persona-chat/transformer/train.py at master · machinereading/persona-chat · 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
import torch
import random
from model.utils import load_openai_weights, set_seed, f1_score
from model.transformer_model import TransformerModel
from model.trainer import Trainer
from model.text import BPEVocab
from model.dataset import FacebookDataset
from config import get_model_config, get_trainer_config
import os


def main():
    model_config = get_model_config()
    trainer_config = get_trainer_config()

    set_seed(trainer_config.seed)
    device = torch.device(trainer_config.device)

    vocab = BPEVocab.from_files(model_config.bpe_vocab_path, model_config.bpe_codes_path)

    transformer = TransformerModel(n_layers=model_config.n_layers,
                                   n_embeddings=len(vocab),
                                   n_pos_embeddings=model_config.n_pos_embeddings,
                                   embeddings_size=model_config.embeddings_size,
                                   padding_idx=vocab.pad_id,
                                   n_heads=model_config.n_heads,
                                   dropout=model_config.dropout,
                                   embed_dropout=model_config.embed_dropout,
                                   attn_dropout=model_config.attn_dropout,
                                   ff_dropout=model_config.ff_dropout,
                                   bos_id=vocab.bos_id,
                                   eos_id=vocab.eos_id,
                                   max_seq_len=model_config.max_seq_len,
                                   beam_size=model_config.beam_size,
                                   length_penalty=model_config.length_penalty,
                                   n_segments=model_config.n_segments,
                                   annealing_topk=model_config.annealing_topk,
                                   annealing=model_config.annealing,
                                   diversity_coef=model_config.diversity_coef,
                                   diversity_groups=model_config.diversity_groups)

    if not trainer_config.load_last:
        load_openai_weights(transformer.transformer_module,
                            trainer_config.openai_parameters_dir,
                            n_special_tokens=vocab.n_special_tokens)
        print('OpenAI weights loaded from {}'.format(trainer_config.openai_parameters_dir))

    train_dataset = FacebookDataset(trainer_config.train_datasets, vocab, transformer.n_pos_embeddings - 1)
    test_dataset = FacebookDataset(trainer_config.test_datasets, vocab, transformer.n_pos_embeddings - 1)

    model_trainer = Trainer(transformer,
                            train_dataset,
                            test_dataset,
                            batch_size=trainer_config.batch_size,
                            batch_split=trainer_config.batch_split,
                            lr=trainer_config.lr,
                            lr_warmup=trainer_config.lr_warmup,
                            lm_weight=trainer_config.lm_weight,
                            risk_weight=trainer_config.risk_weight,
                            n_jobs=trainer_config.n_jobs,
                            clip_grad=trainer_config.clip_grad,
                            device=device,
                            ignore_idxs=vocab.special_tokens_ids)

    start_epoch = 1
    if trainer_config.load_last:
        if trainer_config.load_default:
            ckpt = trainer_config.default_checkpoint_path
            state_dict = torch.load(ckpt, map_location=device)
        else:
            ckpts = os.listdir(trainer_config.trained_checkpoint_path)
            ckpts.sort(reverse=True)
            for ckpt in ckpts:
                if ckpt[:6] == 'epoch_':
                    start_epoch += int(ckpt[6:])
                    break
            state_dict = torch.load(os.path.join(trainer_config.trained_checkpoint_path, ckpt), map_location=device)
        model_trainer.load_state_dict(state_dict)
        print('Weights loaded from {}'.format(ckpt))


    # helpers -----------------------------------------------------
    def save_func(epoch):
        #torch.save(model_trainer.state_dict(), trainer_config.last_checkpoint_path)
        if epoch % 20 == 0:
            save_path = os.path.join(trainer_config.trained_checkpoint_path, 'epoch_{}'.format(epoch))
            torch.save(model_trainer.state_dict(), save_path)

    def sample_text_func(epoch):
        n_samples = 5
        samples_idxs = random.sample(range(len(test_dataset)), n_samples)
        samples = [test_dataset[idx] for idx in samples_idxs]
        for persona_info, dialog, target in samples:
            contexts = [torch.tensor([c], dtype=torch.long, device=model_trainer.device) for c in [persona_info, dialog] if len(c) > 0]
            prediction = model_trainer.model.predict(contexts)[0]

            persona_info_str = vocab.ids2string(persona_info[1:-1])
            dialog_str = vocab.ids2string(dialog)
            dialog_str = dialog_str.replace(vocab.talker1_bos, '\n\t- ').replace(vocab.talker2_bos, '\n\t- ')
            dialog_str = dialog_str.replace(vocab.talker1_eos, '').replace(vocab.talker2_eos, '')
            target_str = vocab.ids2string(target[1:-1])
            prediction_str = vocab.ids2string(prediction)

            print('\n')
            print('Persona info:\n\t{}'.format(persona_info_str))
            print('Dialog:{}'.format(dialog_str))
            print('Target:\n\t{}'.format(target_str))
            print('Prediction:\n\t{}'.format(prediction_str))

    def test_func(epoch):
        if epoch % trainer_config.test_period == 0:
            metric_funcs = {'f1_score': f1_score}
            model_trainer.test(metric_funcs)

    def f1_risk(predictions, targets):
        scores = f1_score(predictions, targets, average=False)
        return [1-s for s in scores]

    # helpers -----------------------------------------------------

    if trainer_config.train:
        try:
            model_trainer.train(trainer_config.n_epochs, after_epoch_funcs=[save_func, sample_text_func, test_func], risk_func=f1_risk, start_epoch=start_epoch)
        except (KeyboardInterrupt, Exception, RuntimeError) as e:
            torch.save(model_trainer.state_dict(), trainer_config.interrupt_checkpoint_path)
            raise e
    else:
        sample_text_func(100)
        test_func(100)


if __name__ == '__main__':
    main()