AI Briefing#diffusers
The latest AI and developer news about #diffusers, with the original source and a short summary.
The latest AI and developer news about #diffusers, with the original source and a short summary.
The new model features significant improvements over the Preview version and includes a detailed model card with training and architecture explanations.
Transformers, Diffusers, and Kernels now run natively on TPUs using the TorchTPU backend, eliminating the need for torch_xla or JAX.
The new model supports text-to-image, multi-reference editing, and bounding-box layout control, with distilled checkpoints generating 1024x1024 images in ~2 seconds on H100 GPUs.
The 7B model generates 2K images and supports natural-language edits using an accelerated 8-step denoising schedule.
The new models include a 29B Pro version and a 3B Lite version, with immediate support for ComfyUI and Diffusers.
A Diffusers PR and demo release suggest that Minimax Music 3 may soon be released with open weights.
Hugging Face Diffusers has officially integrated Nunchaku W4A4 quantization based on SVDQuant, cutting VRAM usage in half and improving inference speed.
Modular Diffusers has been unveiled, allowing diffusion model pipelines to be assembled from reusable blocks.
Black Forest Labs' new image generation model FLUX.2 has been released and is available in the Diffusers library.
A method using Diffusers and PEFT was released that improves LoRA inference speed for Flux models by about 2.3x.
This introduces a method for efficiently fine-tuning the FLUX.1-dev model with under 10GB of VRAM using the diffusers library and QLoRA.
This analyzes the performance and memory efficiency of the various quantization backends supported by Hugging Face Diffusers using the Flux model.
It supports offloading decoding remotely to reduce the memory burden of the VAE decoder during high-resolution image/video generation.
The Diffusers team has summarized the technical status of open video generation models and their future support plans.
Hugging Face's Diffusers library officially supports the Stable Diffusion 3.5 Large model.
Hugging Face's Diffusers library now officially supports the Stable Diffusion 3 (SD3) model.
aMUSEd, an efficient non-diffusion text-to-image generation model that recreates Google's MUSE, has been released.
Hugging Face has released a new script that maximizes SDXL Dreambooth LoRA training performance.
It introduces how to apply fp16 and use memory-efficient attention to speed up inference and reduce memory usage for the SDXL model.
T2I-Adapter, which is much lighter and faster than ControlNet while maintaining SDXL's control performance, has been introduced into diffusers.
AudioLDM 2's inference speed has been improved by more than 10x through Hugging Face Diffusers.
Marking the first anniversary of its release, Hugging Face's Diffusers library summarized its major feature updates and development over the past year.
This introduces a method for fine-tuning Stable Diffusion models by leveraging the AMX accelerator in Intel Sapphire Rapids CPUs.
Hugging Face's diffusers library provides a step-by-step guide for training custom ControlNet models.