Open marued opened 4 years ago
Duplicate with #6
YAML is still in process during the projet. This is a on-going process that will be used for TensorFlow1 model to modularize the configuration including checkpoints and experiences setting.
If we use another FW with YAML, but we will still have these configuration files to makes team work more efficient.
If we want to run on different datasets, we would like to be able to use the same training loop on different datasets that provide the information in the same way so it becomes simply an option to run one or the other dataset in the end.
Either expose the option in the YAML or provide it as a parameter input to the python script.