Search results

Qwen3.5 nothink

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10282 In hiyouga/LlamaFactory;

· ecokayiza opened on Mar 14, 2026

为LlamaFactory 集成 HyperParallel 分布式训练后端

enhancementNew feature or requestNew feature or request

npuThis problem is related to NPU devicesThis problem is related to NPU devices

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

Status: Open.

#10278 In hiyouga/LlamaFactory;

· Cui-yshoho opened on Mar 12, 2026

The difference between the qwen3_5_nothink suffix and the enable_thinking parameter.

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10277 In hiyouga/LlamaFactory;

· Kwen-Chen opened on Mar 12, 2026

Issues with Qwen3-VL quantization

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10275 In hiyouga/LlamaFactory;

· fancyerii opened on Mar 12, 2026

RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10261 In hiyouga/LlamaFactory;

· ImsuperSH opened on Mar 9, 2026

Cannot import name 'load_and_register_kernel' from 'transformers.integrations.hub_kernels'

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10256 In hiyouga/LlamaFactory;

· gxx27 opened on Mar 7, 2026

Support Liger Kernel for Qwen3.5

enhancementNew feature or requestNew feature or request

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

Status: Open.

#10253 In hiyouga/LlamaFactory;

· wyt2000 opened on Mar 6, 2026

[BUG] Ray Demo Unused Keys

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10251 In hiyouga/LlamaFactory;

· SnowCharmQ opened on Mar 5, 2026

TypeError when saving predictions with HuggingFace Dataset

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10250 In hiyouga/LlamaFactory;

· pyxnpyx opened on Mar 5, 2026

华为昇腾使用accelerate调度器微调qwen3-next-80B，但是在配置文件中deepspeed_moe_layer_cls_names不知道应该写什么？

bugSomething isn't workingSomething isn't working

npuThis problem is related to NPU devicesThis problem is related to NPU devices

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

Status: Open.

#10249 In hiyouga/LlamaFactory;

· gongyinfeng0206-design opened on Mar 5, 2026

transformers5.2.0训练moe模型，如果使用torch_npu的moe优化算子，loss会变成nan

bugSomething isn't workingSomething isn't working

npuThis problem is related to NPU devicesThis problem is related to NPU devices

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

Status: Open.

#10248 In hiyouga/LlamaFactory;

· piekey1994 opened on Mar 5, 2026

ChatML template bug in test chat with unsloth

bugSomething isn't workingSomething isn't working

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

Status: Open.

#10247 In hiyouga/LlamaFactory;

· mykeehu opened on Mar 4, 2026