Sci-ZSEL/config.json at main · rasel-isu/Sci-ZSEL · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
{
    "world": "lpt",
    "kb_name": "lpt",
    "kb_file": "lpt_kb.json",
    "data_dir": "../datasets/lpt/",
    "saved_model_dir": "../saved_models/lpt/",
    "synonym_key_on_ontology": "synonyms",
    "has_ground_truth": false,
    "has_ent_alt_id": true,
    "exact_match_file": "m1_e1_unq_ents.json",
    "biencoder_top1_file": "top_1_from_biencoder.json",
    "blink": {
        "split_name": "train",
        "base_models": {
            "biencoder": "../saved_models/biencoder_wiki_large.bin",
            "crossencoder": "../saved_models/crossencoder_wiki_large.bin"
        },
        "candidate_generation": {
            "top_k": 64,
            "max_context_length": 64,
            "encode_batch_size": 8,
            "eval_batch_size": 32,
            "bert_model": "bert-large-uncased",
            "has_gt": false
        },
        "retriever": {
            "exp_list": [
                "synonymU(m1_e1)U(m3_e1)_multi_prime_rm_sm_eU(m4_e2)_multi_prime_rm_sm_e"
            ],
            "negative_selection": [
                "add_prch_in_pos_list"
            ],
            "seeds": [
                0
            ],
            "epochs": 1,
            "train_batch_size": 64,
            "encode_batch_size": 8,
            "learning_rate": "2e-05",
            "dropout_rate": 0.2,
            "max_context_length": 128,
            "max_cand_length": 128,
            "max_seq_length": 192,
            "top_k": 64,
            "bert_model": "bert-large-uncased",
            "type_optimization": "all_encoder_layers"
        },
        "reranker": {
            "exp_list": [
                "synonymU(m1_e1)U(m3_e1)_multi_prime_rm_sm_eU(m4_e2)_multi_prime_rm_sm_e"
            ],
            "negative_selection": "only_bienc_20_neg",
            "seeds": [
                0
            ],
            "epochs": 3,
            "train_batch_size": 16,
            "gradient_accumulation_steps": 2,
            "eval_batch_size": 32,
            "learning_rate": "2e-05",
            "dropout_rate": 0.2,
            "max_context_length": 64,
            "max_cand_length": 128,
            "max_seq_length": 192,
            "bert_model": "bert-large-uncased",
            "type_optimization": "all_encoder_layers"
        }
    },
    "res": {
        "pretrained_model": "../../saved_models/zeshel_disambiguation_attention.pt",
        "hf_model": "roberta-base",
        "split_name": "train",
        "exp_list": [
            "synonymU(m1_e1)U(m3_e1)_multi_prime_rm_sm_eU(m4_e2)_multi_prime_rm_sm_e"
        ],
        "seeds": [
            0
        ],
        "lora": false,
        "use_title_during_testing": true,
        "epochs": 3,
        "learning_rate": 0.0001,
        "batch_size": 16,
        "cand_num_train": 21,
        "cand_num": 64,
        "type_loss": "sum_log_nce",
        "max_len": 512,
        "max_ent_len": 256,
        "max_text_len": 256,
        "warmup_proportion": 0.1,
        "weight_decay": 0.01,
        "adam_epsilon": 1e-06,
        "gradient_accumulation_steps": 1,
        "num_workers": 0,
        "clip": 1,
        "info_token_num": 3,
        "gpus": "0,1,2,3",
        "logging_steps": 100,
        "eval_step": 10000
    }
}