Discussion on the role of some classes

[saman-aghazadeh]

Hi,

I'm trying to understand the structure of the core pruning classes. I wanted to start this thread, and see if you can provide some explanation on the roles of the below classes, and how they generally interact with the main net class:

• Allocator
• Pruner
• Sparsifier
• Tracker

Any other extra information is also much appreciated.

[lucaslie]

Hi,

thanks for reaching out. here is an overview and please feel free to ask me any follow-up questions if you want to go into details about any of those:

The BaseCompressedNet takes care of the overall compression and provides the high-level API for compressing a network.

The network compression itself is split up into multiple parts:

1. Allocator: The allocator takes a overall desired prune ratio for the network and allocates a per-layer prune ratio. So in the simplest case the allocator could just be a constant per-layer prune ratio, for example.

2. Pruner: The pruner takes as input the per-layer prune ratio and determines how many weights should be pruned from each filter or neuron in the layer. So it's almost like a allocator within a layer.

3. Sparsifier: this class actually implements the sparsification. So as input it takes the per-neuron/filter sparsity determined by the pruner and sparsifies the weight tensor.

4. Tracker: is a convenient wrapper for pytorch's forward/backward hook functionality to track layer statistics that might be required during pruning.

[saman-aghazadeh]

Hi,

Thanks for the explanation. I was going through the code with the info you provided above and it helped me get a better understanding of how things works.

I just need a bit more explanation on how base classes for allocator and pruner work. I have hard time to understand the base logic behind these two classes, just by reading the code. Any help will be appreciated.

[saman-aghazadeh]

Also another question. I'm interested to play around with the structured sparsification of the model. Does the code remove the sparsed filters from the model? I would like to be able to get real performance boost on the hardware, after performing the sparsification.

Thanks,

[lucaslie]

This is currently not supported and/or I don't have plans to support it. If you are interested in inference time speed-ups, then I would recommend writing some type of post-processing tool that takes the network graph as input and outputs the slimmed model.

Alternatively, you can also hardcode the slimmed version for a particular architecture.

If you want to implement something like this, I would be happy to take PRs as well.