OOTDiffusion
OOTDiffusion copied to clipboard
levihsu
Official implementation of OOTDiffusion: Outfitting Fusion based Latent Diffusion for Controllable Virtual Try-on
I understand apply_forward_hook is from the older versions of diffusions. Want to know its replacement in the latest version?
Hi @levihsu, do you have any plan to train this model on larger dataset (~1M images)? I can provide you A100 cluster on AWS and also help with data scrapping...
I've noticed a significant difference in results between my locally installed Gradio demo and the online demo at https://ootd.ibot.cn/. Despite changing the seed and cfg multiple times locally, the results...
I have been testing the model for a few hours and a few issues that I have found with custom images I have listed here. This is just meant to...
Recently, I reproduced the OOTD training code and trained it. I would like to ask how low the loss is considered normal. My loss was 1.2 at the beginning, and...
How do I extract image after running the API using gradio_client? It seems to return a temporary image path such as [{'image': '/tmp/gradio/c8ef09d7a7f335611efa026f92fe9998aba924ef/image.png', 'caption': None}]
How much GPU memory do I need for trainning? I have A100 x 8 (40G). I got CUDA oom error when i try trainning. ( i wrote own trainning code...
**environment:**  **cmd:** python run_ootd.py --model_path /home/OOTDiffusion/run/examples/model/model_1.png --cloth_path /home/OOTDiffusion/run/examples/garment/00055_00.jpg --model_type dc --category 2 --scale 2.0 --sample 4 **error:** /opt/conda/lib/python3.10/site-packages/transformers/utils/generic.py:441: UserWarning: torch.utils._pytree._register_pytree_node is deprecated. Please use torch.utils._pytree.register_pytree_node instead. _torch_pytree._register_pytree_node( /opt/conda/lib/python3.10/site-packages/transformers/utils/generic.py:309: UserWarning:...
