Machine Learning

1 readers

1 users here now

Community Rules:

Be nice. No offensive behavior, insults or attacks: we encourage a diverse community in which members feel safe and have a voice.
Make your post clear and comprehensive: posts that lack insight or effort will be removed. (ex: questions which are easily googled)
Beginner or career related questions go elsewhere. This community is focused in discussion of research and new projects that advance the state-of-the-art.
Limit self-promotion. Comments and posts should be first and foremost about topics of interest to ML observers and practitioners. Limited self-promotion is tolerated, but the sub is not here as merely a source for free advertisement. Such posts will be removed at the discretion of the mods.

founded 1 year ago

MODERATORS

communick@academy.garden

Best model choice to finetune on new classes for object detection? [D] (alien.top)

submitted 11 months ago by WingedTorch@alien.top to c/machinelearning@academy.garden

1 comments fedilink hide all child comments

Let’s say I have trained my OD on the Coco dataset. Which choice of OD would be the best to finetune my model on a limited dataset of additional classes in a specific domain of images (Let’s say 1000 instances per class).

Would Yolov8 be a good choice? Are there even any benchmarks for this? Should I just look for the highest AP on Coco/Pascal?

top 1 comments

sorted by: hot top controversial new old

[–] xsouxsou29@alien.top 1 points 11 months ago

From my experience, yolov5 is better on classes with less examples compare to yolov8. I cannot really explain it, but my guess are the weights in the loss that heavily favor bbox parameters over the class loss. Nowadays it is fairly easy to test different architectures, so do the test by yourself, you'll have an answer that will be more precise than what you can find here :)