timmeinhardt / trackformer

Implementation of "TrackFormer: Multi-Object Tracking with Transformers”. [Conference on Computer Vision and Pattern Recognition (CVPR), 2022]
https://arxiv.org/abs/2101.02702
Apache License 2.0
487 stars 113 forks source link

ImportError: libtorch_cuda_cu.so: cannot open shared object file: No such file or directory #89

Closed hustmx721 closed 1 year ago

hustmx721 commented 1 year ago

I follow the instruction from the closed issue #41 my cuda version is 10.2,and pytorch version is 1.7.1,torchersion version is 0.8.1;other third-party dependency libraries are installed based on the instructions provided by install.md. However,whether my pytorch version is 1.5.0 or another (torchvision is the equivalent), I get the same error when I run the demo code: $ python src/track.py with \ dataset_name=DEMO \ data_root_dir=data/snakeboard \ output_dir=data/snakeboard \ write_images=pretty Traceback (most recent call last): File "src/track.py", line 16, in from trackformer.models import build_model File "/home/mist/trackformer/src/trackformer/models/init.py", line 6, in from .deformable_transformer import build_deforamble_transformer File "/home/mist/trackformer/src/trackformer/models/deformable_transformer.py", line 17, in from .ops.modules import MSDeformAttn File "/home/mist/trackformer/src/trackformer/models/ops/modules/init.py", line 1, in from .ms_deform_attn import MSDeformAttn File "/home/mist/trackformer/src/trackformer/models/ops/modules/ms_deform_attn.py", line 11, in from ..functions import MSDeformAttnFunction, ms_deform_attn_core_pytorch File "/home/mist/trackformer/src/trackformer/models/ops/functions/init.py", line 1, in from .ms_deform_attn_func import MSDeformAttnFunction, ms_deform_attn_core_pytorch, ms_deform_attn_core_pytorch_mot File "/home/mist/trackformer/src/trackformer/models/ops/functions/ms_deform_attn_func.py", line 11, in import MultiScaleDeformableAttention as MSDA ImportError: libtorch_cuda_cu.so: cannot open shared object file: No such file or directory

hustmx721 commented 1 year ago

I successfully solved this problem by upgrading my CUDA and pytorch versions. Specifically, cuda version is 11.7 and pytorch version is 1.13.1. pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu117

timmeinhardt commented 1 year ago

Glad you solved the issue. Can we close it?