Search results

qwen3-5模型微调环境支持

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10377 In hiyouga/LlamaFactory;

· voyag8r opened on Apr 10, 2026

mca格式转hf格式的时候会报experts 的 added but not converted warning 这个会影响最终效果嘛

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10375 In hiyouga/LlamaFactory;

· LynnLee1226 opened on Apr 10, 2026

AWQ量化模型进行微调出现AwqExllamaV2QuantLinear相关异常

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10374 In hiyouga/LlamaFactory;

· llz-coder opened on Apr 10, 2026

Failed pt on Gemma4-31B

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10372 In hiyouga/LlamaFactory;

· markmochi200 opened on Apr 10, 2026

使用SFT微调模型后，在llamafactory中的chat进行测试，效果很好，但是使用openai api推理访问，效果就较差

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10371 In hiyouga/LlamaFactory;

· noscarljw opened on Apr 9, 2026

Qwen3.5-27B lora微调时128步报错oom

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10367 In hiyouga/LlamaFactory;

· ENg-122 opened on Apr 8, 2026

Error when loading dataset reason_tool_use_demo_50 with custom dataset_info.

enhancementNew feature or requestNew feature or request

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

Status: Open.

#10365 In hiyouga/LlamaFactory;

· UCC-team opened on Apr 8, 2026

Gemma4目前只支持thinking吗？

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10364 In hiyouga/LlamaFactory;

· FreeLo0op opened on Apr 7, 2026

GLM4应用ligerkernel

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10363 In hiyouga/LlamaFactory;

· YisuZhou opened on Apr 7, 2026

gemma4 finetune loss extreamly high with deepseed Zero stage3

solvedThis problem has been already solvedThis problem has been already solved

Status: Open.

#10355 In hiyouga/LlamaFactory;

· Kuangdd01 opened on Apr 5, 2026

不支持MPS吗？

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10351 In hiyouga/LlamaFactory;

· kekekekekeshi opened on Apr 4, 2026

Ray distributed training: placement group over-allocates resources and NCCL hang when head node has no GPU

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10348 In hiyouga/LlamaFactory;

· ilover311 opened on Apr 3, 2026