VDN-H3
PulseAugur coverage of VDN-H3 — every cluster mentioning VDN-H3 across labs, papers, and developer communities, ranked by signal.
-
H3 Acceleration Arena adds VDN-H3, TaoMate H3, and LightX2V 1.2 models
The H3 Acceleration Arena has expanded its model evaluations with the addition of three new models: VDN-H3, TaoMate H3, and LightX2V 1.2. Users are encouraged to vote on these new models to assess their performance agai…
-
MiniMax H3 video generation achieves 2x real-time speeds
MiniMax AI has announced significant advancements in its video generation model, MiniMax H3. Leveraging the @sgl_project and VDN-H3, the model now achieves over twice real-time denoising speeds. This allows for the gene…
-
MiniMax H3 video generation achieves 2x speedup on Nvidia B200
MiniMax AI has announced significant speed improvements for its H3 video generation model. Utilizing SGLang-Diffusion and VDN-H3 on eight Nvidia B200 GPUs, the system can now generate 14.4 seconds of 768p video in just …
-
VDN-H3 optimized for 24GB VRAM cards released on GitHub
A user has developed a ComfyUI implementation of VDN-H3, specifically optimized for graphics cards with 24GB of VRAM, such as the RTX 3090 Ti. This implementation utilizes INT8 ConvRot VDN weights and has been tested fo…
-
Video DeltaNet accelerates video generation with hybrid attention
Researchers have introduced Video DeltaNet (VDN-H3), a hybrid-attention model designed to accelerate video generation while maintaining high quality. The model utilizes a combination of efficient linear attention and a …