All committers have signed the CLA.

Hi @mwitiderrick 👋🏻! Thanks for your interest in supervision and submitting a PR. Will that connector work with every DeepSparse model or only YOLOv5?

Hey @SkalskiP this is specific for DeepSparse detection models that will return a boxes attribute, such as YOLOv5 and YOLOv8

@onuralpszr I see some problems with the 88-character line length in this PR. We have a line in the docstring that can't be made shorter.

@onuralpszr I see some problems with the 88-character line length in this PR. We have a line in the docstring that can't be made shorter.

Well I was going to fix but @mwitiderrick just did it. So all good :)

@mwitiderrick thank you

@mwitiderrick btw If you don't mind can you squash your commits to make it clean or we can handle in merge as well.

@mwitiderrick btw If you don't mind can you squash your commits to make it clean or we can handle in merge as well.

@onuralpszr okay with the option for handling at merge time

@mwitiderrick and @onuralpszr, but that is not valid line braking 🤣 you cant break string like that. Not to mention that example must look good in docs. @onuralpszr to be honest I care more about code example readability in docs than perfect 88-character lines in docsting.

>>> yolo_pipeline = Pipeline.create(
...     task="yolo",
...     model_path="zoo:cv/detection/yolov5-l/pytorch/
...        ultralytics/coco/pruned80_quant-none")
>>> pipeline_outputs = yolo_pipeline(SOURCE_IMAGE_PATH,
...                         iou_thres=0.6, conf_thres=0.001)

@onuralpszr As for squash, I'm in favor. Let's just make sure @mwitiderrick will get the credit for contribution, and that squashed commit won't be assigned to me or other person who click "merge" button.

@mwitiderrick and @onuralpszr, but that is not valid line braking 🤣 you cant break string like that. Not to mention that example must look good in docs. @onuralpszr to be honest I care more about code example readability in docs than perfect 88-character lines in docsting.

>>> yolo_pipeline = Pipeline.create(
...     task="yolo",
...     model_path="zoo:cv/detection/yolov5-l/pytorch/
...        ultralytics/coco/pruned80_quant-none")
>>> pipeline_outputs = yolo_pipeline(SOURCE_IMAGE_PATH,
...                         iou_thres=0.6, conf_thres=0.001)

Sorry I might be decide too quickly. Let me check it out properly. I will handle squash part too.

Is it all good guys ? :))

lgtm

@onuralpszr As for squash, I'm in favor. Let's just make sure @mwitiderrick will get the credit for contribution, and that squashed commit won't be assigned to me or other person who click "merge" button.

"6a2203c" just his name and all clean.

@mwitiderrick again awesome job and thank you again !

@onuralpszr when does the new feature become available via pip install?

@onuralpszr when does the new feature become available via pip install?

#296 (comment). that's the only answer I can find :) but that is question for @SkalskiP :)

Is it all good guys ? :))

@onuralpszr Good enough for now ;) But I still think we need deeper into that formatting. Perfect solution would be 88 characters for code and 120 for docstrings.

when does the new feature become available via pip install?

@mwitiderrick 1-2 weeks is the timeline for the next release. We might do some pre-releases next week.