Can we do few-shot inference or one-shot inference?

ByungKwanLee / MoAI

[ECCV 2024] Official PyTorch implementation code for realizing the technical part of Mixture of All Intelligence (MoAI) to improve performance of numerous zero-shot vision language tasks.

MIT License

311 stars 32 forks source link

Can we do few-shot inference or one-shot inference? #14

Closed ashutoshIITK closed 7 months ago

ashutoshIITK commented 7 months ago

Is it possible to feed one image and caption and then do another inference as in few shot inference?

ByungKwanLee commented 7 months ago

Few shot inferneces should have ability to understand the multiple images (only texts can be possible). However, MoAI is not traoned with mutiple images and does not support multiple image tokens.

ashutoshIITK commented 7 months ago

Got it! Thank you for the quick response!