Ahmadzei's picture
added 3 more tables for large emb model
5fa1a76
The model obtains state-of-the-art results on
image captioning and visual question answering benchmarks.