Search results

LoRA-Pro and LoRA-Drop support

enhancementNew feature or requestNew feature or request

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8874 In hiyouga/LLaMA-Factory;

· guotong1988 opened on Aug 10, 2025

Indexed dataset entry.

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8873 In hiyouga/LLaMA-Factory;

· WeizhenWang-1210 opened on Aug 10, 2025

gpt-oss 20B 多卡sft

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8872 In hiyouga/LLaMA-Factory;

· bobzhang208 opened on Aug 9, 2025

多机多卡训练一段时间后报错Process group watchdog thread terminated with exception: CUDA error: unspecified launch failure

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8871 In hiyouga/LLaMA-Factory;

· Majiawei opened on Aug 9, 2025

流式数据和num_workers冲突？

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8870 In hiyouga/LLaMA-Factory;

· behrenskong opened on Aug 8, 2025

自定义loss进行训练的时候，训练中途报oom错误。请问是否可以优化？不修改loss的时候训练正常

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8868 In hiyouga/LLaMA-Factory;

· Jesayy opened on Aug 8, 2025

多机训练时卡住

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8864 In hiyouga/LLaMA-Factory;

· Xuchongchao opened on Aug 8, 2025

考虑加入 GSPO 训练方法

enhancementNew feature or requestNew feature or request

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8857 In hiyouga/LLaMA-Factory;

· wojiaoyishang opened on Aug 7, 2025

qwen2.5VL-7b在chat 中的推理的结果和我用transformer 脚本推理的结果差异较大

solvedThis problem has been already solvedThis problem has been already solved

Status: Open.

#8856 In hiyouga/LLaMA-Factory;

· mythzZzZ opened on Aug 7, 2025

Support MPO (Mixed Preference Optimization)

enhancementNew feature or requestNew feature or request

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8852 In hiyouga/LLaMA-Factory;

· xiaosu-zhu opened on Aug 7, 2025

模型部署完之后调用接口一直超时无响应

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8850 In hiyouga/LLaMA-Factory;

· DevilMay-Cry opened on Aug 7, 2025

如何使用自己的代码加载并推理reward model？

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8849 In hiyouga/LLaMA-Factory;

· jinzhuoran opened on Aug 7, 2025