QW21-Turbo-0.1 by Viggle
Steps 4-16 CFG=1 T2i/i2i
演示图片同时使用了 Viggle/Qwen-Image-2.1-viggle-turbo 和 REDQW21 UNLOCKED LoRA (with NV DLSS_NR)
仅作为[ 研究或评估 ]用途内使用 / only for the research or evaluation purposes
The demo images uses both Viggle/Qwen-Image-2.1-viggle-turbo and REDQW21 UNLOCKED LoRA simultaneously
Built with Qwen, Applicable to Qwen-Image-2.1
Viggle/Qwen-Image-2.1-viggle-turbo source repo:
https://huggingface.co/Viggle/Qwen-Image-2.1-viggle-turbo
All credit belongs to the original model author. License is the same as the original model.
Model: Viggle/Qwen-Image-2.1-viggle-turbo · Built with Qwen — distilled from Qwen/Qwen-Image-2.1, which is released under the Qwen RESEARCH LICENSE AGREEMENT (non-commercial: research or evaluation purposes only).
Built with Qwen. A 4-step distilled student of Qwen/Qwen-Image-2.1, trained by Viggle with Distribution Matching Distillation. It does both text-to-image and instruction-driven editing with 1–3 reference images in 4 transformer passes instead of 40, with no classifier-free guidance.
Two students are shipped — pick one:
transformer/— full fine-tuned transformer (bf16, 14.2 GB). Replaces the base transformer; exact, no adapter. This is what the demo Space runs by default and the one we currently recommend — in our qualitative comparison it edits more faithfully than the LoRA.Qwen-Image-2.1-viggle-turbo-4step-lora-r64.safetensors— LoRA adapter (rank 64, 340 MB) loaded on top of the base transformer at runtime. Smaller download, slightly weaker.
Status: v0.1 preview, work in progress — this release still falls short of the base model. On complicated image editing (multi-reference composition, face swaps, identity-preserving edits, instructions with several constraints) it is clearly worse than the 40-step base model. Text-to-image at 4 steps is usable. We are still working on it and will update this repository as the distillation improves; treat the current weights as a preview, not a replacement for the base model.
Minimax-H3-Turbo by LightX2V
Minimax-H3-Turbo provides batch MiniMax-H3 inference and NFE/LoRA comparisons.
Repos
Kijai/MiniMax-H3_comfy Minimax-H3-Turbo ComfyUI v0.1
lightx2v/Minimax-h3-Turbo Minimax-h3-Turbo Diffusers/ComfyUI v1.0
Note
Its current visual quality can be reviewed in the linked examples, and the image details still need improvement. An enhanced version is in development.
This Edition is compatible with the ComfyUI-pruned
MiniMax H3 FL2VA model（20B）.
Roadmap
Improve the visual details and overall quality of FL2V Turbo.
Develop distillation based on Ref2V.
Preview
https://huggingface.co/lightx2v/Minimax-h3-Turbo/discussions/2
Apache-2.0 license
Description
(165.97 MB)is FsstVideo's Hunyuanvideo LoRA
FastHunyuan LoRA is an Accelerator model. It can sample high quality videos with 6-10 diffusion steps. That brings around 8X speed up compared to the original HunyuanVideo with 50 steps.
FAQ
Comments (11)
do i have to use this with a specific hunyuan model? or the 12gb one works fine?
Sure, 12GB can run
Could you re-train them for wan1.2 14b and 1.3b please?
is this just a lora loader node to import into workflow and it will accelerated ? Is it right?
Did it work in FramePack-Studio?
Hi there, this does not work with framepack studio..
an example workflow would help.