Torch-Pruning
Torch-Pruning copied to clipboard
VainF
[CVPR 2023] Towards Any Structural Pruning; LLMs / SAM / Diffusion / Transformers / YOLOv8 / CNNs
I got same MACs and Params between before and after when purning yolov7-tiny, but yolov7 works well. this is my code: ```python import torch from models.experimental import attempt_load import torch_pruning...
/usr/local/miniconda3/lib/python3.8/site-packages/torch_pruning/dependency.py:667: UserWarning: Unwrapped parameters detected: ['cls_token', 'pos_embed']. Torch-Pruning will prune the last non-singleton dimension of these parameters. If you wish to change this behavior, please provide an unwrapped_parameters argument. warnings.warn(warning_str)...
The SplitPruner updates self.offsets instead of layer.offsets causing index errors when using iterative_steps > 1 Reproducible by setting iterative_steps=2 in the test_concat_split.py and test_split.py tests
i use the code pruning offical yolov5m.pt, but it can't detect anything, What might be the cause of this?
Hi @VainF , I'm playing around with sparse training of `GroupNormPruner` for Yolov8. I read your instruction and it said I have to call `pruner.update_regularizer()` to initialize regularizer. However, I...
I used your test code from the " High-level Pruners" section in the documentation: Instead of using "resnet18" I used resnet50 on an easy classification task. the accuracy without pruning...
Hi there, what's the best way to implement a custom element-wise addition method? I aim to eliminate the inter-layer dependency of the standard addition operator, or at least cease tracking...
hello,Thank you for providing the code，After rewriting the YOLOv5 code based on YOLOv7, it can run normally and the values are also good. The map and p values have not...
