Search results

Reward function lacks constraints on reasoning process, causing model to skip CoT

Status: Open.

#117 In JIA-Lab-research/Seg-Zero;

· feivellau opened on Mar 24, 2026

test repeat

Status: Open.

#115 In JIA-Lab-research/Seg-Zero;

· ddzovo opened on Feb 3, 2026

an error occurs：RuntimeError: aten.copy_.default: got mixed torch.Tensor and DTensor, need to convert all torch.Tensor to DTensor before calling distributed operators!

Status: Open.

#105 In JIA-Lab-research/Seg-Zero;

· ddzovo opened on Oct 13, 2025

Bug in GRPO advantage calculation: Extra brackets in std computation

Status: Open.

#104 In JIA-Lab-research/Seg-Zero;

· SCBoy1007 opened on Oct 4, 2025

Any plans or experiments on multi-image (multi-view) reasoning?

Status: Open.

#100 In JIA-Lab-research/Seg-Zero;

· jsn5 opened on Sep 8, 2025