InstructDiffusion
InstructDiffusion copied to clipboard
cientgu
PyTorch implementation of InstructDiffusion, a unifying and generic framework for aligning computer vision tasks with human instructions.
1. Image Editing in the Wild (IEIW), 2. Object replacement based on SA-1B, Open Images 3. PhraseCut with LAMA 4. Web crawl.
Very nice work, can you explain how the web demo works? I often get an error
Hi thanks for opensourcing the awesome work. Is there any specific template to add or remove subject in the instruction ? I tried to add a second subject but it...
Thanks for your great work! I am curious about the the result of detection in your paper in Section 4.9.You mention that you design special prompts for this task. I...
``` |-- gqa-inpaint | |-- images | |-- images_inpainted | |-- masks | |-- train_scenes.json | `-- meta_info.json `-- MagicBrush |-- data |-- processed-train `-- magic_train.json ``` - How to...
``` CUDA_VISIBLE_DEVICES=2 python -m torch.distributed.launch --nproc_per_node=1 main.py --name v0 --base configs/instruct_diffusion.yaml --train --logdir logs/instruct_diffusion ``` I met this error **"[rank0]: RuntimeError: expected scalar type Half but found Float"** ``` [rank0]:...
I found with original training workflow, the loss is not decling, I am not sure this is because I am using a subset of the training set. ``` # File...