issues
search
mle-infrastructure
/
mle-toolbox
Lightweight Tool to Manage Distributed ML Experiments ðŸ›
https://mle-infrastructure.github.io/mle_toolbox/toolbox/
MIT License
3
stars
1
forks
source link
Many small enhancements - Logging, GCS zip, conda/venv
#29
Closed
RobertTLange
closed
3 years ago
RobertTLange
commented
3 years ago
Introduce mle-config variable
use_conda_virtual_env
to allow user to choose between
conda
and
venv
virtual environment setup. Addresses #3.
Introduce mle-config variable
remote_env_name
to allow user to set name for the remote experiment execution environment. Addresses #1.
Fix GCS retrieval unzipping so that there are no redundant sub-directories. Addresses #5.
Make logging even better (different model types and top-k checkpointing). Addresses #12.
Add sklearn SVM classifier + Haiku VAE example with .pkl model saving.
use_conda_virtual_envto allow user to choose betweencondaandvenvvirtual environment setup. Addresses #3.remote_env_nameto allow user to set name for the remote experiment execution environment. Addresses #1.