SANA1.5_4.8B_1024px

<p align="center" style="border-radius: 10px"> <img src="https://raw.githubusercontent.com/NVlabs/Sana/refs/heads/main/asset/logo.png" width="35%" alt="logo"/> </p> <div style="display:flex;justify-content: center"> <a href="https://huggingface.co/collections/Efficient-Large-Model/sana-15-67d6803867cb21c230b780e4"><img src="https://img.shields.io/static/v1?label=Weights&message=Huggingface&color=yellow"></a> &ensp; <a href="https://github.com/NVlabs/Sana"><img src="https://img.shields.io/static/v1?label=Code&message=Github&color=blue&logo=github"></a> &ensp; <a href="https://nvlabs.github.io/Sana/Sana-1.5/"><img src="https://img.shields.io/static/v1?label=Project&message=Github&color=blue&logo=github-pages"></a> &ensp; <!-- <a href="https://hanlab.mit.edu/projects/sana/"><img src="https://img.shields.io/static/v1?label=Page&message=MIT&color=darkred&logo=github-pages"></a> &ensp; --> <a href="https://arxiv.org/abs/2501.18427"><img src="https://img.shields.io/static/v1?label=Arxiv&message=SANA-1.5&color=red&logo=arxiv"></a> &ensp; <a href="https://nv-sana.mit.edu/"><img src="https://img.shields.io/static/v1?label=Demo&message=MIT&color=yellow"></a> &ensp; <a href="https://discord.gg/rde6eaE5Ta"><img src="https://img.shields.io/static/v1?label=Discuss&message=Discord&color=purple&logo=discord"></a> &ensp; </div>

🐱 Sana Model Card

Model

<p align="center" border-raduis="10px"> <img src="https://nvlabs.github.io/Sana/Sana-1.5/asset/content/pipeline.png" width="80%" alt="teaser_page1"/> </p>

We introduce SANA-1.5，an efficient model with scaling of training-time and inference time techniques. SANA-1.5 delivers: efficient model growth from 1.6B Sana-1.0 model to 4.8B, achieving similar or better performance than training from scratch and saving 60% training cost; efficient model depth pruning, slimming any model size as you want; powerful VLM selection based inference scaling, smaller model+inference scaling > larger model; Top-notch GenEval & DPGBench results. Detailed results are shown in the below table.

<p align="center" border-raduis="10px"> <img src="https://nvlabs.github.io/Sana/Sana-1.5/asset/content/geneval_comparison.png" alt="model growth performance on GenEval" class="inserted-image" style="max-width: 45%; height: auto; display: inline-block;"> <img src="https://nvlabs.github.io/Sana/Sana-1.5/asset/content/optimizer_loss_comparison_with_ema.png" alt="8-bit optimizer" class="inserted-image" style="max-width: 45%; height: auto; display: inline-block;"> </p>

Source code is available at https://github.com/NVlabs/Sana.

Model Description

• Developed by: NVIDIA, Sana
• Model type: Scalable Linear-Diffusion-Transformer-based text-to-image generative model
• Model size: 4.8B parameters
• Model precision: torch.bfloat16 (BF16)
• Model resolution: This model is developed to generate 1024px based images with multi-scale heigh and width.
• License: Apache License 2.0. Additional Information: Gemma Terms of Use | Google AI for Developers for Gemma-2-2B-IT, Gemma Prohibited Use Policy | Google AI for Developers.
• Model Description: This is a model that can be used to generate and modify images based on text prompts. It is a Linear Diffusion Transformer that uses one fixed, pretrained text encoders (Gemma2-2B-IT) and one 32x spatial-compressed latent feature encoder (DC-AE).
• Resources for more information: Check out our GitHub Repository and the SANA-1.5 report on arXiv.

Model Sources

For research purposes, we recommend our generative-models Github repository (https://github.com/NVlabs/Sana), which is more suitable for both training and inference and for which most advanced diffusion sampler like Flow-DPM-Solver is integrated. MIT Han-Lab provides free Sana inference.

• Repository: ttps://github.com/NVlabs/Sana
• Demo: https://nv-sana.mit.edu/

🧨 Diffusers

Developing

Uses

Direct Use

The model is intended for research purposes only. Possible research areas and tasks include

• Generation of artworks and use in design and other artistic processes.

• Applications in educational or creative tools.

• Research on generative models.

• Safe deployment of models which have the potential to generate harmful content.

• Probing and understanding the limitations and biases of generative models.

Excluded uses are described below.

Out-of-Scope Use

The model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.

Limitations and Bias

Limitations

• The model does not achieve perfect photorealism
• The model cannot render complex legible text
• fingers, .etc in general may not be generated properly.
• The autoencoding part of the model is lossy.

Bias

While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases.