-
-
Notifications
You must be signed in to change notification settings - Fork 71
[TinyYolo v2] Bug in maxpool? #73
Comments
Does yolo use overlapping pooling? i.e. maxpool with a stride of other than size |
@chewxy the stride of the operator is |
I think I got it; the ONNX spec specifies that a
The I will do different tests, but If this is the case, I will have to find a smart solution to implement this without breaking Gorgonia's API 🧐 |
I did not notice that the implementation of the
Let's fix this before going further. |
Probably related to ONNX issue #1113 |
The problem linked to #74 and because the padding is asymmetric; |
Should be fixed with PR #80 and PR #295 from Gorgonia |
Tiny yolo v2 is working, let's make it fast now! |
This commit allows the model tiny Yolo v2 to be compiled and executed With Gorgonia.
Sadly the execution does not give the expected result:
According to this blog post the architecture should be:
After setting some logs, the architecture of the decoded network is:
The last layer using the
Maxpool
operator does not give the correct output size.The padding used is computed from the
auto_pad
argument but seems ok (padding is[1,1]
).It requires more investigation; maybe a bug in Gorgonia.
Note : the computation is slow, but
Make it work
, thenMake it fast
cc @chewxy
The text was updated successfully, but these errors were encountered: