Fixed BN Extraction

Fixed a bug in the previous version where batch norm layers were being sparsified... this is not good.

Patch

Fixed print in last release.

More General get_W

v0.5

Update setup.py

Added Checkpoints

Added the ability to checkpoint the pruner object without reinitializing random topologies.

Added Accumulated Gradient Scoring

Added the ability to accumulate gradients to simulate larger batch sizes for RigL steps. Larger batch sizes tend to reduce noise & the aim is to emulate a batch size of 4096 using a true batch size of 1024.

Updated Description / Package Details

v0.2

Update setup.py

RigL + Distributed

First release that implements RigL to spec with distributed training available.