n2nmn
n2nmn copied to clipboard
ronghanghu
Code release for Hu et al. Learning to Reason: End-to-End Module Networks for Visual Question Answering. in ICCV, 2017
Hello! I am trying to generate ground truth layouts for the TDIUC dataset (which is very similar to the VQA and VQA 2.0 datasets) and does not contain functional programs....
Hi, Interesting work! I want to test the approach for another dataset(MNIST-dialog). Could you please let me know what would be a good way to go about it? Thanks!
You have mentioned "Install Python 3" in the first step, but you are running all scripts using python2.7. When I do the third step i.e.. install tensorflow fold using pip,...
Hello Ronghang, Thank you for the great paper. I am trying to understand your implementation but could not figure out what does the "average acc with decay" mean in your...
We tried running your code but ended up with cuda compatibility issues. Can you please share any info regarding this.
Your code performs a somewhat unusual L2 normalization across channels in the forward pass ([here](https://github.com/ronghanghu/n2nmn/blob/master/models_shapes/nmn3_modules.py#L60)), which is not mentioned in the paper, I think. Can you please share your motivation...
Hi, I run your code on VQA dataset and get two json results，but I don't know how to evaluate the results?could you give me some advice ? Thank for your...
Metadata
Owner
Metadata
Code release for Hu et al. Learning to Reason: End-to-End Module Networks for Visual Question Answering. in ICCV, 2017