search

sacmehta / ESPNet

ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation
https://sacmehta.github.io/ESPNet/
MIT License
541 stars 112 forks source link

The operation of a neural network on a single board computer (Raspberry, Odroid etc.) #16

Closed DenisN03 closed 5 years ago

DenisN03 commented 6 years ago

Hello, @sacmehta! I've one more question. Have you tried to run your neural network architecture on embedded platforms? Under the embedded platform, I mean Raspberry Pi, Orange Pi, Odroid and others. If so, what results did you get on the speed of work (FPS)? If you did not run, why? Do you think this neural network will work on these platforms? If so, how do I run ESPNet on such a platform? At the moment, I managed to run ENet on a single-board Odroid computer, the speed of work was 3 FPS. I'm counting on the fact that your architecture will work faster. Thank you!

sacmehta commented 6 years ago

I have not tried running on these platforms, but any architecture should work.

Instead of using a PyTorch, you should use Caffe2 which is much more optimized. See below link for more details.

https://pytorch.org/tutorials/advanced/super_resolution_with_caffe2.html

DenisN03 commented 6 years ago

I compiled Pytorch 0.3.0 on a single-board computer. When trying to start test of a neural network, an error occurs: AssertionError: Torch not compiled with CUDA enabled When I run test the gpu parameter was set to False. Is it necessary to use pytorch version 0.3.0 or newer versions, too, should work?

engineer1109 commented 6 years ago

@DenisN03 Seem you don’t use cuda pytorch