Search results

Switching from SGD to AdamW: Significant mIoU boost but concerns about Deformable Offsets stability

Status: Open.

#270 In HuguesTHOMAS/KPConv-PyTorch;

· lxll-0904 opened on Jan 15, 2026

Parameters for training on an S3DIS-like office building dataset

Status: Open.

#269 In HuguesTHOMAS/KPConv-PyTorch;

· Tim-Hoffm opened on Jan 9, 2026

-1 flag when padding

Status: Open.

#267 In HuguesTHOMAS/KPConv-PyTorch;

· TokyoWarfare opened on Oct 17, 2025

KPConv Bias Toward Dominant Classes — How to Break Semantic Collapse?

Status: Open.

#266 In HuguesTHOMAS/KPConv-PyTorch;

· TokyoWarfare opened on Oct 6, 2025

Validation def missing! :D

Status: Open.

#265 In HuguesTHOMAS/KPConv-PyTorch;

· TokyoWarfare opened on Sep 29, 2025

Test model issue

Status: Open.

#264 In HuguesTHOMAS/KPConv-PyTorch;

· ChenxiWang09 opened on Jun 19, 2025

Inquiry Regarding Hyperparameter Settings

Status: Open.

#263 In HuguesTHOMAS/KPConv-PyTorch;

· xxy-nolooker opened on Mar 31, 2025

why it has to use share_memory_

Status: Open.

#262 In HuguesTHOMAS/KPConv-PyTorch;

· liang00fan opened on Jan 23, 2025

Error during my testing phase. I have written the testing program according to my requirement. But i got error in the that.

Status: Open.

#260 In HuguesTHOMAS/KPConv-PyTorch;

· Sayyed-Faizan010189 opened on Dec 11, 2024

About batch_num

Status: Open.

#259 In HuguesTHOMAS/KPConv-PyTorch;

· foxkw opened on Nov 29, 2024

Using intensity as a feature in training results in gradient vanishing

Status: Open.

#258 In HuguesTHOMAS/KPConv-PyTorch;

· RBinSky opened on Nov 26, 2024

export onnx

Status: Open.

#257 In HuguesTHOMAS/KPConv-PyTorch;

· foxkw opened on Nov 26, 2024