This is kind of a merge of #605 and #655 to fix #602.
Unlike #655 this one puts the weight manipulation in the training scripts themselves rather than in separate awk scripts.
Note: if this PR is accepted, the correct flow to perform the transition is, stop training, run net_to_model.py with the most recent training output, then resume training.
This is kind of a merge of #605 and #655 to fix #602. Unlike #655 this one puts the weight manipulation in the training scripts themselves rather than in separate awk scripts.
Note: if this PR is accepted, the correct flow to perform the transition is, stop training, run net_to_model.py with the most recent training output, then resume training.
Basic testing still in progress.