -0
  6. runs/Jul24_08-24-59_DESKTOP-QS6484B/events.out.tfevents.1690169266.DESKTOP-QS6484B.26996.2 +3 -0
  7. runs/Jul24_08-30-22_DESKTOP-QS6484B/events.out.tfevents.1690169426.DESKTOP-QS6484B.26996.3 +3 -0
  8. runs/Jul24_09-01-28_DESKTOP-QS6484B/events.out.tfevents.1690171296.DESKTOP-QS6484B.13552.0 +3 -0
  9. runs/Jul24_09-12-22_DESKTOP-QS6484B/events.out.tfevents.1690171948.DESKTOP-QS6484B.13552.1 +3 -0
  10. runs/Jul24_09-14-56_DESKTOP-QS6484B/events.out.tfevents.1690172102.DESKTOP-QS6484B.13552.2 +3 -0
  11. runs/Jul24_09-16-17_DESKTOP-QS6484B/events.out.tfevents.1690172181.DESKTOP-QS6484B.13552.3 +3 -0
  12. runs/Jul24_09-17-00_DESKTOP-QS6484B/events.out.tfevents.1690172225.DESKTOP-QS6484B.13552.4 +3 -0
  13. runs/Jul24_09-37-00_DESKTOP-QS6484B/events.out.tfevents.1690173424.DESKTOP-QS6484B.18076.0 +3 -0
  14. 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": "4.32.0.dev0",
87
+ "typical_p": 1.0,
88
+ "use_bfloat16": false,
89
+ "use_cache": true,
90
+ "vocab_size": 57561
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": "4.32.0.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 sha256:231733bd627c9a587be679a6228ea787614c471b45a2aadb16bc26a096e8b2ac
3
+ size 809326361
runs/Jul24_08-17-46_DESKTOP-QS6484B/events.out.tfevents.1690168694.DESKTOP-QS6484B.26996.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a56d716744cd6c828aa378901bf743cbb54eba22483a3d83f8710e92e3f12d53
3
+ size 8601
runs/Jul24_08-24-59_DESKTOP-QS6484B/events.out.tfevents.1690169115.DESKTOP-QS6484B.26996.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54e10fd77eccc22f3a30d4bf1ddc02013da28ed9811607f2926613b233b8a0ce
3
+ size 8601
runs/Jul24_08-24-59_DESKTOP-QS6484B/events.out.tfevents.1690169266.DESKTOP-QS6484B.26996.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ba3a9db00ce5950b8d219c593fd42d6d8d02fcd5f21c4de9b8eef89d2374473
3
+ size 8601
runs/Jul24_08-30-22_DESKTOP-QS6484B/events.out.tfevents.1690169426.DESKTOP-QS6484B.26996.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca84ff202c5a32ae4b6c094753c8c2b7f9a395eb5fe63c642481ca37a05fba60
3
+ size 8601
runs/Jul24_09-01-28_DESKTOP-QS6484B/events.out.tfevents.1690171296.DESKTOP-QS6484B.13552.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aaf36e932dffc57ee2e719406b3e57de5f871c1ddd284dda47b4723b73777c8a
3
+ size 25627
runs/Jul24_09-12-22_DESKTOP-QS6484B/events.out.tfevents.1690171948.DESKTOP-QS6484B.13552.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4d14dd70d880161e44b5fd32b8efd916cc4afb9d6461b0588bb63cca2d4bd88
3
+ size 8602
runs/Jul24_09-14-56_DESKTOP-QS6484B/events.out.tfevents.1690172102.DESKTOP-QS6484B.13552.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:882515047551fb66559ae6dce2d00ec34c771bd667215a82c859790e5f3e6fb3
3
+ size 8602
runs/Jul24_09-16-17_DESKTOP-QS6484B/events.out.tfevents.1690172181.DESKTOP-QS6484B.13552.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ef532e14a25e60f58fc67de3571aeaaeae671c729075a4ba594056c1902677d
3
+ size 8601
runs/Jul24_09-17-00_DESKTOP-QS6484B/events.out.tfevents.1690172225.DESKTOP-QS6484B.13552.4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd49292b1608612abe0ab5335f8a246fdaad7ae02e7db611373caeccdd4c73ac
3
+ size 8601
runs/Jul24_09-37-00_DESKTOP-QS6484B/events.out.tfevents.1690173424.DESKTOP-QS6484B.18076.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a0a10fe6871e4eb4e9f72413a3994db4784ae531a24d88cef6d1f89ad6607cd
3
+ size 9854
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:143da3641894a5045c14093d60a08a9c6467d5897a1365b0f2dba0310060b9cd
3
+ size 4155