AtlantixJJ / PVA-CelebAHQ-IDI

Parallel Visual Attention (WACV 2024) and CelebAHQ Identity-Preserving Inpainting dataset repository.
9 stars 1 forks source link

Fine-tuning and inference on custom images #2

Open alasokolova opened 7 months ago

alasokolova commented 7 months ago

Hey, how can I fine-tune and infer the trained model on the custom images? What preprocessing should i conduct? I found create_id_inpaint_dataset script in the repo, but it uses IDInpaintDataset which is not implemented here. Could you please add the implementation or explain how to process data?

AtlantixJJ commented 7 months ago

This script is used to preprocess the existing CelebAHQ dataset into the id-organized dataset, CelebAHQ-IDI dataset. Indeed, this script can be used to process custom data. The IDInpaintDataset is the same as CelebAHQIDIDataset dataset in dataset.py.

AtlantixJJ commented 7 months ago

I uploaded the older dataset with IDInpaintDataset class in script/dataset_old.py

alasokolova commented 7 months ago

Thank you very much for the reply! I will try to run it