cube-studio/myapp/example/pipeline/pytorch/demo.py
2023-12-11 09:59:16 +08:00

187 lines
7.6 KiB
Python

from __future__ import print_function
# pip install tensorboardX torch torchvision --index-url https://mirrors.aliyun.com/pypi/simple
import argparse
import os
import datetime, time
from tensorboardX import SummaryWriter
from torchvision import datasets, transforms
import torch
import torch.distributed as dist
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import pysnooper
WORLD_SIZE = int(os.environ.get('WORLD_SIZE', 1))
# 可以先下载数据到data目录
class MyMNIST(datasets.MNIST):
mirrors = ['https://docker-76009.sz.gfp.tencent-cloud.com/kubeflow/pytorch/example/data/']
# resources = [
# ("https://docker-76009.sz.gfp.tencent-cloud.com/kubeflow/pytorch/example/data/train-images-idx3-ubyte.gz",
# "f68b3c2dcbeaaa9fbdd348bbdeb94873"),
# ("https://docker-76009.sz.gfp.tencent-cloud.com/kubeflow/pytorch/example/data/train-labels-idx1-ubyte.gz",
# "d53e105ee54ea40749a09fcbcd1e9432"),
# ("https://docker-76009.sz.gfp.tencent-cloud.com/kubeflow/pytorch/example/data/t10k-images-idx3-ubyte.gz",
# "9fb629c4189551a2d022fa330f9573f3"),
# ("https://docker-76009.sz.gfp.tencent-cloud.com/kubeflow/pytorch/example/data/t10k-labels-idx1-ubyte.gz",
# "ec29112dd5afa0611ce80d1b7f02629c")
# ]
# 定义模型框架
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 20, 5, 1)
self.conv2 = nn.Conv2d(20, 50, 5, 1)
self.fc1 = nn.Linear(4 * 4 * 50, 500)
self.fc2 = nn.Linear(500, 10)
# 前向计算
def forward(self, x):
x = F.relu(self.conv1(x))
x = F.max_pool2d(x, 2, 2)
x = F.relu(self.conv2(x))
x = F.max_pool2d(x, 2, 2)
x = x.view(-1, 4 * 4 * 50)
x = F.relu(self.fc1(x))
x = self.fc2(x)
return F.log_softmax(x, dim=1)
# 训练过程
def train(args, model, device, train_loader, optimizer, epoch, writer):
model.train()
for batch_idx, (data, target) in enumerate(train_loader):
data, target = data.to(device), target.to(device)
optimizer.zero_grad()
output = model(data)
loss = F.nll_loss(output, target)
loss.backward()
optimizer.step()
if batch_idx % args.log_interval == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tloss={:.4f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader), loss.item()))
niter = epoch * len(train_loader) + batch_idx
writer.add_scalar('loss', loss.item(), niter)
# 测试计算计算损失值
def test(args, model, device, test_loader, writer, epoch):
model.eval()
test_loss = 0
correct = 0
with torch.no_grad():
for data, target in test_loader:
data, target = data.to(device), target.to(device)
output = model(data)
test_loss += F.nll_loss(output, target, reduction='sum').item() # sum up batch loss
pred = output.max(1, keepdim=True)[1] # get the index of the max log-probability
correct += pred.eq(target.view_as(pred)).sum().item()
test_loss /= len(test_loader.dataset)
print('\naccuracy={:.4f}\n'.format(float(correct) / len(test_loader.dataset)))
writer.add_scalar('accuracy', float(correct) / len(test_loader.dataset), epoch)
# 计算是不是分布式
def should_distribute():
return dist.is_available() and WORLD_SIZE > 1
# 计算是不是分布式
def is_distributed():
return dist.is_available() and dist.is_initialized()
@pysnooper.snoop()
def main():
# Training settings
parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
parser.add_argument('--batch-size', type=int, default=64, metavar='N',
help='input batch size for training (default: 64)')
parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
help='input batch size for testing (default: 1000)')
parser.add_argument('--epochs', type=int, default=1, metavar='N',
help='number of epochs to train (default: 10)')
parser.add_argument('--lr', type=float, default=0.01, metavar='LR',
help='learning rate (default: 0.01)')
parser.add_argument('--momentum', type=float, default=0.5, metavar='M',
help='SGD momentum (default: 0.5)')
parser.add_argument('--no-cuda', action='store_true', default=False,
help='disables CUDA training')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
help='how many batches to wait before logging training status')
parser.add_argument('--save-model', action='store_true', default=False,
help='For Saving the current Model')
parser.add_argument('--dir', default='logs', metavar='L',
help='directory where summary logs are stored')
if dist.is_available():
parser.add_argument('--backend', type=str, help='Distributed backend',
choices=[dist.Backend.GLOO, dist.Backend.NCCL, dist.Backend.MPI],
default=dist.Backend.NCCL)
args = parser.parse_args()
print('reveice args:', args)
# args.no_cuda = True
# args.backend = dist.Backend.GLOO
use_cuda = not args.no_cuda and torch.cuda.is_available()
if use_cuda:
print('Using CUDA')
else:
args.backend = dist.Backend.GLOO
writer = SummaryWriter(args.dir)
torch.manual_seed(args.seed)
device = torch.device("cuda" if use_cuda else "cpu")
if should_distribute():
print('Using distributed PyTorch with {} backend'.format(args.backend))
dist.init_process_group(backend=args.backend)
kwargs = {'num_workers': 1, 'pin_memory': True} if use_cuda else {}
print('bengin load train data %s' % str(datetime.datetime.now()))
train_loader = torch.utils.data.DataLoader(
datasets.MNIST('data', train=True, download=True,
transform=transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])),
batch_size=args.batch_size, shuffle=True, **kwargs)
print('bengin load test data %s' % str(datetime.datetime.now()))
test_loader = torch.utils.data.DataLoader(
datasets.MNIST('data', train=False, transform=transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])),
batch_size=args.test_batch_size, shuffle=False, **kwargs)
print('bengin make net model %s' % str(datetime.datetime.now()))
model = Net().to(device)
if is_distributed():
Distributor = nn.parallel.DistributedDataParallel if use_cuda \
else nn.parallel.DistributedDataParallelCPU
model = Distributor(model)
optimizer = optim.SGD(model.parameters(), lr=args.lr, momentum=args.momentum)
print('bengin train model %s' % str(datetime.datetime.now()))
for epoch in range(1, args.epochs + 1):
train(args, model, device, train_loader, optimizer, epoch, writer)
test(args, model, device, test_loader, writer, epoch)
print('bengin save model %s' % str(datetime.datetime.now()))
if (args.save_model):
torch.save(model.state_dict(), "mnist_cnn.pt")
if __name__ == '__main__':
print('begin python shell %s' % str(datetime.datetime.now()))
main()
print('end python shell %s' % str(datetime.datetime.now()))