Feature/sg 1488 yolo nas r integration model #2001
Conversation
…feature/SG-1488-YoloNAS-R-integration-model
| ] | ||
|
|
||
|
|
||
| class AbstractOBBDetectionDecodingModule(nn.Module): |
There was a problem hiding this comment.
Should we keep this entire interface if the only difference between it and AbstractObjectDetectionDecodingModule is the returned item from get_output_names?
There was a problem hiding this comment.
Fixed that in fde191e, thanks!
src/super_gradients/module_interfaces/exportable_obb_detector.py
Outdated
Show resolved
Hide resolved
src/super_gradients/training/models/detection_models/yolo_nas_r/yolo_nas_r_dfl_head.py
Outdated
Show resolved
Hide resolved
There was a problem hiding this comment.
Looks great!
Apart from inline comments and ongoing review, some suggestionsfor the notebook:
- Limitations part: maybe its best to put those at the top?
- I didn't see any mentioning of TRT support, maybe I missed it (ctrl+f tensorrt / trt to verify myself).
Also
- Work on albu support - I think this is a valuable feature. Maybe add a section in our Albu suport botebook for rotated boxes too ?
…ect with bool tensors
There was a problem hiding this comment.
So the only comment I have left is https://github.com/Deci-AI/super-gradients/pull/2001/files#r1615586708
Didn't we say we can make it inherit from AbstractObjectDetectionDecodingModule?
Sorry if Im mixing some of our talks up...
That's correct. I forgot to add that, thanks for noticing. Fixed in fde191e |
BloodAxe
merged commit e9e48d5
into
feature/SG-1488-YoloNAS-R-integration
This PR contains: