Self-pruning MLP-Mixer on CIFAR-10 every Linear is a custom PrunableLinear with a learnable sigmoid gate, trained with CE + λ·Σσ(g). 83.86 % accuracy, up to 128× compression, verified by hard-pruning every sub-threshold weight. Submitted for the Tredence AI Engineering Internship 2026 case study.
sparsity deep-learning pytorch case-study model-compression cifar-10 network-pruning l1-regularization mlp-mixer self-pruning learnable-gates tredence
-
Updated
Apr 19, 2026 - Jupyter Notebook