Ryuk

我对多输入模型进行量化时发现维度有问题，使用的模型在[这里](https://github.com/alibaba/MNN/issues/1946)。 我的模型的两个输入是[1, 400, 80]和[1]，config文件也是这么设置的  但是运行时报  我batch_size设置的是32，而903168/28224 = 32，这显然是把一个batch的输入([32, 1, 400, 80])当成单个网络模型输入([1, 400, 80])了

pyMNN doesn't exit

1

运行官方程序mobilenet_demo.py程序不会自动退出，除非使用ctrl+c

when i dump weights with dump_percepnet.py, i meet `missing argument:activation` in this [line](https://github.com/jzi040941/PercepNet/blob/80e2e74af742d56b782745f1b75bb02827651ccd/dump_percepnet.py#L147). It seems that for the dense or conv1d layer, there should be an activation function name, isn't...

The exported onnx model has some useless module

1

**Describe the bug** 我使用如下的参数对非流式的conformer进行转换  得到的onnx模型中的r_cnn_cache并没有参与任何计算，说明这部分是没有用的 

from ._aubio import __version__ as version UnicodeDecodeError: 'utf-8' codec can't decode bytes in position 945-946: invalid continuation byte

1

When i use `import aubio` The following error is reported ``` File "D:\software\python\lib\site-packages\aubio\__init__.py", line 26, in from ._aubio import __version__ as version UnicodeDecodeError: 'utf-8' codec can't decode bytes in position...

For example, what's the following is a comment in in function estnoisem ` # empirical formula for standard error based on Fig 15 of [2] ` but there is no...

打开拼音数据库时候报错误 ``` File "D:\software\python\lib\site-packages\ChineseTone\chinesetone.py", line 68, in getPinyinResource for line in open(os.path.join(CURRENT_DIR, 'data', 'pinyin.db')): UnicodeDecodeError: 'gbk' codec can't decode byte 0x87 in position 2: illegal multibyte sequence ```