HunyuanDiT V1.1 Diffusers Distilled is a text-to-image diffusion model published on Hugging Face by Tencent Hunyuan. It provides a distilled variant of the HunyuanDiT v1.1 architecture packaged for compatibility with the Diffusers library.
The model card supplies ready-to-run code that loads the pipeline from the repository using PyTorch and the DiffusionPipeline class. Example inference creates an image from a text prompt such as "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" after setting the data type to bfloat16 and mapping the model to a CUDA device. The same card lists instructions for switching the device to MPS for Apple hardware.
It is distributed as a set of Safetensors files under the tencent-hunyuan-community license. The repository also links to local applications that can consume the model, specifically Draw Things and DiffusionBee. Notebooks for Google Colab and Kaggle are referenced for further experimentation.
The model belongs to the class of text-to-image diffusion transformers and is accompanied by the paper arxiv:2405.08748. It is made available for download and local execution without stated usage fees.
In the Image generation space, HunyuanDiT V1.1 Diffusers Distilled takes a focused approach. It focuses on generating high-quality images from text prompts using a distilled, efficient diffusion transformer model. HunyuanDiT V1.1 Diffusers Distilled is an open-source project aimed at developers. HunyuanDiT V1.1 Diffusers Distilled is open source under the Open Source license. It runs on the web, the command line, and API, and it can be self-hosted.
Tencent Hunyuan builds and maintains HunyuanDiT V1.1 Diffusers Distilled, and it first shipped in 2024. The project is developed in the open on GitHub with 4.3k stars. Among its 3 catalogued features are text-to-Image, Diffusers Pipeline, and multi-Resolution.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do