All Tools
M
Fine-tuningFreeOpen Source
MMRAZOR
OpenMMLab toolbox for pruning, distillation, and NAS
Apache-2.0
ABOUT
Compressing OpenMMLab detectors and classifiers usually means a separate pruning or distillation repo. MMRazor plugs pruning, KD, and NAS into MMEngine so teams slim models for edge deployment without rewriting the training loop.
INSTALL
pip install mmrazorINTEGRATION GUIDE
1. Prune MMDetection models for faster edge inference
2. Distill a large classifier into a smaller student with MMRazor recipes
3. Search compact architectures with NAS on OpenMMLab backbones
4. Compress segmentation models before mobile or embedded deploy
TAGS
pythonpytorchpruningdistillationnascompressionopenmmlabfine-tuning