Search results

sinkhorn的计算问题

Status: Open.

#59 In AkaliKong/MiniOneRec;

· xym0601 opened on Feb 4, 2026

强化学习（RL）后模型性能下降

Status: Open.

#58 In AkaliKong/MiniOneRec;

· cywuuuu opened on Feb 1, 2026

测试的时候非常多UserWarning: No valid tokens found for hash_key

Status: Open.

#56 In AkaliKong/MiniOneRec;

· SuperTom99 opened on Jan 29, 2026

Fluctuations in Reinforcement Learning Rewards

Status: Open.

#55 In AkaliKong/MiniOneRec;

· fukioston opened on Jan 26, 2026

贴一个用BPE分词的BUG

Status: Open.

#54 In AkaliKong/MiniOneRec;

· lyzltysgithub opened on Jan 21, 2026

Which model is the priority choice for reproduction: Base or Instruct?

Status: Open.

#52 In AkaliKong/MiniOneRec;

· cywuuuu opened on Jan 14, 2026

您好，为什么不用训练好的 codebook hidden 做LLM新增token维度的初始化呢

Status: Open.

#51 In AkaliKong/MiniOneRec;

· yangyan1227 opened on Jan 8, 2026

关于文章结果复现的问题

Status: Open.

#49 In AkaliKong/MiniOneRec;

· rijinjjjj opened on Jan 3, 2026

Is Qwen2.5-1.5B-Instruction Required to Reproduce Results?

Status: Open.

#45 In AkaliKong/MiniOneRec;

· AQingRoad opened on Dec 10, 2025

Reproducing -- test performance degrades after RL step

Status: Open.

#44 In AkaliKong/MiniOneRec;

· ashleys0 opened on Dec 9, 2025

Problems Encountered in Result Reproduction

Status: Open.

#43 In AkaliKong/MiniOneRec;

· newwindhxx opened on Dec 9, 2025

rqkmeans 不支持 码本三层不同的num_level吗？

Status: Open.

#42 In AkaliKong/MiniOneRec;

· luyao-cv opened on Dec 8, 2025