Dataset

We will use Hugging Face Dataset to demonstrate Text Classification. You can checkout this course for more details on how to use Hugging Face.

For this course, we are using “emotion” dataset from Hugging Face

from datasets import load_dataset
emotion = load_dataset("emotion")

The emotions object is an example of the DatasetDict class, functioning akin to a Python dictionary, and it features split keys.
To extract different split of data, we can use following code

train_ds = emotions["train"]
validation_ds = emotions["validation"]
test_ds = emotions["test"]

Each object in the above split is instance of class Dataset
To find the column Name in the above dataset, we can use following code

train_ds.column_names

Although Dataset provide a lot of APIs for data manipulation, but due to familiarity with Pandas it is a standard way to convert Dataset into Dataframe

import pandas as pd
emotions.set_format(type="pandas")
df = emotions["train"][:]