Obtaining visual features for the ConceptualCaptions data set

jackroos / VL-BERT

Code for ICLR 2020 paper "VL-BERT: Pre-training of Generic Visual-Linguistic Representations".

MIT License

735 stars 110 forks source link

Obtaining visual features for the ConceptualCaptions data set #50

Closed jules-samaran closed 4 years ago

jules-samaran commented 4 years ago

Hi, I'd like to take a look at the pretraining but I was wondering if there was a way to have access to the train_frcnn.zip files without having to run the caffe model, are they available somewhere? Best,

jackroos commented 4 years ago

Sorry about that. The files are too big to upload to online disks like Google Drive. You need to run the caffe code to extract the visual features.

yonatanbitton commented 3 years ago

+1 - I need this as well, and having problems with the caffe model. Opening an issue for it. @jules-samaran were you manage to extract the features?