Prepare datasets for fine tuning MT, STS, NLI and Classification

[zeynepyirmibes]

We should prepare the scripts and datasets for:

• Translation
• Semantic Textual Similarity
• Natural Language Inference
• Text Classification

[zeynepyirmibes]

STS and NLI scripts have been added with PR #12.

Train validation & test sizes are recorded in this spreadsheet. , and baselines in the Baseline sheet.

[gokceuludogan]

Closing the issue as the classification datasets have been included in PR #27