Lightweight inference library for ONNX files, written in C++. It can run Stable Diffusion XL 1.0 on a RPI Zero 2 (or in 298MB of RAM) but also Mistral 7B on desktops and servers. ARM, x86, WASM, RISC-V supported. Accelerated by XNNPACK.
Please consider implementing text-to-video using onnxstream. Text2Video-Zero does not run out of the box on CPU but can be made to run on CPU by changing 4 files and installing the appropriate version(and not the pinned version) of python dependencies. On my machine image generation took around 4GB RAM whereas video generation took around 6GB RAM(both case had only CPU, no GPU present). So, I guess you will be able to fit video generation in ~450MB on RaspberryPi.
Please consider implementing text-to-video using onnxstream. Text2Video-Zero does not run out of the box on CPU but can be made to run on CPU by changing 4 files and installing the appropriate version(and not the pinned version) of python dependencies. On my machine image generation took around 4GB RAM whereas video generation took around 6GB RAM(both case had only CPU, no GPU present). So, I guess you will be able to fit video generation in ~450MB on RaspberryPi.