roguedarkjedi/dogcam-ai

Detection processes very slowly

roguedarkjedi opened this issue · 4 comments

We should look into potentially doing some image resizing operations to lower the amount of process time that this takes. I feel like with this addition, images will use less memory and CPU process time.

Additionally, we can drop additional filtration methods as it won't be strictly necessary.

A better more focused model would also help too

More focused model will really help, also some small optimizations.

Also the current method we're using is really jank and does a couple extra cycles that are likely unnecessary.

New code is much faster, the model is a bit tighter using SSDs.