[Bug] 安装错误

[Keshawn]

Prerequisite

• I have searched Issues and Discussions but cannot get the expected help.
• I have read the FAQ documentation but cannot get the expected help.
• The bug has not been fixed in the latest version.

Task

I'm using the official example scripts/configs for the officially supported tasks/models/datasets.

Environment

当前环境

NVIDIA Driver Version: 510.47.03
CUDA Version: 11.6
torch 1.13.1+cu116
torchvision 0.14.1+cu116
mmcv-full 1.7.1
openmim 0.3.9
mmgen 0.7.3 /home/keshawn/mmgen/mmgeneration

Reproduces the problem - code sample

from mmgen.apis import init_model, sample_unconditional_model

config_file = 'configs/styleganv2/stylegan2_c2_lsun-church_256_b4x8_800k.py'

you can download this checkpoint in advance and use a local file path.

checkpoint_file = 'https://download.openmmlab.com/mmgen/stylegan2/official_weights/stylegan2-church-config-f-official_20
210327_172657-1d42b7d1.pth'
device = 'cuda:0'

init a generatvie

model = init_model(config_file, checkpoint_file, device=device)

sample images

fake_imgs = sample_unconditional_model(model, 4)

Reproduces the problem - command or script

执行此文件报错：

python 1.py

Reproduces the problem - error message

Traceback (most recent call last):
File "1.py", line 1, in
from mmgen.apis import init_model, sample_unconditional_model
File "/home/keshawn/mmgen/mmgeneration/mmgen/init.py", line 21, in
mmcv_version = digit_version(mmcv.version)
AttributeError: module 'mmcv' has no attribute 'version'

Additional information

如果同时安装 mmcv 和 mmcv-full
也一样会报错

[mm-assistant]

We recommend using English or English & Chinese for issues so that we could have broader discussion.