Closed adietrichs closed 1 year ago
Implements support for an optional negative prompt that is used to initialize the unconditional context. This is a feature that has been added to many sd forks, and has been requested in https://github.com/divamgupta/stable-diffusion-tensorflow/issues/48.
Example Colab: https://colab.research.google.com/drive/1QsTve3BTzOUZDIAdzrzur487USloZvjX?usp=sharing Note that tf.random.set_seed(0) in the colab examples is required to make seeds consistent across generations (see https://github.com/divamgupta/stable-diffusion-tensorflow/issues/25)
tf.random.set_seed(0)
Also updates text2image.py and img2img.py to expose a --negative-prompt argument. Example usage:
text2image.py
img2img.py
--negative-prompt
python text2image.py --prompt="hilly landscape" --negative-prompt="grass"
😎
Implements support for an optional negative prompt that is used to initialize the unconditional context. This is a feature that has been added to many sd forks, and has been requested in https://github.com/divamgupta/stable-diffusion-tensorflow/issues/48.
Example Colab: https://colab.research.google.com/drive/1QsTve3BTzOUZDIAdzrzur487USloZvjX?usp=sharing Note that
tf.random.set_seed(0)
in the colab examples is required to make seeds consistent across generations (see https://github.com/divamgupta/stable-diffusion-tensorflow/issues/25)Also updates
text2image.py
andimg2img.py
to expose a--negative-prompt
argument. Example usage: