Closed BioinfoTongLI closed 1 month ago
Hi @BioinfoTongLI, thanks for bringing this up! i just released a new version to PyPi which allows changing the cache directory where models (and datasets) are saved. You can just set an env var now (SPOTIFLOW_CACHE_DIR
) and, if it exists, it will use that directory instead (see here). That should facilitate the HPC runs :)
Current the pretrained model are saved to the user home folder https://github.com/weigertlab/spotiflow/blob/main/spotiflow/model/pretrained.py#L24-L25
Describe the solution you'd like Would be great to support custume directory to facilitate HPC runs.