XME/run.py at master · XME-anonymous/XME · 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
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
import copy
import random
import importlib
import logging

import hydra
from omegaconf import OmegaConf
import numpy as np
import torch
import utils


# from trainer import EditTrainer
import models


OmegaConf.register_new_resolver("uuid", lambda: utils.uuid())


logging.basicConfig(format='%(asctime)s - %(levelname)s [%(filename)s:%(lineno)d] %(message)s',
                    level=logging.INFO)
LOG = logging.getLogger(__name__)


def add_padding(tokenizer, model):
    tokenizer.add_special_tokens({'pad_token': '[PAD]'})
    model.resize_token_embeddings(len(tokenizer))
    model.transformer.word_embeddings.weight.data[-1] = model.transformer.word_embeddings.weight.data.mean(0) # Bloom
    # model.transformer.wte.weight.data[-1] = model.transformer.wte.weight.data.mean(0) # GPT Style


@hydra.main(config_path='config', config_name='config')
def run(config):
    # print(config.test_file)
    LOG.info(f"\n\n{OmegaConf.to_yaml(config)}\n")
    base_dir = hydra.utils.get_original_cwd()
    LOG.info(f"Project base directory: {base_dir}")

    random.seed(config.seed)
    np.random.seed(config.seed)
    torch.manual_seed(config.seed)

    model = models.get_model(config)
    # for name, _ in model.named_parameters():
    #     print(name)
    tokenizer = models.get_tokenizer(config)

    import transformers
    if isinstance(model, transformers.GPT2ForSequenceClassification):
        model.config.pad_token_id = model.config.eos_token_id

    if config.task == "gen" or config.task == "wiki":
        add_padding(tokenizer, model)
        from data_classes.wiki import GenDataset

        train_set = GenDataset("train", tokenizer, config, config.data.path, pct=10)
        val_set = GenDataset("validation", tokenizer, config, config.data.path, pct=10)
    elif config.task == "fc" or config.task == "fever":
        from data_classes.fever import BinaryAugmentedKILT

        if config.tests:
            # from data_classes.fever_test import BinaryAugmentedKILT
            train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/{config.train_set}", config)
            val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/{config.val_set}", config)
        else:
            if config.lang == "english":
                # English
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever-train-kilt.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever-dev-kilt.jsonl", config)
            elif config.lang == "hindi":
                # Hindi
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - hindi-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - hindi-10K.jsonl", config)

            elif config.lang == "spanish":
                # Spanish
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - spanish-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev-spanish-10K.jsonl", config)

            elif config.lang == "french":
                # French
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - french-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - french-10K.jsonl", config)

            elif config.lang == "bengali":
                # Bengali
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - bengali-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - bengali-10K.jsonl", config)

            elif config.lang == "gujarati":
                # Gujarati
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - gujarati-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - gujarati-10K.jsonl", config)

            elif config.lang == "malayalam":
                # Malayalam
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - malayalam-1L_lang_code.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - malayalam-10K_lang_code.jsonl", config)

            elif config.lang == "tamil":
                # Tamil
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - tamil-1L_lang_code.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - tamil-10K_lang_code.jsonl", config)

            elif config.lang == "kannada":
                # Kannada
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - kannada-1L_lang_code.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - kannada-10K_lang_code.jsonl", config)

            elif config.lang == "chinese":
                # Chinese
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - chinese-1L_lang_code.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - chinese-10K_lang_code.jsonl", config)

            elif config.lang == "arabic":
                # Arabic
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - arabic-1L_lang_code.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - arabic-10K_lang_code.jsonl", config)

            elif config.lang == "mixed":
                # Mixed
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - mixed-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - mixed-10K.jsonl", config)

            elif config.lang == "inverse":
                # Mixed
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - mixed_inv-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - mixed_inv-10K.jsonl", config)

            elif config.lang == "inverse-xlm":
                # Mixed
                train_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_train - inverse-xlm-1L.jsonl", config)
                val_set = BinaryAugmentedKILT(tokenizer, f"{base_dir}/data/fever/fever_dev - inverse-xlm-10K.jsonl", config)

    elif config.task == "qa" or config.task == "zsre":
        from data_classes.zsre import Seq2SeqAugmentedKILT

        train_set = Seq2SeqAugmentedKILT(tokenizer, f"{base_dir}/data/zsre/structured_zeroshot-train-new_annotated_final.jsonl",
                                         config)
        val_set = Seq2SeqAugmentedKILT(tokenizer, f"{base_dir}/data/zsre/structured_zeroshot-dev-new_annotated_final.jsonl",
                                       config)

    else:
        raise ValueError(f"Unrecognized task {config.task}")

    alg_module = importlib.import_module(f"algs.{config.alg}")
    LOG.info(f"Loading class {config.alg.upper()} from module {alg_module}")
    AlgClass = getattr(alg_module, config.alg.upper())
    alg = AlgClass(model, config, lambda: copy.deepcopy(model))

    if config.alg == "ft" and config.ft.locality.enabled:
        if config.ft.locality.oracle:
            alg.loc_sampler = train_set.edit_generator(config.ft.locality.batch_size + 1)
        else:
            state = np.random.get_state()
            np.random.seed(0)
            loc_batch = next(train_set.edit_generator(config.ft.locality.batch_size + 1))["loc"]
            np.random.set_state(state)
            alg.loc_ids = loc_batch["input_ids"]
            alg.loc_masks = loc_batch["attention_mask"]

    if config.tests:
        from tester import EditTrainer
    else:
        from trainer import EditTrainer

    trainer = EditTrainer(alg, config, train_set, val_set)
    trainer.run()


if __name__ == "__main__":
    run()