-
Notifications
You must be signed in to change notification settings - Fork 8
/
train.py
70 lines (62 loc) · 2.73 KB
/
train.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
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
# coding=utf-8
# Copyright 2018 jose.fonollosa@upc.edu
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import print_function, division
import torch
import torch.nn as nn
def train(loader, model, optimizer, epoch, cuda, log_interval, weight=None, verbose=True):
model.train()
global_epoch_loss = 0
for batch_idx, (_, data, target) in enumerate(loader):
criterion = nn.CrossEntropyLoss(weight=weight)
if cuda:
data, target = data.cuda(), target.cuda()
criterion = criterion.cuda()
optimizer.zero_grad()
output = model(data)
loss = criterion(output, target)
loss.backward()
optimizer.step()
global_epoch_loss += loss.data.item()
if verbose:
if batch_idx % log_interval == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(loader.dataset), 100. * batch_idx / len(loader), loss.data.item()))
return global_epoch_loss / len(loader.dataset)
def test(loader, model, cuda, verbose=True, data_set='Test', save=None):
model.eval()
test_loss = 0
correct = 0
if save is not None:
csv = open(save, 'wt')
print('fname,label', file=csv)
with torch.no_grad():
for keys, data, target in loader:
criterion = nn.CrossEntropyLoss(reduction='sum')
if cuda:
data, target = data.cuda(), target.cuda()
criterion = criterion.cuda()
output = model(data)
test_loss += criterion(output, target).data.item() # sum up batch loss
pred = output.data.max(1, keepdim=True)[1] # get the index of the max log-probability
correct += pred.eq(target.data.view_as(pred)).cpu().sum()
if save is not None:
for i, key in enumerate(keys):
print(key+'.wav,'+loader.dataset.classes[int(pred[i])], file=csv)
test_loss /= len(loader.dataset)
accuracy = float(correct) / len(loader.dataset)
if verbose:
print('\n{} set: Average loss: {:.4f}, Accuracy: {}/{} ({:.1f}%)\n'.format(
data_set, test_loss, correct, len(loader.dataset), 100 * accuracy))
return test_loss, accuracy