Awesome

Comparing Rewinding and Fine-tuning in Neural Network Pruning

This is the code and data for the paper Comparing Rewinding and Fine-tuning in Neural Network Pruning.

@inproceedings{renda2020comparing,
    title={Comparing Rewinding and Fine-tuning in Neural Network Pruning},
    author={Renda, Alex and Frankle, Jonathan and Carbin, Michael},
    booktitle={International Conference on Learning Representations},
    year={2020},
    url={https://openreview.net/forum?id=S1gSj0NKvB}
}

For comparing against the data included in the paper, data and plotting code are located in vision/plot/.

For replicating results and using the trained networks released with the paper, vision networks are located in the vision subdirectory, and the GNMT network is located in the gnmt subdirectory.

Some example networks (original initializations, trained, and pruned and re-trained) can be found here.