Hi,
have deepstack 2022.01.1 installed. Deepstack would start and work properly for a while. After a few min it would crash with:
The size of tensor a (32) must match the size of tensor b (40) at non-singleton dimension 2
Other pytorch posts mention this happens if the image is too big.
I reduced the images sizes of my 8mp cams and it's working fine now.
Is there a way to limit / reduce the hi-res jpeg size in Blue iris before it gets sent to Deepstack ?
Ossible Bug with Deepstack - KJpeg Image size too big with 8mp cams
Re: Ossible Bug with Deepstack - KJpeg Image size too big with 8mp cams
Just spitballin' here, but check whether AI is using the main stream or sub-stream on the trigger tab. If it's main stream, disable it for Deepstack. If using the sub stream and the image is still too large, changing it at the camera as you've done is the most obvious way. That said, there are options on the Record tab to resize the image, but it's unclear to me if that will affect what's fed to Deepstack.
- "Whenever I take something apart to fix it and put it back together again, I end up with like six really important looking pieces left over" -Tim Allen
- If you know what your after, you'll recognize it when you see it.