Skip to content

Issues: hiyouga/LLaMA-Factory

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

475 Open 6,357 Closed
475 Open 6,357 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
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Gemma3 finetune后model.safetensors.index.json和原来不一样，导致无法使用vllm推理 bug Something isn't working pending This problem is yet to be addressed
#8243 opened May 31, 2025 by junleiz
1 task done
4
使用VLLM推理internvl3-8b-hf报错:ValueError: Attempted to assign 2816 = 2816 multimodal tokens to 2817 placeholders bug Something isn't working pending This problem is yet to be addressed
#8238 opened May 30, 2025 by junleiz
1 task done
@Kuangdd01
2
使用vllm部署internvl3-8b-hf，在转换后报错ValueError: There is no module or parameter named 'model' in InternVLChatModel bug Something isn't working pending This problem is yet to be addressed
#8237 opened May 30, 2025 by junleiz
1 task done
@Kuangdd01
QLora Finetune quantized GGUF model? enhancement New feature or request pending This problem is yet to be addressed
#8229 opened May 30, 2025 by kyang-06
1 task done
3
ValueError: Number of images does not match number of special image tokens in the input text. Got 0 image tokens in the text but 256 tokens from image embeddings. bug Something isn't working pending This problem is yet to be addressed
#8226 opened May 30, 2025 by Dod-o
1 task done
3
How to generate multi results with beamsearch size? enhancement New feature or request pending This problem is yet to be addressed
#8218 opened May 29, 2025 by LinguaLogician
1 task done
after lora fintuing qwen2.5_7b_omni Qwen2_5OmniThinkerConfig' object has no attribute 'vision_start_token_id bug Something isn't working pending This problem is yet to be addressed
#8214 opened May 29, 2025 by Qiny-dl
1 task done
@Kuangdd01
1
Confusion About Data Shuffling for Pretraining bug Something isn't working pending This problem is yet to be addressed
#8213 opened May 29, 2025 by LinaZAlyahya
1 task done
2
在已有模型基础上加新的层，并重新定义loss function进行训练 bug Something isn't working pending This problem is yet to be addressed
#8208 opened May 29, 2025 by Felixvillas
1 task done
1
error using deepspeed AutoTP training qwen3 moe bug Something isn't working pending This problem is yet to be addressed
#8206 opened May 29, 2025 by neiblegy
1 task done
1
训练进度条不更新/卡住（Qwen2.5-omni）：Training-Bug under specific bs /accumulate_steps/ num_examples bug Something isn't working pending This problem is yet to be addressed
#8204 opened May 29, 2025 by Eureka-Maggie
1 task done
@Kuangdd01
1
cuda error bug Something isn't working pending This problem is yet to be addressed
#8200 opened May 28, 2025 by SISTMrL
1 task done
lora的rm继续训练报错，加载checkpoint出现error bug Something isn't working pending This problem is yet to be addressed
#8185 opened May 28, 2025 by WRAllen
1 task done
1
当前最新版本代码使用内置数据集c4_demo，在华为NPU上对qwen3-0.6b做增量预训练报错argument of type "NoneType" is not iterable bug Something isn't working npu This problem is related to NPU devices pending This problem is yet to be addressed
#8175 opened May 27, 2025 by garyyang85
1 task done
1
RuntimeError: generator raised StopIteration enhancement New feature or request pending This problem is yet to be addressed
#8168 opened May 27, 2025 by cs-mshah
1 task done
PPO ds3报错问题 bug Something isn't working pending This problem is yet to be addressed
#8158 opened May 26, 2025 by Mango17adjz
1 task done
视频DPO训练报错 bug Something isn't working pending This problem is yet to be addressed
#8157 opened May 26, 2025 by zhanghang-official
1 task done
@BUAADreamer
2
Qwen-Omni在混合模态数据上dpo训练时，训练卡住 bug Something isn't working pending This problem is yet to be addressed
#8151 opened May 25, 2025 by wwfnb
1 task done
@Kuangdd01
4
Qwen3-8b模型全参数预训练过程中，grad_norm突然增大，模型训练中止 bug Something isn't working pending This problem is yet to be addressed
#8150 opened May 24, 2025 by hummingbird2030
1 task done
The performance decreases seriously after finetuning on qwen2.5-Omni model with lora bug Something isn't working pending This problem is yet to be addressed
#8146 opened May 23, 2025 by humble-gambler
1 task done
@Kuangdd01
13
Expects torch.Size([525336576]) but got torch.Size([128256, 4096]) bug Something isn't working pending This problem is yet to be addressed
#8142 opened May 23, 2025 by Abhivadan
1 task done
111 bug Something isn't working pending This problem is yet to be addressed
#8141 opened May 23, 2025 by XiaYifen
dlopen: cannot load any more object with static TLS bug Something isn't working pending This problem is yet to be addressed
#8140 opened May 23, 2025 by wangsikuan
1 task done
Loss becomes 0 when using DeepSpeed Zero2 with multi-node training (Zero3 works fine) bug Something isn't working pending This problem is yet to be addressed
#8137 opened May 22, 2025 by JackLingjie
1 task done
1
how to train with vqa bug Something isn't working pending This problem is yet to be addressed
#8132 opened May 22, 2025 by lleye
1 task done
3
ProTip! no:milestone will show everything without a milestone.