Description
There is an inconsistency between the PyTorch version specified in the installation instructions in the README and the version listed in the requirements.txt file. This can cause confusion for users trying to set up the environment.
Steps to Reproduce
In the README, the installation command specifies:
pip install torch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 --index-url https://download.pytorch.org/whl/cu118
However, in requirements.txt, the same packages are listed without the CUDA-specific URL:
torch==2.1.1
torchvision==0.16.1
...
Proposed Solution
- Make the requirements consistent between the README and requirements.txt
- Consider using a more flexible approach for requirements specification:
- Use a
base_requirements.txt for common dependencies
- Provide separate requirement files for different CUDA versions
- Update the README to guide users on choosing the appropriate requirements file
- Add comments in the requirements.txt file to clarify CUDA compatibility
Description
There is an inconsistency between the PyTorch version specified in the installation instructions in the README and the version listed in the requirements.txt file. This can cause confusion for users trying to set up the environment.
Steps to Reproduce
In the README, the installation command specifies:
However, in requirements.txt, the same packages are listed without the CUDA-specific URL:
Proposed Solution
base_requirements.txtfor common dependencies