Zeyi Sun
@zhanghongyong123456 this doesn't work for me...still OOM, can you share more about you GPU?
I believe [ComfyUI](https://comfyanonymous.github.io/ComfyUI_examples/video/) is a better choice for GPU with memory less than 80G, as it use `xformer` to reduce memory cost.
Paper name/title: Alpha-CLIP: A CLIP Model Focusing on Wherever You Want Paper link: https://arxiv.org/abs/2312.03818 Code link: https://github.com/SunzeY/AlphaCLIP
@yuedajiong Could you show the code of this? I am trying the solution mentioned in [here](https://github.com/mikedh/trimesh/issues/1399). But I am still having trouble writing the correct code...
I have recently written a [script](https://github.com/SunzeY/awesome-blender-script/blob/main/convert_to_std_obj.py) to convert TripoSR exported `.obj` file into standard obj file with separated `.png` for kd_diffuse texture using Blender. Check it out if you have...
Thank you for you kind reply! Is it possible to achieve this automatically using script or package(like `trimesh`)? we are now researching some operations directly on texture.png and want to...
Still looking forward!
We have launched training code as well as our data. Sorry for keep you waiting for so long.
We have launched training code as well as our data. you can finetuning bigger model on your own, we currently don't have plan to finetune due to limited resources.
the hugging face demo is in fact pointing to openxlab demo, where an A10 is allocated for running demo. You can go to openxlab link to awake the demo when...