In this model I have used two datasets from kaggle, one is a collection of random tweets and one is just depressive tweets. I'm pasting the drive link for them both below. I have mounted them from my drive in this notebook to work with them, you can download and use it however you want.
Random tweets: https://drive.google.com/file/d/1ISWSL2nngKGEAak9309tGDlgPaXDgUbo/view?usp=share_link Depressive tweets: https://drive.google.com/file/d/1_sO-8MX3ujgE0YrQoDFdGCmdcgDh0eQC/view?usp=share_link