When I cycled through multiple images, and for each image, I cycled through multiple layers of the model, I saw an abnormal increase in memory.

[LeonNerd]

When I cycled through multiple images, and for each image, I cycled through multiple layers of the model, I saw an abnormal increase in memory. I just increased the outer loop, and I found that every time I run self.model(images), it causes it to grow

[AdrianCiotinga]

There is a memory leak in the forward function of YOLOV5TorchObjectDetector when assigning sliced tensors back to the original tensor, since the computational graph is kept. Going to submit a PR to fix it (since I need the computational graph), but if you don't need the computational graph wrapping your code in with torch.no_grad(): should fix the issue.