Issues - deepseek-ai/DeepSeek-MoE - Forgejo: Beyond coding. We Forge.

deepseek-ai/DeepSeek-MoE

mirror of https://github.com/deepseek-ai/DeepSeek-MoE.git synced 2026-11-05 00:57:43 +00:00

Search syntax

`assignee:<username>`	Filter by assigned user
`author:<username>`	Filter by the author
`is:all`	All
`is:closed`	Closed
`is:open`	Open
`mentions:<username>`	Filter by mentioned user
`modified:[>\|<]<date>`	Filter by last modified date
`review:<username>`	Filter by user who reviewed
`sort:<by>:[asc\|desc]`	Sort by: `created\|comments\|updated\|deadline`

New issue

[PR #49] Ensure transformers version is lower than v4.49.0

#52 opened 2026-10-26 21:02:09 +00:00 by bsakaguchi

[PR #54] fix: use the tokenizer's eos_token as the fine-tuning target terminator

#51 opened 2026-10-26 21:02:09 +00:00 by bsakaguchi

[PR #50] fix: correct logger.info() syntax in finetune.py

#50 opened 2026-10-26 21:02:09 +00:00 by bsakaguchi 1 / 2

[PR #40] Finetune patch

#49 opened 2026-10-26 21:02:07 +00:00 by bsakaguchi

[PR #45] Fix HuggingFace path for text completion quickstart

#48 opened 2026-10-26 21:02:07 +00:00 by bsakaguchi

[PR #44] Fix typo in README.md

#47 opened 2026-10-26 21:02:06 +00:00 by bsakaguchi

[GH-ISSUE #48] 数据格式怎么确定

#44 opened 2026-10-26 21:02:02 +00:00 by bsakaguchi

[GH-ISSUE #53] Router-swappable experts: has profile-based routing been considered for DeepSeekMoE?

#43 opened 2026-10-26 21:02:02 +00:00 by bsakaguchi

[GH-ISSUE #46] zero3单机多卡前向传播会有问题

#41 opened 2026-10-26 21:02:00 +00:00 by bsakaguchi

[GH-ISSUE #42] Ablation studies for DeepSeekMoE

#39 opened 2026-10-26 21:01:58 +00:00 by bsakaguchi

[GH-ISSUE #41] 用来训练deepseek-v2-coder-lite-instruction会挂起

#38 opened 2026-10-26 21:01:58 +00:00 by bsakaguchi

[GH-ISSUE #39] 训练MoE的时候会出现loss = 0 的情况

#37 opened 2026-10-26 21:01:58 +00:00 by bsakaguchi

[GH-ISSUE #38] 专家并行是怎么配置的？有配置代码吗

#36 opened 2026-10-26 21:01:56 +00:00 by bsakaguchi

[GH-ISSUE #37] why <|EOT|> ?

#34 opened 2026-10-26 21:01:56 +00:00 by bsakaguchi

[GH-ISSUE #35] Finetune with deepspeed: type mismatch

#33 opened 2026-10-26 21:01:54 +00:00 by bsakaguchi

[GH-ISSUE #34] 单卡A100-80G推理速度慢

#31 opened 2026-10-26 21:01:54 +00:00 by bsakaguchi

[GH-ISSUE #30] 模型结果测评复现

#30 opened 2026-10-26 21:01:52 +00:00 by bsakaguchi

[GH-ISSUE #32] 能添加modelscope链接吗，这样可以更方便一些不能连hg的情况

#29 opened 2026-10-26 21:01:52 +00:00 by bsakaguchi

[GH-ISSUE #31] MOE 并行怎么实现的？

#28 opened 2026-10-26 21:01:52 +00:00 by bsakaguchi

[GH-ISSUE #27] deepseek-moe模型在进行lora微调训练时loss值会突然变为0一直到最后，导致推理异常。

#25 opened 2026-10-26 21:01:49 +00:00 by bsakaguchi

1 2