Xueyan Zou
Xueyan Zou
The checkpoint is different, you could see the RED highlight in the HTML webpage of your running table. We release seem_focalt_v1.pt and the demo we are running is on large...
I am running a large model with publishable checkpoint. But expected to be slightly worse than the demo checkpoint.
That would be better than focalt checkpoint. It is trained with in21k pretrained focal-large-xdecoder.
As you may noticed, this one is still training. Expected 1-2 days to be finished without interrupt.
We have released the focal-large checkpoint and config, feel free to play with it.
I haven't test the fps. this also depends on the gpu you are using. On one V100 gpu, a 20s video will take 20s to process in total. Something like...
We do not support inference code currently. @bratjay01 mentioned please refer to those documents.
Thanks so much for your reminder, I just updated the new files at: https://huggingface.co/xdecoder/SEEM/blob/main/panoptic_train2017_filtrefgumdval.json https://huggingface.co/xdecoder/SEEM/blob/main/grounding_train2017_filtrefgumd.json https://huggingface.co/xdecoder/SEEM/blob/main/coco_train2017_filtrefgumdval_lvis.json https://huggingface.co/xdecoder/SEEM/blob/main/captions_train2017_filtrefgumdval.json
This could be download from the official website