quantization

• Create a new examples directory at the top-level
• Use the API doc examples to apply Channel Pruning model compression to a Keras resnet18 model

If you are interested in working on this issue - please indicate via a comment on this issue. It should be possible for us to pair you up with an existing contributor to help you get started.

From a complexity perspective, this ticket is at an ea

FINN has a Vivado version requirements, e.g. 2019.1 in the 0.2b release. The available Vivado version should be checked before any Vivado-related commands are launched, and an assertion should be raised if there is a version mismatch.

The idea is to have a more advanced Filter Pruning method to be able to show SOTA results in model compression/optimization.

I suggest reimplementing the method from here: https://github.com/cmu-enyac/LeGR and reproduce baseline results for MobileNet v2 on CIFAR100 as the first step.

cc'ed @vshampor, @vanyalzr.

Blueoil inference doesn't work with python2 any longer.
We need to update the pip packages for inference but enum34 package would be the block for some recent packages.
Now the CI has pip uninstall enum34 in the pipeline for the test backward compatibility, but this is no longer required for future.

Related to
blue-oil/blueoil#1139

We also need to benchmark the Lottery-tickets Pruning algorithm and the Quantization algorithms. The models used for this would be the student networks discussed in #105 (ResNet18, MobileNet v2, Quantization v2).

Pruning (benchmark upto 40, 50 and 60 % pruned weights)

• Lottery Tickets

Quantization

• Static
• QAT