Upload Pix2StructForConditionalGeneration

#1
Files changed (3) hide show
  1. config.json +180 -0
  2. generation_config.json +8 -0
  3. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,180 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_commit_hash": "b46b9db89c0990d571322cc327a4633e5aa78162",
3
+ "_name_or_path": "google/deplot",
4
+ "architectures": [
5
+ "Pix2StructForConditionalGeneration"
6
+ ],
7
+ "decoder_start_token_id": 0,
8
+ "eos_token_id": 1,
9
+ "initializer_factor": 1.0,
10
+ "initializer_range": 0.02,
11
+ "is_encoder_decoder": true,
12
+ "is_vqa": false,
13
+ "model_type": "pix2struct",
14
+ "pad_token_id": 0,
15
+ "text_config": {
16
+ "_name_or_path": "",
17
+ "add_cross_attention": false,
18
+ "architectures": null,
19
+ "bad_words_ids": null,
20
+ "begin_suppress_tokens": null,
21
+ "bos_token_id": null,
22
+ "chunk_size_feed_forward": 0,
23
+ "cross_attention_hidden_size": null,
24
+ "d_ff": 2048,
25
+ "d_kv": 64,
26
+ "decoder_start_token_id": 0,
27
+ "dense_act_fn": "gelu_new",
28
+ "diversity_penalty": 0.0,
29
+ "do_sample": false,
30
+ "dropout_rate": 0.1,
31
+ "early_stopping": false,
32
+ "encoder_hidden_size": 768,
33
+ "encoder_no_repeat_ngram_size": 0,
34
+ "eos_token_id": 1,
35
+ "exponential_decay_length_penalty": null,
36
+ "finetuning_task": null,
37
+ "forced_bos_token_id": null,
38
+ "forced_eos_token_id": null,
39
+ "hidden_size": 768,
40
+ "id2label": {
41
+ "0": "LABEL_0",
42
+ "1": "LABEL_1"
43
+ },
44
+ "initializer_factor": 1.0,
45
+ "initializer_range": 0.02,
46
+ "is_decoder": true,
47
+ "is_encoder_decoder": false,
48
+ "label2id": {
49
+ "LABEL_0": 0,
50
+ "LABEL_1": 1
51
+ },
52
+ "layer_norm_epsilon": 1e-06,
53
+ "length_penalty": 1.0,
54
+ "max_length": 20,
55
+ "min_length": 0,
56
+ "model_type": "pix2struct_text_model",
57
+ "no_repeat_ngram_size": 0,
58
+ "num_beam_groups": 1,
59
+ "num_beams": 1,
60
+ "num_heads": 12,
61
+ "num_layers": 12,
62
+ "num_return_sequences": 1,
63
+ "output_attentions": false,
64
+ "output_hidden_states": false,
65
+ "output_scores": false,
66
+ "pad_token_id": 0,
67
+ "prefix": null,
68
+ "problem_type": null,
69
+ "pruned_heads": {},
70
+ "relative_attention_max_distance": 128,
71
+ "relative_attention_num_buckets": 32,
72
+ "remove_invalid_values": false,
73
+ "repetition_penalty": 1.0,
74
+ "return_dict": true,
75
+ "return_dict_in_generate": false,
76
+ "sep_token_id": null,
77
+ "suppress_tokens": null,
78
+ "task_specific_params": null,
79
+ "temperature": 1.0,
80
+ "tf_legacy_loss": false,
81
+ "tie_encoder_decoder": false,
82
+ "tie_word_embeddings": false,
83
+ "tokenizer_class": null,
84
+ "top_k": 50,
85
+ "top_p": 1.0,
86
+ "torch_dtype": null,
87
+ "torchscript": false,
88
+ "transformers_version": "4.29.0",
89
+ "typical_p": 1.0,
90
+ "use_bfloat16": false,
91
+ "use_cache": false,
92
+ "vocab_size": 50244
93
+ },
94
+ "tie_word_embeddings": false,
95
+ "torch_dtype": "float32",
96
+ "transformers_version": null,
97
+ "vision_config": {
98
+ "_name_or_path": "",
99
+ "add_cross_attention": false,
100
+ "architectures": null,
101
+ "attention_dropout": 0.0,
102
+ "bad_words_ids": null,
103
+ "begin_suppress_tokens": null,
104
+ "bos_token_id": null,
105
+ "chunk_size_feed_forward": 0,
106
+ "cross_attention_hidden_size": null,
107
+ "d_ff": 2048,
108
+ "d_kv": 64,
109
+ "decoder_start_token_id": null,
110
+ "dense_act_fn": "gelu_new",
111
+ "diversity_penalty": 0.0,
112
+ "do_sample": false,
113
+ "dropout_rate": 0.0,
114
+ "early_stopping": false,
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_size": 768,
122
+ "id2label": {
123
+ "0": "LABEL_0",
124
+ "1": "LABEL_1"
125
+ },
126
+ "initializer_factor": 1.0,
127
+ "initializer_range": 0.02,
128
+ "is_decoder": false,
129
+ "is_encoder_decoder": false,
130
+ "label2id": {
131
+ "LABEL_0": 0,
132
+ "LABEL_1": 1
133
+ },
134
+ "layer_norm_bias": false,
135
+ "layer_norm_eps": 1e-06,
136
+ "length_penalty": 1.0,
137
+ "max_length": 20,
138
+ "min_length": 0,
139
+ "model_type": "pix2struct_vision_model",
140
+ "no_repeat_ngram_size": 0,
141
+ "num_attention_heads": 12,
142
+ "num_beam_groups": 1,
143
+ "num_beams": 1,
144
+ "num_channels": 3,
145
+ "num_hidden_layers": 12,
146
+ "num_return_sequences": 1,
147
+ "output_attentions": false,
148
+ "output_hidden_states": false,
149
+ "output_scores": false,
150
+ "pad_token_id": null,
151
+ "patch_embed_hidden_size": 768,
152
+ "patch_size": 16,
153
+ "prefix": null,
154
+ "problem_type": null,
155
+ "projection_dim": 768,
156
+ "pruned_heads": {},
157
+ "relative_attention_max_distance": 128,
158
+ "relative_attention_num_buckets": 32,
159
+ "remove_invalid_values": false,
160
+ "repetition_penalty": 1.0,
161
+ "return_dict": true,
162
+ "return_dict_in_generate": false,
163
+ "sep_token_id": null,
164
+ "seq_len": 4096,
165
+ "suppress_tokens": null,
166
+ "task_specific_params": null,
167
+ "temperature": 1.0,
168
+ "tf_legacy_loss": false,
169
+ "tie_encoder_decoder": false,
170
+ "tie_word_embeddings": true,
171
+ "tokenizer_class": null,
172
+ "top_k": 50,
173
+ "top_p": 1.0,
174
+ "torch_dtype": null,
175
+ "torchscript": false,
176
+ "transformers_version": "4.29.0",
177
+ "typical_p": 1.0,
178
+ "use_bfloat16": false
179
+ }
180
+ }
generation_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "decoder_start_token_id": 0,
4
+ "eos_token_id": 1,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.29.0",
7
+ "use_cache": false
8
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f33903b8e13d8162ca0f1a48429c49007dd0caf3b4d4ad2fa5580cf76b8853e
3
+ size 1129238081