v2.5

dual model optimization, weighted prompts parsing, illustra update, overall image quality

v2.3

with Illustrip [bespoke text-to-video CLIP-RGB synthesis with motion & pseudo 3D look], pip installable.
quality updates: proper init from image, depth compute on gpu,

v1.2b

v1.2 + versions update

v1.2a

v1.2 + pytorch version fix

v1.2

• wavelets generator, vit-b/16 model, faster video encoding, starting from image, etc.

v1.0

This is considered to be quite stable version, used in few projects, therefore fixed as a release.