Skip to content

Issues: open-mmlab/mmrazor

New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

#

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

141 Open 133 Closed
141 Open 133 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Recorder for gradients of certian modules
#653 opened Sep 23, 2024 by lanqz7766
Problematic difference between FakeQuantize for activations and weights
#651 opened Aug 29, 2024 by Veccoy
1
[Docs] A100算力加持！书生大模型实战营第3期全面升级，趣味闯关模式等你开启
#649 opened Jul 16, 2024 by boshallen
MAP is stucked at 0 for Mobilenet V2 SSD QAT without pretrained model [Bug] bug Something isn't working
#648 opened Jul 11, 2024 by HeleMartin
Is this a dead project ?
#643 opened Apr 25, 2024 by taofuyu
1
[Bug] (suggested fix) mmrazor.models.algorithms.mm_architecture.MMArchitectureQuant.get_deploy_model() fails if predict mode lacks nodes from the model.quantizer.tracer.skipped_methods configuration, but the architecture quantizer.prepare(fp32_model) has these nodes. bug Something isn't working
#642 opened Apr 23, 2024 by elisa-aleman
4
Regarding tables and accuracy
#641 opened Apr 15, 2024 by PriyanshuPandeyNA
When I use methodoutputs to access the results of assigner, I only obtain one sample
#639 opened Apr 12, 2024 by pkyzh2006
[Bug] Custom Distillation MMSeg CWD loss nan problem bug Something isn't working
#638 opened Apr 12, 2024 by Priyanshu88
1
[Bug] mmrazor.engine.runner.quantization_loops.QATValLoop calls after_val_epoch hook twice with different keys, causing mmengine.hooks.checkpoint_hook._save_best_checkpoint() to fail with KeyError for the save_best config bug Something isn't working
#637 opened Apr 10, 2024 by elisa-aleman
[Bug] (suggested fix) mmrazor.models.algorithms.quantization.mm_architecture.MMArchitectureQuant.sync_qparams() fails if there are modules present in other modes but not in forward mode='tensor' bug Something isn't working
#634 opened Apr 5, 2024 by elisa-aleman
4
[Bug] (suggested fix) nn.Parameter are not added to root after being traced in mmrazor.models.task_utils.tracer.fx.custom_tracer.build_graphmodule() bug Something isn't working
#633 opened Apr 2, 2024 by elisa-aleman
2
[Bug] (suggested temporary fix) Pytorch >= 2 causes mmrazor.engine to fail bug Something isn't working
#632 opened Mar 26, 2024 by elisa-aleman
4
Try to reproduce CWD in VOC data set
#631 opened Mar 20, 2024 by nameLiMing
[Bug] TypeError: 'NoneType' object is not iterable bug Something isn't working
#630 opened Mar 13, 2024 by Priyanshu88
How to get started??
#629 opened Mar 8, 2024 by milinbhade1214
I can't reproduce dfad results
#628 opened Feb 27, 2024 by c122-ode
Was support for GN ever added to mmrazor?
#627 opened Feb 18, 2024 by qraleq
Error: assert len(module.recorded_input) > 0 bug Something isn't working
#626 opened Feb 10, 2024 by yourbikun
How to use the mmrazor tool to lightweight the end2end.onnx model generated by mmdeploy
#625 opened Jan 28, 2024 by xqc-qc
[Bug] use mmrazor/main/tools/visualizations/feature_diff_visualization.py error: inference_mot() missing 2 required positional arguments: 'frame_id' and 'video_len' bug Something isn't working
#624 opened Jan 21, 2024 by zhanghay
4
generate a yolov5_prune_config file
#623 opened Jan 19, 2024 by mayilin0714
torch.fx.proxy.TraceError: class MMArchitectureQuant bug Something isn't working
#621 opened Jan 5, 2024 by Levi-zhan
20
KD connector
#620 opened Dec 25, 2023 by FPPMXB
The student model maintains an mAP of 0 throughout the distiller training process.
#619 opened Dec 22, 2023 by Rax-Lie
ProTip! Follow long discussions with comments:>50.