Skip to content

Issues: alibaba/Pai-Megatron-Patch

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

18 Open 192 Closed
18 Open 192 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

H800 flashattention-3不支持deepseek
#473 opened Feb 23, 2025 by leo-ztjht
deepseek-v3模型微调支持fp8训练么？
#471 opened Feb 21, 2025 by renjie130
1
deepseekv2转换误差
#469 opened Feb 20, 2025 by liux750
训练75模型，TP8,PP4，4张A800sft微调训练，总是报错torch.distributed.DistStoreError: Socket Timeout
#468 opened Feb 19, 2025 by Chunfeng1994
2
DeepSeek V3 负载均衡 bias
#467 opened Feb 17, 2025 by yzlnew
对于多模态大模型的训练并行优化支撑
#466 opened Feb 13, 2025 by liyang-7
为什么llama-3.1权重转换时70B需要cpu_options="--use-cpu-initialization"，不会速度过慢吗
#465 opened Feb 13, 2025 by kkkeepgoing
9
求助！Qwen2-VL-7B-Instruct 训练时报错：TypeError: Attention.forward() got an unexpected keyword argument 'attention_bias'
#464 opened Feb 12, 2025 by zhaicunqi
1
新模型支持
#462 opened Feb 11, 2025 by fxnie
请问如何在自己的电脑上配置环境？
#461 opened Feb 10, 2025 by Derek-Kun
请问有Qwen2.5-VL支持的计划吗？
#458 opened Feb 10, 2025 by zhaicunqi
1
qwen2.5不支持moe吗？
#450 opened Feb 6, 2025 by DumoeDss
2
qwen2-vl sft
#447 opened Feb 1, 2025 by whc688
[Question] Reason for excluding fusion in deepseek-v2 training
#446 opened Jan 24, 2025 by wavy-jung
使用json数据格式报下面问题
#445 opened Jan 24, 2025 by QianguoS
训练参数求助
#433 opened Jan 17, 2025 by ltm920716
PAI-Megatron-Patch llama3.1脚本128k上下文预训练速度慢
#410 opened Dec 30, 2024 by kkkeepgoing
optimizer offloading 太强了
#311 opened Aug 2, 2024 by 154912369
2
ProTip! Find all open issues with in progress development work with linked:pr.