RankEF/train_hard.py at main · sssszh/RankEF · 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
import io
import logging
import math
import os
os.environ["TOKENIZERS_PARALLELISM"] = "false"
import pprint
import sys
import time
import json
import pickle as pkl
import pdb
import numpy as np
import scipy
from tqdm import tqdm
from datetime import datetime

import transformers
from transformers import EvalPrediction, WEIGHTS_NAME
import torch

from Datasets.CodeT5BaseDataset import CodeT5BaseDataset
from trainers.trainer_hard import Trainer
from models.ModelBasedCodeT5 import CriticModelBasedCodeT5

import torch.multiprocessing
torch.multiprocessing.set_sharing_strategy('file_system')

def run_training(args, train_data, val_data):

    model_path = args.model_path if args.model_path is not None else '{}'.format(args.model)
    print("Loading model from {}...".format(model_path))

    model = CriticModelBasedCodeT5.from_pretrained(model_path)


    print('Finished loading model {}'.format(args.model))

    start_iteration = 0
    train_data.start_iteration = start_iteration
    print(f"Starting main loop")

    training_args = transformers.TrainingArguments(
        label_names=['labels', 'class_labels'],
        output_dir=args.save_dir,
        overwrite_output_dir=True,
        remove_unused_columns=False,
        do_train=True,
        do_eval=False,
        do_predict=False,
        save_strategy='steps',

        num_train_epochs=args.epochs,
        per_device_train_batch_size=args.batch_size_per_replica,
        per_device_eval_batch_size=64,
        gradient_accumulation_steps=args.grad_acc_steps,

        learning_rate=args.lr,
        weight_decay=0.05,
        warmup_steps=200,
        lr_scheduler_type='constant_with_warmup',

        logging_dir=args.save_dir,
        logging_first_step=True,
        logging_steps=args.log_freq,
        save_steps=args.save_freq,
        save_total_limit=args.save_total_limit,

        dataloader_drop_last=True,
        dataloader_num_workers=8,

        local_rank=args.local_rank,

        deepspeed=args.deepspeed,
        fp16=args.fp16,

    )
    trainer = Trainer(
        alpha=0.5,
        model=model,
        args=training_args,
        train_dataset=train_data
    )
    trainer.train()

    if args.local_rank == 0:
        model.save_pretrained(os.path.join(args.save_dir, "final_checkpoint"))

def get_dataset(args, mode="train"):

    if mode == "train":
        dataroot = args.train_path
        with open(args.train_path, 'r') as f:
            problems_1 = f.readlines()
    elif mode == "val":
        dataroot = args.val_path
        with open(args.val_path, 'r') as f:
            problems_1 = f.readlines()

    # problems_2 = problems_2[:100]
    # train in debugging mode with small data split
    if args.db and mode == "train":
        problems_1 = problems_1[:640]
    elif args.db and mode == "val":
        problems_1 = problems_1[:640]

    train_data = CodeT5BaseDataset(
        dataroot=dataroot,
        problems=problems_1,
        model=args.model,
        max_tokens=150,
        max_src_tokens=1024,
    )

    return train_data

def main(args):

    argsdict = vars(args)
    print(pprint.pformat(argsdict))

    os.makedirs(args.save_dir, exist_ok=True)

    # Load dataset
    train_data = get_dataset(args, "train")
    val_data = get_dataset(args, "val")

    # Save args to file
    json.dump(argsdict, open(os.path.join(args.save_dir, "args.json"), 'w'))

    # Load and train model; save model checkpoints
    run_training(args, train_data, val_data)


if __name__ == "__main__":
    from configs.train_codet5_three_configs import *

    main(args)