Search results

mnn hiai后端不能执行int8推理

Status: Open.

#3962 In alibaba/MNN;

· szh6 opened on Oct 20, 2025

MNN对qat量化的支持情况咨询

Status: Open.

#3961 In alibaba/MNN;

· yll3518974 opened on Oct 20, 2025

Themed/adaptive app icon

Status: Open.

#3960 In alibaba/MNN;

· NielSync opened on Oct 19, 2025

I ask stable diffusion v1 5 mnn to generate a picture about"一条狗", but it gives me a photo of lots of scaring monkeys.

Status: Open.

#3959 In alibaba/MNN;

· zcr0312 opened on Oct 19, 2025

[ QUESTION ] Why apk release don't appears in github release section

Status: Open.

#3958 In alibaba/MNN;

· bi4key opened on Oct 18, 2025

2507系列18G模型在手机上运行内存不够闪退

Status: Open.

#3957 In alibaba/MNN;

· vallageandboatandtree opened on Oct 18, 2025

3.2.5Android预编译版本vulkan崩溃

Status: Open.

#3956 In alibaba/MNN;

· justinamei opened on Oct 18, 2025

gemma3_text config没有layer_types这个字段

Status: Open.

#3955 In alibaba/MNN;

· tongwei0828 opened on Oct 17, 2025

Why Replace the KleidiAI macro with the hint config

Status: Open.

#3952 In alibaba/MNN;

· andyt9527 opened on Oct 16, 2025

TaoAvatar能将部分模型（ASR，TTS等）由本地推理替换为通过网络调用云端的API接口吗

Status: Open.

#3949 In alibaba/MNN;

· joeshow79 opened on Oct 16, 2025

[Bug] llm_bench 在 ARM CPU 上测试 gemma-3-1b-it 模型时出现异常高的解码速度（1500 toks/s）

Status: Open.

#3947 In alibaba/MNN;

· jialilve opened on Oct 16, 2025

模型量化是FP16 设备ARMV8.2以上 如何确定CPU模式下precision_low就是按照FP16推理呢。实际测试内存占用发现precision_low和precision_normal相差不大

Status: Open.

#3946 In alibaba/MNN;

· FSRyan opened on Oct 15, 2025