search

weijiawu / TransDETR

[IJCV 2024] TransDETR: End-to-end Video Text Spotting with Transformer
101 stars 11 forks source link

not Set FP track Instances' obj_idx as -2 #12

Open lsabrinax opened 2 years ago

lsabrinax commented 2 years ago

https://github.com/weijiawu/TransDETR/blob/b6c73c5fd773e5e0d87f394631cfcab6e167b070/models/qim.py#L109