shivam-sood / memnet_wk_vqa Goto Github PK

Below are the folders/files and their description that are included in the zip file

Folder : Code
-------------
CustomQATransformers.ipynb  - Code for new approach with transformers for the visual question answering
memnetParaphrasedQNoWikicapV1.ipynb  - Code for memnet implementation with paraphrased questions and without wiki captions
memnetParaphrasedQV1.ipynb  - Code for memnet implementation with paraphrased questions and with wiki captions
memnetWithoutWikicapV1.ipynb - Code for memnet implementation with original questions and without wiki captions
memnetWithWikicapV1.ipynb  - Code for memnet implementation with original questions and with wiki captions
faceRecogCV.ipynb  - Code for face recognition for visual entity linking

Folder : Reports
-----------------
1. accuracynw.png - Accuracy of model with original quesions and without wikicaptions
2. accuracyp.png  - Accuracy of model with paraphrased questions and with wikicaptions
3. accuracypnw.png - Accuracy of model with paraphrased question and without wikicaptions
4. accuracyw.png - Accuracy of model with original questions and with wikicaptions
5. lossnw.png - Loss of model with original quesions and without wikicaptions
6. lossp.png - Loss of model with paraphrased questions and with wikicaptions
7. losspnw.png - Loss of model with paraphrased question and without wikicaptions
8. lossw.png - Loss of model with original questions and with wikicaptions


Folder : Models
---------------
Models are saved in google drive. Here is the link for the saved models

https://drive.google.com/file/d/1-DXkwFR7T_whabnwXq1nfyAOAXIK2TQ1/view?usp=share_link
https://drive.google.com/file/d/1-8t6F697oLURx-ip-L1zQegyMTVnax2n/view?usp=sharing