-
Notifications
You must be signed in to change notification settings - Fork 9
/
pretrain.py
51 lines (42 loc) · 1.11 KB
/
pretrain.py
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
from pathlib import Path
import pandas as pd
from tqdm import tqdm # type: ignore
import torch
import os
from other import create_features, Trainer, RNN, Transformer, NN_17, GRU_P
model_name = os.environ.get("MODEL", "FSRSv3")
short_term = os.environ.get("SHORT")
secs_ivl = os.environ.get("SECS_IVL")
file_name = (
model_name + ("-short" if short_term else "") + ("-secs" if secs_ivl else "")
)
if model_name == "GRU":
model = RNN
elif model_name == "GRU-P":
model = GRU_P
elif model_name == "Transformer":
model = Transformer
elif model_name == "NN-17":
model = NN_17
total = 0
for param in model().parameters():
total += param.numel()
print(total)
df_list = []
for i in tqdm(range(1, 101)):
file = Path(f"../FSRS-Anki-20k/dataset/1/{i}.csv")
dataset = pd.read_csv(file)
dataset = create_features(dataset, model_name=model_name)
df_list.append(dataset)
df = pd.concat(df_list, axis=0)
trainer = Trainer(
model(),
df,
None,
n_epoch=32,
lr=4e-2,
wd=1e-4,
batch_size=65536,
)
trainer.train()
torch.save(trainer.model.state_dict(), f"./{file_name}_pretrain.pth")