Dataset: mnist
Model: basiccnn
Number of classes: 10
Train samples: 42000
Validation samples: 18000
Test samples: 10000
Validation accuracy: 0.9895
Test accuracy: 0.9901
Github repository: ....

Data Loading Instructions:
The extracted features, predictions, targets, and logits are saved in .npz files.
To load the data in Python:

import numpy as np

# Load training data
train_data = np.load('mnist_basiccnn_train_out.npz')
train_features = train_data['features']
train_predictions = train_data['predictions']
train_targets = train_data['targets']
train_logits = train_data['logits']

# Load validation data
val_data = np.load('mnist_basiccnn_val_out.npz')
val_features = val_data['features']
val_predictions = val_data['predictions']
val_targets = val_data['targets']
val_logits = val_data['logits']

# Load test data
test_data = np.load('mnist_basiccnn_test_out.npz')
test_features = test_data['features']
test_predictions = test_data['predictions']
test_targets = test_data['targets']
test_logits = test_data['logits']

Note: features are the output of the feature extractor (before the final classification layer),
predictions are softmax probabilities, targets are the true labels, and logits are the raw model outputs.