About 38k image-text pairs(10k from LAION and 28k from nsfw_detect), and captions are generated by LLaVA-NeXT with prompt "Describe the photo in detail (attributes of person)".
The "txt" column shown in the dataset viewer is originated from LAION, not the captions yielded by LLaVA-NeXT.
pretrained = "lmms-lab/llama3-llava-next-8b"
model_name = "llava_llama3"
device = "cuda:2"
device_map = "auto"
tokenizer, model, image_processor… See the full description on the dataset page:
https://huggingface.co/datasets/K00B404/NSFW-T2I.
