CV fine-tuning: dataset prep
In this exercise, you will prepare the Stanford Cars dataset for training. This will involve using the datasets
library to split the dataset and applying the preprocessing transformations. The dataset consists of 8k labeled images of 196 car models:
The dataset has been loaded as dataset
. The transformations have been defined for you as transforms
, and consist of renormalization and type conversion.
Diese Übung ist Teil des Kurses
Multi-Modal Models with Hugging Face
Anleitung zur Übung
- Create an 80/20 train/test split from
dataset
using the.train_test_split()
method. - Apply the transformations (
transforms
) todata_splits
. - Plot the augmented image from the first set of pixel values in
dataset_transformed
.
Interaktive Übung
Vervollständige den Beispielcode, um diese Übung erfolgreich abzuschließen.
# Create a train/test split within the HF dataset
data_splits = ____(test_size=____, seed=42)
# Apply the transformations
dataset_transformed = ____
# Plot the transformed image
plt.imshow(dataset_transformed["train"][0]["____"].permute(1, 2, 0))
plt.show()