cannot import name 'c_lib' from partially initialized module 'byteps.torch'

[njqconquer]

Describe the bug
Traceback (most recent call last):
File "/media/sdb1/niejinquan/compression-code/deep-gradient-compression/test.py", line 5, in
import byteps.torch as bps
File "/home/niejinquan/.conda/envs/GC/lib/python3.9/site-packages/byteps/torch/init.py", line 24, in
from byteps.torch.ops import push_pull_async_inplace as byteps_push_pull
File "/home/niejinquan/.conda/envs/GC/lib/python3.9/site-packages/byteps/torch/ops.py", line 29, in
from byteps.torch import c_lib
ImportError: cannot import name 'c_lib' from partially initialized module 'byteps.torch' (most likely due to a circular import) (/home/niejinquan/.conda/envs/GC/lib/python3.9/site-packages/byteps/torch/init.py)

To Reproduce
Steps to reproduce the behavior:

1. pip install byteps

2. run the test.py

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
import byteps.torch as bps

初始化

bps.init()
torch.manual_seed(42)

定义模型

class Net(nn.Module):
def init(self):
super(Net, self).init()
self.fc1 = nn.Linear(784, 128)
self.fc2 = nn.Linear(128, 10)

def forward(self, x):
    x = torch.relu(self.fc1(x))
    x = self.fc2(x)
    return x

数据准备

transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])

train_dataset = datasets.MNIST('data', train=True, download=True,
transform=transform)
train_loader = torch.utils.data.DataLoader(
train_dataset, batch_size=64, shuffle=True)

模型和优化器

model = Net()
optimizer = optim.SGD(model.parameters(), lr=0.01 * bps.size())
optimizer = bps.DistributedOptimizer(optimizer)

广播参数

bps.broadcast_parameters(model.state_dict(), root_rank=0)

训练循环

def train(epoch):
model.train()
for batch_idx, (data, target) in enumerate(train_loader):
optimizer.zero_grad()
output = model(data.view(-1, 784))
loss = nn.functional.cross_entropy(output, target)
loss.backward()
optimizer.step()

for epoch in range(1, 11):
train(epoch)

3. See error

Expected behavior
A clear and concise description of what you expected to happen.

Screenshots
If applicable, add screenshots to help explain your problem.

Environment (please complete the following information):

• OS: Ubuntu 18.04
• GCC version: 9.4
• CUDA and NCCL version: 11.4
• Framework (TF, PyTorch, MXNet): PyTorch

Additional context
Add any other context about the problem here.