Mactarvish comments

Results 13 comments of


                                            Mactarvish

Not getting input_raspicam when compiling on 64bit raspbian

`sudo apt-get install libraspberrypi-dev` That works! Note that I'm on 4B+ with `Linux ubuntu 5.4.0-1069-raspi #79-Ubuntu SMP PREEMPT Thu Aug 18 18:15:22 UTC 2022 aarch64 aarch64 aarch64 GNU/Linux` instead of...

can't compile on raspbian x64

same issue

How to set pylance memory max limitation?

Otherwise, if there anyway to let pylance NOT analysis special directory or soft link? I've found that when I create a soft link to my HUGE dataset that causes pylance...

How to set pylance memory max limitation?

Thanks for your helps. Seems the key point is not depth of directory but num of files. In the soft link under my workspace there's more than 50000000 .jpg files...

转换好的模型进行模型推导extractor.extract卡在matmul层，请问是什么问题有什么办法解决吗

检查一下matmul的输入形状是否正确

谷歌浏览器创建流失败

Error: Running the contributed command: 'remote-x11-ssh.connect' failed.

Thanks, I'll try your method first.

[feature] Support multi-gpu test multiple datasets separately

> Hi @Mactarvish , thank you for your contribution. I think this feature can be implemented in a new Evaluator `MultiDatasetEvaluator` rather than in the `Metric`. We can discuss this...

实时语音识别报错：terminate called after throwing an instance of 'std::out_of_range'what(): _Map_base::at

> @Mactarvish 可以回复一下么？不然我们要 close 这个 issue 了. 不好意思我一直在排查嗷，平台是x86 Ubuntu22.04 在戴尔笔记本上运行的线程数是1 代码基于 [8af2af8](https://github.com/fengyongronglu/sherpa-onnx/commit/8af2af84664d3285ba452bf453bb928a3eb6e978) 我根据`c-api-alsa.cc`改写的代码，目前来看好像是因为我定义了一个全局的std::unordered_map对象（没有声明为static），然后就会引起这个问题（相同的代码在Ubuntu 20.04上编译运行是正常的），怀疑是堆结构被破坏了我这两天整理一下最小可复现代码发上来。多谢支持

实时语音识别报错：terminate called after throwing an instance of 'std::out_of_range'what(): _Map_base::at

澄清了😅是我这边的一个问题两个cpp同时include了一个hpp，这个hpp里定义了一个全局的unordered_map导致析构的时候重复析构不过只有在被访问到的时候才会出错，误以为是sherpa的问题了