10,000 USD₮
DescriptionThe LTX-2 family of models (specifically LTX 2.3) is a state-of-the-art open-weights video generation model (14B-parameter DiT, Gemma 3 text encoder, spatiotemporal Video-VAE) supporting text-to-video (T2V) and image-to-video (I2V). Its official stack is Python/PyTorch-only, making it impractical for edge and embedded use.
qvac-ext-stable-diffusion.cpp already provides ggml-based inference for image diffusion models (SD, Flux, Wan, etc.) across CPU, Vulkan, and Metal, but does not yet support any video diffusion. This grant funds the addition of LTX-2 T2V and I2V support to that fork, plus a Bare runtime addon to expose video generation to JavaScript applications in the QVAC ecosystem.
qvac-ext-stable-diffusion.cpp already provides ggml-based inference for image diffusion models (SD, Flux, Wan, etc.) across CPU, Vulkan, and Metal, but does not yet support any video diffusion. This grant funds the addition of LTX-2 T2V and I2V support to that fork, plus a Bare runtime addon to expose video generation to JavaScript applications in the QVAC ecosystem.