Seppo Enarvi
Seppo Enarvi
I can try to do that.
Here's some kind of a benchmark script: ```python import time import torch try: from torchmetrics.detection import MeanAveragePrecision except ImportError: from torchmetrics.detection import MAP MeanAveragePrecision = MAP total_time = dict() class...
Hi @stancld , sorry I didn't have time to respond earlier. I wrote my observations in that pull request. It was indeed a lot faster, but in one case the...
@stancld Hmm. The data's not public. I wonder if you could debug it using random boxes, like in the speed test. I modified it to make the task a little...
Would Peter like to take a look at the small change in CMake configuration?
> Overall, I like the code, though I have to say I haven't combed through it with a very fine comb 😄 One major thing at the moment is if...
I noticed that the iou and giou losses are also included in Bolts: https://github.com/Lightning-AI/lightning-bolts/blob/master/pl_bolts/losses/object_detection.py But the plan was to reduce code duplication between different PyTorch projects, so I would remove...
> > I noticed that the iou and giou losses are also included in Bolts: https://github.com/Lightning-AI/lightning-bolts/blob/master/pl_bolts/losses/object_detection.py > > But the plan was to reduce code duplication between different PyTorch projects,...
> Yes, @senarvi, sorry for not being clearer. It would be great if you could manage the removing in this PR. I could help, but it seems I don't have...
> could you please merge master and then we should be good to go @otaj Done.