guidance for developing a LSTM model

declare-lab / MELD

MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversation

GNU General Public License v3.0

788 stars 200 forks source link

guidance for developing a LSTM model #8

Closed Thani011 closed 5 years ago

Thani011 commented 5 years ago

Hi, I am struggling to find any support for developing a model for a multi-model dataset . Can you please guide or give me some reference for using this dataset and developing a LSTM or CNN model. I am new to this field but for now i am able to develop model for images and text separately but having trouble in using a merged input (image+text or image+audio ).

Please provide some direction or explain with respect to the baseline model you provided. Thanks

soujanyaporia commented 5 years ago

Bimodal features have been provided in the repo. Please download and load using -

import pickle train_bimodal_emb, val_bimodal_emb, test_bimodal_emb = pickle.load(open(filepath, 'rb'))