Skip to content

Issues: hiyouga/LLaMA-Factory

🚨FAQs | 常见问题🚨
#4614 opened Jun 28, 2024 by hiyouga
Open
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

267 Open 5,115 Closed
267 Open 5,115 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

Translation dataset pending This problem is yet to be addressed
#6205 opened Dec 1, 2024 by nichellehouston
Help to get the generated output: The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's attention_mask to obtain reliable results. Setting pad_token_id to eos_token_id:None for open-end generation. pending This problem is yet to be addressed
#6203 opened Dec 1, 2024 by Wwwendyy
1 task done
🐛 Bug：不支持A800 NVLINK pending This problem is yet to be addressed
#6201 opened Dec 1, 2024 by Evi233
1 task done
在双卡上进行GPTQ量化时出现异常 pending This problem is yet to be addressed
#6200 opened Nov 30, 2024 by czhcc
1 task done
Question on the relation between cutoff_len and max_position_embeddings pending This problem is yet to be addressed
#6199 opened Nov 30, 2024 by mces89
1 task done
我有个类似问题。假定基础模型Model 1.我在基础模型Model 的基础上用数据集A微调了LoRA权重LoRA_A 2.我想继续在LoRA权重A的基础上，用数据集B微调LoRA权重LoRA_B 我不希望将LoRA_A存成一个基础模型，我希望直接从LoRA_A的权重微调LoRA_B，获得的B_adapter直接能用在原始模型Model的权重上，我该如何操作？ pending This problem is yet to be addressed
#6198 opened Nov 30, 2024 by DataNinja42298
1 task done
no_sync context manager is incompatible with gradient partitioning logic of ZeRO stage 3 pending This problem is yet to be addressed
#6194 opened Nov 29, 2024 by ioir123ju
1 task done
2
Problems using VRAM when learning models pending This problem is yet to be addressed
#6187 opened Nov 29, 2024 by rhqudco
1 task done
1
使用Badam多卡训练中途弹出Invalidate trace cache @ step 2: expected module 1, but got module 332并停止训练 pending This problem is yet to be addressed
#6184 opened Nov 28, 2024 by 66RomanReigns
1 task done
8
Increasing the number of epochs results in an OOM (Out of Memory) error. pending This problem is yet to be addressed
#6183 opened Nov 28, 2024 by cola-1210
1 task done
HQQ量化报错 pending This problem is yet to be addressed
#6176 opened Nov 28, 2024 by xinyubai1209
1 task done
如何得到每条数据的 loss pending This problem is yet to be addressed
#6165 opened Nov 27, 2024 by Word2VecT
1 task done
dora微调后模型进行dpo，模型输出复读机 pending This problem is yet to be addressed
#6164 opened Nov 27, 2024 by immusferr
1 task done
qwen2.5-coder-33B 相比 qwen2.5-33B-instruct 训练需要消耗更多的显存么？ pending This problem is yet to be addressed
#6163 opened Nov 27, 2024 by yangyang6666
1 task done
Qwen2-VL-7B 图文微调数据与纯文本微调数据混合训练,loss跌为0 pending This problem is yet to be addressed
#6159 opened Nov 27, 2024 by VincentVanNF
1 task done
两台机器全参数微调Qwen2.5-14B-Instruct挂起不动 pending This problem is yet to be addressed
#6143 opened Nov 26, 2024 by zhaoxjmail
1 task done
2
qwen2_vl 用 lora 训练慢，有什么工具可以排查慢的原因，有 torch.profiler 可以用吗？ pending This problem is yet to be addressed
#6134 opened Nov 25, 2024 by enerai
1 task done
pt模式下，关于样本数量、训练步数的不一致的现象 pending This problem is yet to be addressed
#6133 opened Nov 25, 2024 by kascas
1 task done
3
httpx.ConnectError: [Errno 111] Connection refused pending This problem is yet to be addressed
#6119 opened Nov 23, 2024 by cuiMidAutumn
1 task done
再次求助。。wandb断点延伸曲线 pending This problem is yet to be addressed
#6110 opened Nov 22, 2024 by Saberlve
1 task done
求助：模型qwen2.5-7b-instruct全量sft的时候，训练过程中loss突然变为0。 pending This problem is yet to be addressed
#6109 opened Nov 22, 2024 by Chtholly1
1 task done
2
使用API调用奖励模型得到的奖励值与PPO过程中的奖励值差距巨大/The disparity between the reward values obtained from calling the reward model using the API and the reward values from the PPO process is huge pending This problem is yet to be addressed
#6100 opened Nov 21, 2024 by LuRenjias
1 task done
mllm 数据格式如果存在问题，如何设置忽略该样本 pending This problem is yet to be addressed
#6096 opened Nov 21, 2024 by DietDietDiet
1 task done
LoRA微调Qwen2-VL-2B时，Loss一直为0，grad_norm为nan pending This problem is yet to be addressed
#6092 opened Nov 20, 2024 by Tian-ye1214
1 task done
1
BAdam算法finetune的迭代轮次和论文不是很符合 pending This problem is yet to be addressed
#6088 opened Nov 20, 2024 by PhzCode
1 task done
4
ProTip! no:milestone will show everything without a milestone.