Irreproducible conda environment from yml file
Opened this issue · 1 comments
As is, running the given yml file on conda 23.11.0 in my Linux system installed a cpu-only pytorch. Here's a fixed yml file that specifies a GPU-compatible Pytorch:
name: RF2NA
channels:
- pytorch
- nvidia
- defaults
- conda-forge
dependencies:
- python=3.10
- pip
- pytorch::pytorch=2.0.1=py3.10_cuda11.7_cudnn8.5.0_0
- requests
- pytorch-cuda=11.7
- dglteam/label/cu117::dgl
- pyg::pyg
- bioconda::mafft
- bioconda::hhsuite
- bioconda::blast
- bioconda::hmmer>=3.3
- bioconda::infernal
- bioconda::cd-hit
- bioconda::csblast
- pip:
- psutil
- tqdm
and output from conda env export
for a reproducible yml file is attached. Issue is similar to #36 , but here we want to run on GPU and not CPU. Additionally, you may want to add aria2 to the dependency list, as aria2c is recommended for downloading large dbs like bfd.
env.txt
As is, running the given yml file on conda 23.11.0 in my Linux system installed a cpu-only pytorch. Here's a fixed yml file that specifies a GPU-compatible Pytorch:
name: RF2NA channels: - pytorch - nvidia - defaults - conda-forge dependencies: - python=3.10 - pip - pytorch::pytorch=2.0.1=py3.10_cuda11.7_cudnn8.5.0_0 - requests - pytorch-cuda=11.7 - dglteam/label/cu117::dgl - pyg::pyg - bioconda::mafft - bioconda::hhsuite - bioconda::blast - bioconda::hmmer>=3.3 - bioconda::infernal - bioconda::cd-hit - bioconda::csblast - pip: - psutil - tqdm
and output from
conda env export
for a reproducible yml file is attached. Issue is similar to #36 , but here we want to run on GPU and not CPU. Additionally, you may want to add aria2 to the dependency list, as aria2c is recommended for downloading large dbs like bfd. env.txt
This was the only solution that worked for me when trying to get the code to run on GPU. Note, I used the version of this repo as of the commit of Dec 26, 2023. For generating the conda env, I used the provided env.txt file.