Skip to content

Issues: hiyouga/LLaMA-Factory

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

99 Open 3,271 Closed
99 Open 3,271 Closed
Author
Filter by author
Label
Filter by label
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Milestones
Filter by milestone
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

请问推理的时候，instruction与input是按照如下方式拼接的吗？
#3682 opened May 10, 2024 by ironlionliu
1 task done
旧版本和新版本都无法部署web界面的交互
#3680 opened May 10, 2024 by caoyaru123
1 task done
如何实现docker多机多卡 implement multi-machine and multi-gpu training in Docker
#3677 opened May 10, 2024 by magnificent1208
1 task done
请问llamafactory-cli eval支持的评估方式有哪些？ solved This problem has been already solved.
#3676 opened May 10, 2024 by moyaom
2
训练完后没有loss图 wontfix This will not be worked on
#3675 opened May 10, 2024 by Vencin-Z
1 task done
2
auto_gptq 0.7.1，利用工程量化出来的模型不可用，报错：(RayWorkerVllm pid=6035) ERROR 05-10 03:37:55 ray_utils.py:44] ValueError: torch.bfloat16 is not supported for quantization method gptq. Supported dtypes: [torch.float16] [repeated 2x across cluster]
#3674 opened May 10, 2024 by camposs1979
1 task done
TF-TRT Warning: Could not find TensorRT wontfix This will not be worked on
#3673 opened May 10, 2024 by SpikeShaun
1 task done
2
Packing的问题咨询
#3672 opened May 10, 2024 by 447428054
1 task done
全量微调Qwen1.5模型 报错 'weight' must e 2-D
#3671 opened May 10, 2024 by Leekinxun
1 task done
Ascend 910B卡只能跑通单GPU LoRA 微调的（增量）预训练，跑其他任务全部报如下相同的错误，帮忙看下什么原因，感谢
#3670 opened May 10, 2024 by guoyjalihy
1 task done
1
LLam3-8B预测很慢，比lora微调都慢，如果加上adapter_name_or_path就很快
#3669 opened May 10, 2024 by xielinzhen
1 task done
添加--group_by_length后训练一段时间后OOM solved This problem has been already solved.
#3668 opened May 10, 2024 by fst813
1 task done
2
关于react模板thought 和action 写法的问题
#3667 opened May 10, 2024 by lr13953301387
1 task done
自定义的数据集，一直报错，sad.. solved This problem has been already solved.
#3665 opened May 9, 2024 by MissQueen
1
When apply SFT for llama3-8b-base. the results generated duplicated or not relevant information solved This problem has been already solved.
#3664 opened May 9, 2024 by hnhoangdz
1 task done
3
咨询：4 bit量化Qwen72B的模型，需要多大的GPU？我采用4 * A40 （4 * 48GB），量化进度到46%的时候OOM了。
#3663 opened May 9, 2024 by camposs1979
1 task done
1
Mixture-of-Depths Finetune，IndexError: too many indices for tensor of dimension 0
#3662 opened May 9, 2024 by AlexYoung757
1 task done
请问网页版能否断点续训
#3660 opened May 9, 2024 by 153534345
1 task done
0.7.0 版本生成的trainer_log.json不完整
#3658 opened May 9, 2024 by Cucunnber
1 task done
1
How to use accelerate to SFT with multiple machines?
#3656 opened May 9, 2024 by may012345
1 task done
Unable to use ports other than 7860
#3653 opened May 9, 2024 by cocktailpeanut
1 task done
1
请教llama pro扩展后保存的模型文件会变小
#3652 opened May 9, 2024 by LiuChen19960902
glaive_toolcall 数据集Checksum failed solved This problem has been already solved.
#3650 opened May 9, 2024 by cxjtju
1 task done
1
1
部署服务 solved This problem has been already solved.
#3648 opened May 9, 2024 by cxjtju
1 task done
1
您好，如何在日志增加输出Tokens/gpu/s和TFLOPS
#3646 opened May 9, 2024 by Liuweixiong0118
1
ProTip! Find all open issues with in progress development work with linked:pr.