Live Discussion: How to Use Sparsification Recipes to Simplify Pruning & Quantization of PyTorch Models
Sparse models1 are the future of deep learning. They require less footprint and are able to run more efficiently on commodity CPUs2. But popular belief is that achieving sparsity is hard. Using a PyTorch example, we’ll discuss that it doesn’t have to be when using sparsification recipes.
Register below to join us for a two-way discussion. Benjamin Fineran, Neural Magic’s Sr. ML Engineer, will demo how you can apply ready-to-use sparsification recipes to prune and quantize deep learning models. During the session, Benjamin will:
- Discuss the benefits of using recipes to sparsify your deep learning models
- Show how you can easily sparsify PyTorch models within your existing training flows
- Demo speedups on CPUs that result from model sparsification
- Open the floor for a live community discussion.
While recommended but not necessary for this discussion, we encourage you to check out our sparse model zoo, recipes, sparsification tools, and our CPU inference engine on GitHub.
Date: May 19, 2021
Speaker: Benjamin Fineran, Sr. ML Engineer, Neural Magic
1Sparse models = pruned and quantized models
2See our conference-approved research papers to learn about the impact of sparsity on deep learning performance.