Open JunzheJosephZhu opened 8 months ago
yes - see the colab. it's released there :)
The alignment fine tuning subset seems to be similar in scale compared to objaverse 1.0. Is this a superset of objaverse 1.0? On average, does the alignment fine tuning dataset have comparable quality to objaverse 1.0? @mattdeitke
@mattdeitke Btw, could you tell us how is alignment subset selected? Is it handpicked?
Also, is there overlap between objaverse and the alignment fine-tuning subset of objaverse-xl?
Alignment Fine-tuning Annotations
Is this just a subset of objaverse-xl that's ensured to be aligned with text description?