Reproduce the results on Nuscenes

[huangzhengxiang]

Thanks for your great work.
I reproduced the results on Nuscenes dataset with your pre-trained checkpoint, following your instructions in README.md, and got the following outputs:
iou
57.5 & 52.1
pq
51.5 & 47.7
sq
75.5 & 73.1
rq
68.3 & 65.3
Is this result expected? Is there any error? Is this the result of StretchBEV-P?
Looking forward to your response. Thank you in advance.

[kaanakan]

Hello,

Sorry for the very late response. The results can change from training to training (also, from inference to inference) since our model is stochastic. You are right, these are the results of StretchBEV-P.

Feel free to ask if you have further questions.