Looking to replicate these kind of effects in stable diffusion. Anyone know what prompts/techniques would be involved? I'd guess they used img2img + ebsynth?

• Dreambooth-SD-optimized

  26 341 1.8 Jupyter Notebook

  Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) with Stable Diffusion

I use this dreambooth repo to train the SD model. https://github.com/gammagec/Dreambooth-SD-optimized Here's the video that shows how to install it in very good detail. https://youtu.be/TwhqmkzdH3s He uses it to train in a face but you can use it to train in a style as well. I suggest taking about 15 or 20 detailed frames of the video and training it in as a style for the class name. You'll have to experiment with how many training steps to take. I suggest doing 1,000 steps at a time and testing out the model. Also don't leave the default "sks", the researchers forgot that that's a common acronym if you know what I mean. Do something like my_style1 so the model doesn't get confused with something else.

• InfluxDB

  www.influxdata.com featured

  Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.

  

Suggest a related project