Dataset
Hello. Thanks for sharing such an amazing model! I am very interested on which dataset you trained this model ?. I want to add images in the training set and retrain the model on my collected images. Can you share information about 1)Which datasets did you combine ? 2)Did you use all images from each dataset or some fraction ?
The datasets have been collectively taken from Hugging Face Datasets and Kaggle. The datasets used here were published almost 4–5 years ago. I recommend searching for more recent datasets to achieve better results for current tasks.
For more, visit:
Concatenated Same-Type Dataset – Combining multiple datasets of the same type into a single dataset for better training efficiency.
Modular Fraction Combined Dataset with Full Train Split – Merging modular dataset fractions while ensuring a complete training split for balanced learning.
This is a new dataset based on recent curation : https://huggingface.co/datasets/prithivMLmods/Deepfake-vs-Real