News for hackers. Stuff that matters.
Re:Show HN: Text-to-video model from scratch (2 brothers, 2 ... by schopra909 on Friday January 23, 2026 @09:49AM (#46733136)
That all being said, you can just delete the T5 from memory after encoding the text so save on memory.

The 2B parameters will take up 4 Gb of memory but activations will be a lot more given size of context windows for video.

A 720p 5 second video is roughly 100K tokens of context

There are no comments on this comment yet.