FapMachine Alpha
An experiment on training a model by feeding the network with data created by another AI
Description
FapMachine is an experiment, as stated above, with the goal of recognizing naked or dressed women without being feeded with any real world image. Be aware: it can be considered NSFW even if there are no NSFW images included.
Dataset used
50 Images of naked women generated by Stable Diffusion (through DiffusionBee) 50 Images of dressed women generated by Stable Diffusion (through DiffusionBee)
Training method
Liner.ai training with Image Classification mode
Type of network
EfficientNet with Early Stop, 1000 iterations
Result
70% Accuracy and 0.3 loss values
How to test
You can clone this repository and rename 20d.png as image.png or use any image you want renaming it as image.png, then run the python file to see the prediction result
Disclaimer
This model is intended to show the possibility of autofeeding a network with ai generated data
- Downloads last month
- 14
# Note: 'keras<3.x' or 'tf_keras' must be installed (legacy) # See https://github.com/keras-team/tf-keras for more details. from huggingface_hub import from_pretrained_keras model = from_pretrained_keras("tcsenpai/FapMachine")