davidnvq / grit

GRIT: Faster and Better Image-captioning Transformer (ECCV 2022)
185 stars 28 forks source link

training with my own dataset #50

Open luyuzhe11 opened 11 months ago

luyuzhe11 commented 11 months ago

Hello author, thank you for your excellent work. I want to train my dataset, how do I generate a ‘coco train ids.npy’ file for my own dataset

1702349913604