What is Phenaki?

Phenaki is an AI mannequin to generate movies that may be a number of minutes lengthy straight from textual content. It’s also possible to generate video from a nonetheless picture and a immediate. The proposed video encoder-decoder outperforms all per-frame baselines presently used within the literature when it comes to spatio-temporal high quality and variety of tokens per video. To generate video tokens from textual content, they’re utilizing bidirectional masked transformer conditioned on pre-computed textual content tokens. The generated video tokens are subsequently de-tokenized to create the precise video.