Tina J.

It's 2:30am! And still awake, googling for this problem. I hope you can do it with opencv. So after knowing the mesh coordinates, you use opencv for transformation?

Let's make nose smaller. That's the task.

How to use laplacian transformation? Should we pass the nose coordinates or the whole face mesh?

Prob image? Let's assume we stream from laptop camera.

@glenn-jocher Btw, how is the prediction? Should we also specify the range, like `imgsz=[640, 1280]`? Also, I notice the best.pt model produced has is like 24MB while the nano model...

@glenn-jocher Got it. So I trained the multi-res nano model with `imgsz=[650, 1280]`. Now for prediction, the imgsz MUST be within this range? I tried to test the model on...

I just set it the same way you did and it did the training. Does it throw an error for you? How did you realize it is set to a...

Hmm I will check and see if I get any warnings. Also make sure you update ultralyitics to the latest version.

@faketifosi You are right. I just checked and realized I get same Warning. All my experiments are now invalidated :( `WARNING ⚠️ updating to 'imgsz=1200'. 'train' and 'val' imgsz must...

I hope it gets addressed ASAP. It shouldn't be anything more than just a string parsing.