[feature/improvements] List of metrics to implement for NLP

[NISH1001]

Following are the metrics we need to implement based on jury

• Bleu score
• sequence evaluation (NER)
• ROUGE
• Meteor
• Perplexity

All these metrics could be derived through evalem.metrics._base.JuryBasedMetric (see evalem.metrics.basics module for some general implementation)

---

cc: @muthukumaranR @xhagrg @code-geek