Ryuk

Results 10 comments of Ryuk

请问解决了吗,我也遇到了reshape error的问题[1950](https://github.com/alibaba/MNN/issues/1950)

使用 pip install MNN 安装

What's about the CPU loading?

仔细定位了下问题,发现从createFromFile就有内存泄漏 ![image](https://user-images.githubusercontent.com/22525811/192276088-da06355d-60c3-401f-87bc-030aa88bde55.png) ![image](https://user-images.githubusercontent.com/22525811/192275855-576ea9bc-d547-4809-8e9c-ac59b8cf9631.png)

> macos平台:使用下面这两行代码,得到的类型是MNN.tensor类型,这种类型能转化成torch.tensor然后接着执行模型下面的代码吗? > > ``` > tmp_output = MNN.Tensor((1, 240, 384), MNN.Halide_Type_Float, np.ones([1, 240, 384]).astype(np.float32), > MNN.Tensor_DimensionType_Caffe) > output_tensor.copyToHostTensor(tmp_output) > ``` > > 如果可以的话,怎么转换啊,求助!!! 我记得好像可以转numpy的,然后numpy转torch.tensor

内存泄漏解掉了: 1)申请的Tensor要及时delete 2)Interpreter releaseModel后也要delete 3)但是还有少部分内存处于still reachable状态

Have you solved the problem? I cannot feed the r_cnn_cache with an uncertain dimension to the encoder.

看了下代码,如果使用非流式的模型应该是不包含cache部分,但是模型转出来包括att_cache和图中的cnn_cache

> I solved Can you tell me how you solve it?Thanks in advanced.