mr-lab

they use a different way of training i think, they concatenate the color image to depth image just like inpainting training , but i think different ... have you covered...

you may be on to something  reading https://github.com/VAST-AI-Research/TripoSR/blob/main/tsr/system.py line 163 and after will see if the marching cubes not working correctly or simply an issue with the resolution 256...

kind of better results  system.py line 171 change threshold from 25 to 5 i think this works for solid objects like an apple resolutions wont fix it 512 takes...

> > simply change model.extract_mesh(scene_codes) to model.extract_mesh(scene_codes, resolution=some_integer). > > Any chance this could be exposed in the Gradio UI? here is how to fully expose it in the UI...

> I think it could be due to marching cubes with resolution 256. Could you please try with higher resolutions and see if things get better? To use different resolutions,...

> Threshold should actually be more like 0. It's probably intended as a basic de-noising mechanism; but it's going to cause weird cutoffs. > > It might be worth testing...

> > kind of better results  system.py line 171 change threshold from 25 to 5 i think this works for solid objects like an apple resolutions wont fix it...

put more effort will you . it's clearly objaverse . just look at "Dataset used to train" on huggingface .

> we realize the value of SUPIR and the power it can have to help our future research. Because keep improve SUPIR can be hardly done with the resource in...

> > > we realize the value of SUPIR and the power it can have to help our future research. Because keep improve SUPIR can be hardly done with the...