Eduard Zamfir

@eduard__zamfir

CS PhD with Prof. Radu Timofte, JMU Würzburg

following
followers

MiniT2I fine-tune runs twice as fast with half the tokens, a user says

The user says their fine-tuned model uses region tokens instead of treating every image patch as a token in every layer, while maintaining the same quality.

2 Sources, 39d ago, first seen 39d ago

Featured in

MiniT2I fine-tune runs twice as fast with half the tokens, a user says

Featured in

MiniT2I fine-tune runs twice as fast with half the tokens, a user says