search

NVlabs / ODISE

Official PyTorch implementation of ODISE: Open-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models [CVPR 2023 Highlight]
https://arxiv.org/abs/2303.04803
Other
845 stars 45 forks source link

checkpoint for trained model without Implicit Captioner #12

Closed Albert0147 closed 1 year ago

Albert0147 commented 1 year ago

Hi, do you have the checkpoint for the ODISE trained without Implicit Captioner? Thanks.

shalinidemello commented 1 year ago

Sorry! We don't have that checkpoint readily available.

However, we have provided detailed instructions for training the model from scratch, if you'd like to reproduce it.