search

pytorch / ao

PyTorch native quantization and sparsity for training and inference
BSD 3-Clause "New" or "Revised" License
1.58k stars 173 forks source link

Self compressing neural networks #658

Open msaroufim opened 3 months ago

msaroufim commented 3 months ago

Self-Compressing Neural Networks is dynamic quantization-aware training that puts the size of the model in the loss

Paper: https://arxiv.org/pdf/2301.13142 Code: https://github.com/geohot/ai-notebooks/blob/master/mnist_self_compression.ipynb

aiepm commented 1 week ago

I'm taking this one.