I have tried Emoji prediction on the following dataset:
The link to my code is: https://colab.research.google.com/drive/1LjCJWORIK_6CycY7S3Rjjq-9vJZVCi2Y#scrollTo=-uf7vHZFNeZJ
Even the training accuracy is very low(around 30%). I have tried changing the number of units in the LSTM layer but cannot figure out any improvements. Any suggestions would be helpful.
Please check. Thanks