system HF staff commited on
Commit
f28b2fd
1 Parent(s): a161146

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +174 -0
config.json ADDED
@@ -0,0 +1,174 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "RagSequenceForGeneration"
4
+ ],
5
+ "dataset": "wiki_dpr",
6
+ "dataset_split": "train",
7
+ "do_marginalize": false,
8
+ "doc_sep": " // ",
9
+ "exclude_bos_score": false,
10
+ "generator": {
11
+ "_num_labels": 3,
12
+ "activation_dropout": 0.0,
13
+ "activation_function": "gelu",
14
+ "add_bias_logits": false,
15
+ "add_cross_attention": false,
16
+ "add_final_layer_norm": false,
17
+ "architectures": [
18
+ "BartForConditionalGeneration"
19
+ ],
20
+ "attention_dropout": 0.1,
21
+ "bad_words_ids": null,
22
+ "bos_token_id": 0,
23
+ "chunk_size_feed_forward": 0,
24
+ "classif_dropout": 0.0,
25
+ "d_model": 1024,
26
+ "decoder_attention_heads": 16,
27
+ "decoder_ffn_dim": 4096,
28
+ "decoder_layerdrop": 0.0,
29
+ "decoder_layers": 12,
30
+ "decoder_start_token_id": 2,
31
+ "do_sample": false,
32
+ "dropout": 0.1,
33
+ "early_stopping": false,
34
+ "encoder_attention_heads": 16,
35
+ "encoder_ffn_dim": 4096,
36
+ "encoder_layerdrop": 0.0,
37
+ "encoder_layers": 12,
38
+ "eos_token_id": 2,
39
+ "extra_pos_embeddings": 2,
40
+ "finetuning_task": null,
41
+ "force_bos_token_to_be_generated": false,
42
+ "id2label": {
43
+ "0": "LABEL_0",
44
+ "1": "LABEL_1",
45
+ "2": "LABEL_2"
46
+ },
47
+ "init_std": 0.02,
48
+ "is_decoder": false,
49
+ "is_encoder_decoder": true,
50
+ "label2id": {
51
+ "LABEL_0": 0,
52
+ "LABEL_1": 1,
53
+ "LABEL_2": 2
54
+ },
55
+ "length_penalty": 1.0,
56
+ "max_length": 20,
57
+ "max_position_embeddings": 1024,
58
+ "min_length": 0,
59
+ "model_type": "bart",
60
+ "no_repeat_ngram_size": 0,
61
+ "normalize_before": false,
62
+ "normalize_embedding": true,
63
+ "num_beams": 1,
64
+ "num_hidden_layers": 12,
65
+ "num_return_sequences": 1,
66
+ "output_attentions": false,
67
+ "output_hidden_states": false,
68
+ "output_past": false,
69
+ "pad_token_id": 1,
70
+ "prefix": " ",
71
+ "pruned_heads": {},
72
+ "repetition_penalty": 1.0,
73
+ "return_dict": true,
74
+ "scale_embedding": false,
75
+ "static_position_embeddings": false,
76
+ "task_specific_params": {
77
+ "summarization": {
78
+ "early_stopping": true,
79
+ "length_penalty": 2.0,
80
+ "max_length": 142,
81
+ "min_length": 56,
82
+ "no_repeat_ngram_size": 3,
83
+ "num_beams": 4
84
+ }
85
+ },
86
+ "temperature": 1.0,
87
+ "tie_encoder_decoder": false,
88
+ "tie_word_embeddings": true,
89
+ "tokenizer_class": null,
90
+ "top_k": 50,
91
+ "top_p": 1.0,
92
+ "torchscript": false,
93
+ "use_bfloat16": false,
94
+ "use_cache": true,
95
+ "vocab_size": 50265,
96
+ "xla_device": null
97
+ },
98
+ "index_name": "legacy",
99
+ "is_encoder_decoder": true,
100
+ "label_smoothing": 0.0,
101
+ "max_combined_length": 300,
102
+ "model_type": "rag",
103
+ "n_docs": 5,
104
+ "passages_path": null,
105
+ "question_encoder": {
106
+ "add_cross_attention": false,
107
+ "architectures": [
108
+ "DPRQuestionEncoder"
109
+ ],
110
+ "attention_probs_dropout_prob": 0.1,
111
+ "bad_words_ids": null,
112
+ "bos_token_id": null,
113
+ "chunk_size_feed_forward": 0,
114
+ "decoder_start_token_id": null,
115
+ "do_sample": false,
116
+ "early_stopping": false,
117
+ "eos_token_id": null,
118
+ "finetuning_task": null,
119
+ "gradient_checkpointing": false,
120
+ "hidden_act": "gelu",
121
+ "hidden_dropout_prob": 0.1,
122
+ "hidden_size": 768,
123
+ "id2label": {
124
+ "0": "LABEL_0",
125
+ "1": "LABEL_1"
126
+ },
127
+ "initializer_range": 0.02,
128
+ "intermediate_size": 3072,
129
+ "is_decoder": false,
130
+ "is_encoder_decoder": false,
131
+ "label2id": {
132
+ "LABEL_0": 0,
133
+ "LABEL_1": 1
134
+ },
135
+ "layer_norm_eps": 1e-12,
136
+ "length_penalty": 1.0,
137
+ "max_length": 20,
138
+ "max_position_embeddings": 512,
139
+ "min_length": 0,
140
+ "model_type": "dpr",
141
+ "no_repeat_ngram_size": 0,
142
+ "num_attention_heads": 12,
143
+ "num_beams": 1,
144
+ "num_hidden_layers": 12,
145
+ "num_return_sequences": 1,
146
+ "output_attentions": false,
147
+ "output_hidden_states": false,
148
+ "pad_token_id": 0,
149
+ "prefix": null,
150
+ "projection_dim": 0,
151
+ "pruned_heads": {},
152
+ "repetition_penalty": 1.0,
153
+ "return_dict": false,
154
+ "task_specific_params": null,
155
+ "temperature": 1.0,
156
+ "tie_encoder_decoder": false,
157
+ "tie_word_embeddings": true,
158
+ "tokenizer_class": null,
159
+ "top_k": 50,
160
+ "top_p": 1.0,
161
+ "torchscript": false,
162
+ "type_vocab_size": 2,
163
+ "use_bfloat16": false,
164
+ "use_cache": true,
165
+ "vocab_size": 30522,
166
+ "xla_device": null
167
+ },
168
+ "reduce_loss": false,
169
+ "retrieval_batch_size": 8,
170
+ "retrieval_vector_size": 768,
171
+ "title_sep": " / ",
172
+ "use_dummy_dataset": false,
173
+ "vocab_size": null
174
+ }