Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
MonsterMMORPG 
posted an update 7 days ago
Post
3285
SANA: Ultra HD Fast Text to Image Model from NVIDIA Step by Step Tutorial on Windows, Cloud & Kaggle — Generate 2048x2048 Images

Below is YouTube link for step by step tutorial and a 1-Click to installer having very advanced Gradio APP to use newest Text-to-Image SANA Model on your Windows PC locally and also on cloud services such as Massed Compute, RunPod and free Kaggle.

https://youtu.be/KW-MHmoNcqo

This above tutorial covers the newest SANA 2K model and I predict SANA 4K model will be published as well. Sana 2K model is 4 MegaPixel so it can generate the following aspect ratio and resolutions very well:

“1:1”: (2048, 2048), “4:3”: (2304, 1792), “3:4”: (1792, 2304),
“3:2”: (2432, 1664), “2:3”: (1664, 2432), “16:9”: (2688, 1536),
“9:16”: (1536, 2688), “21:9”: (3072, 1280), “9:21”: (1280, 3072),
“4:5”: (1792, 2240), “5:4”: (2240, 1792)

I have developed an amazing Gradio app with so many new features :

VAE auto offloading to reduce VRAM usage significantly which is not exists on official pipeline

Gradio APP built upon official pipeline with improvements so works perfect

Batch size working perfect

Number of images working perfect

Multi-line prompting working perfect

Aspect ratios for both 1K and 2K models working perfect

Randomized seed working perfect

1-Click installers for Windows (using Python 3.10 and VENV — isolated), RunPod, Massed Compute and even a free Kaggle account notebook

With proper latest libraries working perfect speed on Windows too

Automatically properly saving every generated image into accurate folder

🔗 Full Instructions, Configs, Installers, Information and Links Shared Post (the one used in the tutorial) ⤵️
▶️ https://www.patreon.com/posts/click-to-open-post-used-in-tutorial-116474081

🔗 SECourses Official Discord 9500+ Members ⤵️
▶️ https://discord.com/servers/software-engineering-courses-secourses-772774097734074388

If it has non-commercial license, that is proprietary license and not free, generating large confusion for the user on how to use the software and works generated.

Please, why not make full free software license and follow the path that Microsoft or IBM have taken with their models like Phi or Granite, or Alibaba Group Qwen models licensed under Apache 2.0 or Mistral models, so please consider changing it to full free software license.

·

@JLouisBiz valid points but if you ask my personal opinion, as long as your not a SaaS , i never saw these companies care. Moreover they are also using proprietary copyrighted data to train so i don't think they have strong hand against you. Also still AI outputs are not copyrightable as far as i know