Christian Bartz
Christian Bartz
Did you look at the predictions (my first point of the last answer?) Those images are meant as a help to determine what the network does over time. This really...
Hi, 1. You can verify that training is going correctly, by having a look at the images located in the `bboxes` folder in your `log_dir`. Those images show the predictions...
alright: 1. I think its not working because the images could be too large (meaning width and/or height), to be fit into a video container. You could set the keyword...
Yes, the output of the spatial transformer sampling mechanism always has the same spatial size. This could be 50x50 or 100x32, depends on what you want to do. The sampled...
Yes, there are some tricks you can try: 1. decrease initial learning rate, you should get good results with learning rates around `1e-4` to `1e-6` 2. if the network starts...
See my answer to one of your questions [here](https://github.com/Bartzi/see/issues/53#issuecomment-471486061)...
Did you try to install cupy with verbose logging on? Did you check that cupy is able to find NCCL? How do you install cupy?
I don't know what to say except that chainer can not find your GPU. Have you tried turning it off and on again?
hmm, I'm not sure but `chainer==3.2.0` and `cupy==5.3.0` does not sound like a good combination, you should try to use a chainer version that is similar to the cupy version...
did you try installing cupy with `pip install cupy`? And letting the system compile it for you?