/
main.py
53 lines (44 loc) · 1.93 KB
/
main.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
import sys
import os
from os.path import join
import torch
import random
import numpy as np
project_dir = './'
data_dir = join(project_dir, 'data')
src_dir = join(project_dir, 'src')
model_dir = join(project_dir, 'model')
output_dir = join(project_dir, 'output')
sys.path.append(src_dir)
import Autoencoder
import Data
import Solver
def loss_function(output, x):
recon_x, mu, logVar = output
batchSize = mu.shape[0]
rl = (recon_x - x).pow(2).sum() / batchSize
kld = -0.5 * torch.sum(1 + logVar - mu.pow(2) - logVar.exp()) / batchSize
return rl + kld
if __name__ == '__main__':
# data
dataset = Data.PancreasDataset(data_dir=data_dir, lod=6)
split, samples = 0.9, len(dataset)
dataset_train, dataset_validate = torch.utils.data.random_split(dataset, [int(split * samples), samples - int(split * samples)])
# model, optimizer
model = Autoencoder.VariationalAutoencoder(imageShape=(1, 2 ** 7, 2 ** 7), firstFilterCount=16, act=torch.nn.functional.elu, layerwise=False)
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
# Solver
loadWeights = True
solver = Solver.Solver(model=model, modelDir=model_dir, loadWeights=loadWeights, optimizer=optimizer, criterions=[loss_function, torch.nn.BCELoss()], iouThreshold=0.2)
# routine: train on lod=6, hence (64, 64)
epochs = 101
lod = 6
batchSize = 1024
iterations = 1
dataloader_train = torch.utils.data.DataLoader(dataset=dataset_train, batch_size=batchSize, shuffle=True, num_workers=2)
dataloader_validate = torch.utils.data.DataLoader(dataset=dataset_validate, batch_size=batchSize, shuffle=True, num_workers=2)
for i in range(iterations):
solver.train(dataloader=dataloader_train, epochs=epochs, lod=lod)
solver.evaluate(dataloader=dataloader_validate, lod=lod)
solver.saveReconstructions(dataloader=dataloader_validate, lod=lod, count=10, output_dir=output_dir)
#solver.saveCheckpoint()