ACM AI Pytorch Review

Given the following code, what's the shape of the tensor produced?

• [3, 5]

• [3, 3]

• [5, 5]

• [5, 1]

What does this code produce?

• [4, 5, 6]

• [1, 2, 1]

• [4, 5, 1]

• [1, 2, 3]

Is this nn.module constructed correctly?

• Yes

• No

Answer explanation

Let's say we have a tensor with shape [5, 12, 32, 64]. What command should we use to create a tensor of shape [32, 5, 64, 12]?

Fill in the blank without spaces or negative indices: A._______(_,_,_,_)

If we have X = [[1, 2, 3], [4, 5, 6]], what does X[-1,:1] yield?

• 4

• 5

• 2

• 3

I have the following training loop, as well as the following 4 operations. Can you help me order them?

What line comes first?

1. 1. optimizer.step()

2. 2. loss = loss_fn(output, target)

3. 3. output = model(data)

4. 4. loss.backward()

• 1

• 2

• 3

• 4

I have the following training loop, as well as the following 4 operations. Can you help me order them?

The first line is output = model(data).

What line comes after?

1. 1. optimizer.step()

2. 2. loss = loss_fn(output, target)

3. 4. loss.backward()

• 1

• 2

• 4

I have the following training loop, as well as the following 4 operations. Can you help me order them?

The code so far is:

output = model(data)

1. loss = loss_fn(output, target)

What line comes after?

1. 1. optimizer.step()

2. 4. loss.backward()

• 4. loss.backward()

• 1. optimizer.step()

• 4. l0ss.backward()

• 4. model.forward()

A = [[2,3],

[1,2]]

B = [[1,0],

[0,0]]

A = torch.tensor(A)

B = torch.tensor(B)

What is torch.matmul(A, B)?

• [[2,3],

  [1,2]]

• [[2,0],

  [0,1]]

• [[2,3],

  [0,0]]

• [[2,0],

  [1,0]]

Which modern model is the largest (in param count?)

• GPT4

• Largest Llama-3.1

• Largest Mixtral

• Stable Diffusion 3

• AlphaGo