search

Cadene / vqa.pytorch

Visual Question Answering in Pytorch
716 stars 177 forks source link

can you report a single model MUTAN + Att performance? #16

Closed gaopeng-eugene closed 6 years ago

gaopeng-eugene commented 6 years ago

In your paper, you report ensemble model. Because you are the state-of-the-art VQA method right now, can you report a single model MUTAN+Att performance? It is easier to compare with.

Cadene commented 6 years ago

You should be able to download a pretrained model (the open ended accuracy is included in the zip).

mkdir -p logs/vqa
cd logs/vqa
wget http://webia.lip6.fr/~cadene/Downloads/vqa.pytorch/logs/vqa/mutan_noatt_train.zip 
wget http://webia.lip6.fr/~cadene/Downloads/vqa.pytorch/logs/vqa/mlb_att_trainval.zip 
wget http://webia.lip6.fr/~cadene/Downloads/vqa.pytorch/logs/vqa/mutan_att_trainval.zip

(The last one)

Our fusion scheme should be state of the art (as far as we know), but other fusion schemes are competitive (MCB, MLB, MFB). However multiple tricks and different features and architectures can be used outside the fusion scheme to improve the accuracy further more (please, look at the CVPR workshop about VQA).

aronsar commented 6 years ago

Hi, it seems like the link is dead now (for mutan_att_trainval), would you mind please posting another one?