configu VS generative-models

To join Configu and start building up, Checkout: Github : Configu

As we celebrate Hacktoberfest, we at Configu invite you to be part of our mission to redefine software configuration management. We've set out to tackle the persistent challenge of configuration chaos, and we're making strides every day. If you're searching for a place to make a significant impact this Hacktoberfest, consider Configu. Delve into our open-source repository, understand our vision, and contribute to shaping our journey. If you're unsure where to begin or need some help along the way, our Configu Discord community is always here to guide you. For newcomers, we recommend starting with issues labeled 'good-first-issues'.

Enter Configu, an open source implementation of the concept of Configuration-as-Code, ensuring that the code remains the source of truth. But we didn't stop there. We've just launched a new feature called configu find that takes configuration management to the next level.

Thanks Geva for sharing my article.

In this piece, I delve into the world of application configuration, discussing the drawbacks of using environment variables stored in env files and introducing the powerful alternative - Configuration as Code (CaC).

Looking forward to your thoughts! and If you're intrigued, I invite you to explore our open-source project on GitHub: https://github.com/configu/configu

git clone https://github.com/Stability-AI/generative-models.git && cd generative-models

Generative Models by Stability AI Github Repository

It looks like the huggingface page links their github that seems to have python scripts to run these: https://github.com/Stability-AI/generative-models

However, when using the example code from HuggingFace or setting up stuff from the StabilityAI/generative-models repo in a jupyter notebook, I end up using 21 GB of VRAM just for running the default pipeline (with no base model output). If I try to run the extra `base.vae.decode(base_latents)` after generation to get unrefined outputs, I get a CUDA out of memory error as it blows past the 24GB of my NVIDIA RTX 3090.

SDXL uses cropping information as part of the conditioning. Images were randomly cropped during training and the coordinates of the crop were included as two integers at the end of the conditioning vector. If you're using ComfyUI you can use the CLIPTextEncodeSDXL node to specify where the upper left corner of the image should appear to be in relation to some hypothetical uncropped image. Here's a figure with examples from the report on SDXL: