emilianogagliardiemanueleghelfi / cnn-compression-performance

A python script that automatise the training of a CNN, compress it through tensorflow (or ristretto) plugin, and compares the performance of the two networks

Introduction

For real world application, convolutional neural network(CNN) model can take more than 100MB of space and can be computationally too expensive. Therefore, there are multiple methods to reduce this complexity in the state of art. Ristretto is a plug-in to Caffe framework that employs several model approximation methods. For this projects, first a CNN model will be trained for Cifar-10 dataset with Caffe, then Ristretto will be use to generate multiple approximated version of the trained model using different schemes. The goal of this projects is comparison of the models in terms of execution performance, model size and cache utilizations in the test or inference phase. The same steps are done with Tensorflow and Quantisation tool. The quantisation schemes of Tensorflow and Ristretto are then compared.

Folder description

• ristretto: contains the tools for caffe ristretto quantization
• tf_quantize: contains the tools for tensorflow quantization

A more detailed description is inside the readme of each folder.

Results

Authors

• Emanuele Ghelfi
• Emiliano Gagliardi

cnn-compression-performance's Issues

Testing model

Hello,
I am trying to test this model and facing issue while running workflow.py. it giving me error:
user_module = importlib.import_module(args.module_name[0])
TypeError: 'NoneType' object is not subscriptable

Can you please tell me what is missing?
Also, for mnist_models file 2conv_2fc.py is having below variables:
checkpoint_prefix = 'CNNs/mnist_models/net_serializations/2conv_2fc/net'
checkpoint_path = 'CNNs/mnist_models/net_serializations/2conv_2fc'
I am not able to get this path.

How to save the model after quantized?

I implement linear quantization by PyTorch,but after quantized, model's size unchanged,it still save in float format, how to save the quantized model? Thanks for your help.

a bytes-like object is required, not 'str'

Hello,
While Testing I am getting this error:

Traceback (most recent call last):
File "workflow.py", line 294, in
main(model_instance, strtobool(args.train[0]), strtobool(args.quantize[0]), strtobool(args.evaluate[0]))
File "workflow.py", line 252, in main
o_current_dict = perf_stderr2dict(o_perf_std_err)
File "workflow.py", line 186, in perf_stderr2dict
perf_output_list = perf_output.split("\n")
TypeError: a bytes-like object is required, not 'str'