Just trying to get a sense of what was used so I don't end up training the model on a classification with images it was already trained on, as much as I can. Also would be helpful building a new dataset for that class since it's been 5 years since training so perhaps those sources have new images to use.
Thank you @GantMan, much appreciated all the work you put into this!
Just wondering what was the source for the Neutral class?
I noticed that https://github.com/alex000kim/nsfw_data_scraper and https://github.com/EBazarov/nsfw_data_source_urls don't have a Neutral dataset, so what did you use for that training? Do you have a list of subreddits you scraped, or any other sources you used?
Just trying to get a sense of what was used so I don't end up training the model on a classification with images it was already trained on, as much as I can. Also would be helpful building a new dataset for that class since it's been 5 years since training so perhaps those sources have new images to use.
Thank you @GantMan, much appreciated all the work you put into this!