ML TensorsWhat is a tensor? A multi-dimensional array that generalizes scalars, vectors, and matrices. What is a 0D tensor called? A scalar (e.g. 5) What is a 1D tensor called? A vector (e.g. [1, 2, 3]) What is a 2D tensor called? A matrix (e.g. [[1, 2],

Réussis tes devoirs et examens dès maintenant avec Quizwiz!

What does a shape [16, 3, 5, 5] mean in a CNN?

16 filters, 3 input channels (RGB), each with a 5×5 kernel

What does the shape [3, 4] typically represent in a dense layer?

3 input features mapped to 4 output neurons (weight matrix)

What is a 2D tensor called?

A matrix (e.g. [[1, 2], [3, 4]])

What is a tensor?

A multi-dimensional array that generalizes scalars, vectors, and matrices.

What is a 0D tensor called?

A scalar (e.g. 5)

What is a 1D tensor called?

A vector (e.g. [1, 2, 3])

In deep learning, what type of data is usually stored as tensors?

Inputs, weights, biases, outputs, and gradients.

Are all tensors matrices?

No — only 2D tensors are matrices.

Why are tensors used in ML libraries like PyTorch or TensorFlow?

They're efficient for GPU computation and support automatic differentiation.

Are all matrices tensors?

Yes — all matrices are 2D tensors.

What is the shape of a tensor representing 2 samples, each with 3 features?

[2, 3] — a 2×3 matrix

What is the result of multiplying a tensor of shape [2, 3] by [3, 4]?

[2, 4] — 2 samples, 4 outputs each

What shape does a batch of 8 RGB images of size 32×32 have?

[8, 3, 32, 32] — batch size, channels, height, width

What is the shape of a 4D weight tensor in a convolutional layer?

[num_filters, in_channels, kernel_height, kernel_width]

What function multiplies two tensors (matrix multiplication)?

np.dot() in NumPy or torch.matmul() in PyTorch


Ensembles d'études connexes

Chapter 10 - Global Strategy: Competing Around the world

View Set

lacharity chapter 7 exam questions for management exam 1

View Set

Labeling the contractile unit of the muscle

View Set

ОГЭ Part 2 Travel, Traditions, Holidays

View Set

Chapter 02 Quiz - Introduction to the VDV Industry

View Set