search

MILVLG / openvqa

A lightweight, scalable, and general framework for visual question answering research
Apache License 2.0
318 stars 64 forks source link

Supplement for CLEVR #57

Closed nbgao closed 4 years ago

nbgao commented 4 years ago
  1. Add clevr_extract_feat.py to extract image features from origin image files and transform .h5 files into multiple .npz files.
  2. Add CLEVR part of install.md.