search

Vanint / DatasetExpansion

The repository of Expanding Small-Scale Datasets with Guided Imagination (NeurIPS 2023).
73 stars 2 forks source link

Details about finetuning SD on medical image #9

Open won548 opened 6 months ago

won548 commented 6 months ago

Hello.

I am currently trying your work on medical domain based on Dreambooth.

But there are a lot of missing details on finetuning with SD in this paper (Only number of training image mentioned).

Can you tell about more details based on the repository above? (e.g. class word, prompt, training time, etc.)

Vanint commented 6 months ago

We use the huggingface implementation for dreambooth training. Please refer to https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth.py

Moreover, please refer to the training scripts and config files in the corresponding dreambooth_breastmnist, dreambooth_organsmnist, dreambooth_pathmnist files, where you can find class prompts.

won548 commented 6 months ago

I just checked config files and the code in the link above.

But, i found some missing arguments not defined in the code. (configs_list, pretrained_vae_name_or_path, etc.)

These arguments are defined in this repository, maybe the different version or branch of huggingface diffusers.

Also, some of arguments are defined in this file, which is custom diffusion proposed for multi concepts.

I am really confused which repository or code you used for finetuning process in this work.

Could you clarify which version or code you used?

jeeyung commented 5 months ago

I am confused as well.

Vanint commented 5 months ago

I just checked config files and the code in the link above.

But, i found some missing arguments not defined in the code. (configs_list, pretrained_vae_name_or_path, etc.)

These arguments are defined in this repository, maybe the different version or branch of huggingface diffusers.

Also, some of arguments are defined in this file, which is custom diffusion proposed for multi concepts.

I am really confused which repository or code you used for finetuning process in this work.

Could you clarify which version or code you used?

Hi, You may refer to files in GIF_SD/dreambooth_organsmnist, GIF_SD/dreambooth_breastmnist, and GIF_SD/dreambooth_pathmnist for the information (e.g., configs_list, pretrained_vae_name_or_path, etc.).

jeeyung commented 5 months ago

Could you check it again? If you try to run the code, you'll find that what you provided doesn't work. :(