add new dataset kan_hope

[adeepH]

Adding a Dataset

• Name: KanHope
• Description: A code-mixed English-Kannada dataset for Hope speech detection
• Task: Binary Text Classification
• Paper: https://arxiv.org/abs/2108.04616
• Data: https://github.com/adeepH/kan_hope/tree/main/dataset
• Motivation: The dataset is amongst the very few resources available for code-mixed low-resourced Dravidian languages of India

[lhoestq]

Thanks for adding this dataset !
And good job with the dataset script and dataset card

My only concern is that I think it's better if we have the labels in the other order, i.e. "Not-Hope" (0), "Hope" (1). I'm doing the change if you don't mind