-
Looking at [the dropout code](https://github.com/benanne/nntools/blob/2840cbc657ff341ef2f1654c1a5916d0fbe3931e/nntools/layers/base.py#L414-428), I see two potential performance problems:
``` python
r…
-
Hi all,
I have just modified caffe to run for dense pixel labelling for my semantic scene understanding task. I am testing that on a very simplistic network with only one convolution layer followed …
-
When training big networks (over 500Mb proto buffers) protobuf cannot read them.
```
A protocol message was rejected because it was too big (more than 536870912 bytes).
To increase the limit (or t…
-
To the best of my understanding there is nothing stopping Caffe to be used as a sliding window detector (instead of using selective search crop proposals as in http://caffe.berkeleyvision.org/imagenet…
-
It took me several hours to finally find this problem.
In my own implementation of dropout in cuda-convnet, I randomly drop half of the nodes during training time, and multiply by one half during test…