All committers have signed the CLA.

@lalitpagaria PR is complete from my side, one possible improvement that can be added. is mapping of labels, transformers map labels as {"LABEL_0", "LABEL_1",....,"LABEL_N"}, what we can do is use the labels parameter and map values from labels to transformers labels, however, if the user passes the wrong sequence of labels, they'd get altogether wrong predictions. So there must be some mechanism for enforcing the right sequence in my knowledge it won't be possible to enforce such a check. what do you think?

@shahrukhx01 I have refactored you changes a little. Also added functionality to support user provided label map. Check test and let me know if you have any concern. Post that I will merge this.
Also it would be great if we plan a demo/article/colab showing capabilities of https://huggingface.co/obsei-ai/sell-buy-intent-classifier-bert-mini

Or we can show demo of newly added classifier to detect hate speech via https://huggingface.co/Hate-speech-CNERG/dehatebert-mono-english
So workflow would be Twitter or Reddit -> Analyzer -> Elastic Search or Slack

WDYT?

@lalitpagaria I can write a small collab tutorial using Obsei and then post-it on LinkedIn about the buy-sell intent model.

@shahrukhx01 We can add tutorial here https://github.com/obsei/obsei#tutorials and article here https://github.com/obsei/obsei#articles