haochen wang

Results 6 comments of haochen wang

I think maybe the feature extractor is not well-trained. Please check the performance with only the feature extractor. The mAP should be about 76 without NFormer. The training of the...

> First of all, thank you for your training code, but the accuracy obtained in this way is about 69.4. Do you have a pre-train code? If it is convenient...

Sorry, the visualization is not in the repo

Please check this https://pytorch.org/docs/stable/generated/torch.sparse.softmax.html

https://github.com/haochenheheda/segment-anything-annotator

https://github.com/haochenheheda/segment-anything-annotator