File size: 1,678 Bytes
4321e7e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os
import sys
import argparse

import torch
import shutil

# Append the utils module path
sys.path.append("../")
from models import PLTNUM


def parse_args():
    """
    Parse command line arguments.
    """
    parser = argparse.ArgumentParser(
        description="Convert the model implemented with nn.Module to a model implemented with transformers' PreTrainedModel."
    )
    parser.add_argument(
        "--model_path",
        type=str,
        help="The path to a model weight which you want to convert.",
    )
    parser.add_argument(
        "--config_and_tokenizer_path",
        type=str,
        help="The path to a config and tokenizer of the model which you want to convert.",
    )
    parser.add_argument(
        "--model",
        type=str,
        help="The name of the base model of the finetuned model",
    )
    parser.add_argument(
        "--output_dir",
        type=str,
        default="./",
        help="Directory to save the prediction.",
    )
    parser.add_argument(
        "--task",
        type=str,
        default="classification",
    )

    return parser.parse_args()


if __name__ == "__main__":
    config = parse_args()

    if not os.path.exists(config.output_dir):
        os.makedirs(config.output_dir)

    model = PLTNUM(config)
    model.load_state_dict(torch.load(config.model_path, map_location="cpu"))

    torch.save(model.state_dict(), os.path.join(config.output_dir, "pytorch_model.bin"))
    for filename in ["config.json", "special_tokens_map.json", "tokenizer_config.json", "vocab.txt"]:
        shutil.copy(os.path.join(config.config_and_tokenizer_path, filename), os.path.join(config.output_dir, filename))