Ostris(@ostrisai):I removed all the critics from my turbo time continuous training method and switched to using the model itself as an evolving semi frozen critic. I also switched to Video Sparse Attention for a speed boost. This is the result from training over night. These are all 1 step.

Ostris

@ostrisai

AI / ML researcher and developer. Creator of AI Toolkit - Support my work -

加入 August 2023

338 正在关注 15K 粉丝

Ostris@ostrisai

2026.08.31 13:20

I removed all the critics from my turbo time continuous training method and switched to using the model itself as an evolving semi frozen critic. I also switched to Video Sparse Attention for a speed boost. This is the result from training over night. These are all 1 step.

显示更多