Search results

Gemma-4-E2B-it-MNN local model causes ANR on Android 15 / OnePlus PKX110 in MNN Chat 0.8.2.2

Status: Open.

#4396 In alibaba/MNN;

· Freeman086 opened on Apr 22, 2026

MNN.podspec 求一个3.5.0版本

Status: Open.

#4395 In alibaba/MNN;

· IDEAL-YANG opened on Apr 22, 2026

MNN 3.5.0 bug: CPUConvolutionDepthwise enables KleidiAI depthwise kernels without an sme2 guard

Status: Open.

#4393 In alibaba/MNN;

· burningfireplace opened on Apr 21, 2026

相同视觉模型, 鸿蒙平台mate60上, cpu占用接近跑满, 安卓机器上未出现

Status: Open.

#4392 In alibaba/MNN;

· bolun365 opened on Apr 21, 2026

qwen2.5 0.5b的模型转换成qnn之后，短的回答正常，但长一点就会有问题，最后的结果是重复某个词。

Status: Open.

#4390 In alibaba/MNN;

· fromwhzz opened on Apr 21, 2026

In version 0.8.2.2 some models Qwen-2.5 Coder and Gemma 3 incorrect work.

Status: Open.

#4388 In alibaba/MNN;

· iyiviry opened on Apr 20, 2026

关于多个具有视觉能力的模型识别问题

Status: Open.

#4387 In alibaba/MNN;

· l00jj opened on Apr 20, 2026

llm_demo 交互式和通过文件推理相同提示词结果不同

Status: Open.

#4385 In alibaba/MNN;

· yuimo opened on Apr 20, 2026

在华为平板edge上尝试使用MNN Chat加载本地模型时，提示需要通过命令行移动模型路径

Status: Open.

#4383 In alibaba/MNN;

· erlingmijiang opened on Apr 17, 2026

如何通过裁剪 op 来压缩 libMNN 库

Status: Open.

#4382 In alibaba/MNN;

· Hughheyu opened on Apr 17, 2026

MnnLlmChat编译问题

Status: Open.

#4380 In alibaba/MNN;

· sys2008 opened on Apr 16, 2026

CUDA Backend Compilation Issue

Status: Open.

#4377 In alibaba/MNN;

· EpicFacepalm94 opened on Apr 16, 2026