Search results

多卡并行时，queue的更新情况

Status: Open.

#70 In tfzhou/ContrastiveSeg;

· DingHuilsp opened on Jan 6, 2025

CKPT

Status: Open.

#69 In tfzhou/ContrastiveSeg;

· geek-APTX4869 opened on Apr 20, 2024

Running the t code shows loss=nan when calculating the coco dataset.

Status: Open.

#68 In tfzhou/ContrastiveSeg;

· liaochuanlin opened on Aug 4, 2023

Dataloader distributed or not

Status: Open.

#67 In tfzhou/ContrastiveSeg;

· zhd2rng opened on Apr 3, 2023

@tfzhou So, the class 0 is the ignore index and not the background class? I mean, for the background class is applied also the contrastive loss. Right?

Status: Open.

#66 In tfzhou/ContrastiveSeg;

· zhangyuxuan321 opened on Feb 27, 2023

Code bug

Status: Open.

#65 In tfzhou/ContrastiveSeg;

· eraserNut opened on Jan 3, 2023

About the dimension of feature embedding

Status: Open.

#64 In tfzhou/ContrastiveSeg;

· chen-wolfberry opened on Nov 28, 2022

loss_contrast与loss_contrast_mem两个版本中的PixelContrastLoss区别

Status: Open.

#63 In tfzhou/ContrastiveSeg;

· IceHowe opened on Nov 26, 2022

problem with resuming training from checkpoint

Status: Open.

#62 In tfzhou/ContrastiveSeg;

· mailtohrishi opened on Nov 16, 2022

Example for max_views, max_samples?

Status: Open.

#61 In tfzhou/ContrastiveSeg;

· lennart-maack opened on Oct 29, 2022

Semi-Hard Example Sampling implementation

Status: Open.

#59 In tfzhou/ContrastiveSeg;

· nysp78 opened on Oct 10, 2022

How to reproduce the SoTA result of Cityscapes?

Status: Open.

#58 In tfzhou/ContrastiveSeg;

· JasiRose opened on Oct 3, 2022