config.json
1.5 KB · 68 lines · json Raw
1 {
2 "_name_or_path": "facebook/bart-large-xsum",
3 "_num_labels": 3,
4 "activation_dropout": 0.0,
5 "activation_function": "gelu",
6 "add_bias_logits": false,
7 "add_final_layer_norm": false,
8 "architectures": [
9 "BartForConditionalGeneration"
10 ],
11 "attention_dropout": 0.0,
12 "bos_token_id": 0,
13 "classif_dropout": 0.0,
14 "classifier_dropout": 0.0,
15 "d_model": 1024,
16 "decoder_attention_heads": 16,
17 "decoder_ffn_dim": 4096,
18 "decoder_layerdrop": 0.0,
19 "decoder_layers": 12,
20 "decoder_start_token_id": 2,
21 "dropout": 0.1,
22 "early_stopping": true,
23 "encoder_attention_heads": 16,
24 "encoder_ffn_dim": 4096,
25 "encoder_layerdrop": 0.0,
26 "encoder_layers": 12,
27 "eos_token_id": 2,
28 "eos_token_ids": [
29 2
30 ],
31 "forced_eos_token_id": 2,
32 "gradient_checkpointing": false,
33 "id2label": {
34 "0": "LABEL_0",
35 "1": "LABEL_1",
36 "2": "LABEL_2"
37 },
38 "init_std": 0.02,
39 "is_encoder_decoder": true,
40 "label2id": {
41 "LABEL_0": 0,
42 "LABEL_1": 1,
43 "LABEL_2": 2
44 },
45 "max_length": 62,
46 "max_position_embeddings": 1024,
47 "min_length": 11,
48 "model_type": "bart",
49 "no_repeat_ngram_size": 3,
50 "normalize_before": false,
51 "normalize_embedding": true,
52 "num_beams": 6,
53 "num_hidden_layers": 12,
54 "output_past": true,
55 "pad_token_id": 1,
56 "prefix": " ",
57 "replacing_rate": 0,
58 "scale_embedding": false,
59 "static_position_embeddings": false,
60 "student_decoder_layers": null,
61 "student_encoder_layers": null,
62 "task_specific_params": {},
63 "torch_dtype": "float32",
64 "transformers_version": "4.10.0",
65 "use_cache": true,
66 "vocab_size": 50264
67 }
68