from mmpose.models.builder import build_pose_estimator
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmpose/models/init.py", line 8, in
from .heads import * # noqa
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmpose/models/heads/init.py", line 11, in
from .transformer_heads import EDPoseHead
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmpose/models/heads/transformer_heads/init.py", line 2, in
from .edpose_head import EDPoseHead
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmpose/models/heads/transformer_heads/edpose_head.py", line 14, in
from mmcv.ops import MultiScaleDeformableAttention
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmcv/ops/init.py", line 3, in
from .active_rotated_filter import active_rotated_filter
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmcv/ops/active_rotated_filter.py", line 10, in
ext_module = ext_loader.load_ext(
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/site-packages/mmcv/utils/ext_loader.py", line 13, in load_ext
ext = importlib.import_module('mmcv.' + name)
File "/opt/nas/p/conda/envs/MuseTalk/lib/python3.10/importlib/init.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
ModuleNotFoundError: No module named 'mmcv._ext'
If you are running on mac, there is a repo which showcases how to run this.
When I tried it was instanely slow though on macos, since this repo doesn't support MPS too well. Most of this repo requires CUDA.
https://github.com/rrfunde/MuseTalk/blob/61ef17f0fc5b522ef660ac08a9b81de374b630e5/MACOS_COMPATIBILITY_NOTE.md?plain=1#L205
I faced the same error you have above on a GPU machine with CUDA and I was able to solve it by downgrading the mmcv package and used this version instead of 2.1.0:
mim install "mmcv==2.0.1"
I solved it by building from source
mim install "mmcv==2.0.1"
not work
I solved it by building from source
not work
TMElyralab
