• A

• B

• C

• D

• E

What is the purpose of setting the model to evaluation mode with model.eval() in PyTorch?

• To initialize the model parameters.

• To disable gradient computation during training.

• To ensure layers like dropout and batch normalization behave correctly during inference.

• To load the pre-trained weigths of the model.

Which of the following code snippets correctly initializes a ResNet50 model without pre-trained weights in PyTorch?

• 
• 
• 
• 

What does the following code snippet do in the context of PyTorch model inferencing?

• It disables gradient computation and performs inference.

• It initializes the model parameters without gradient computation.

• It enables gradient computation and performs inference.

• It computes gradients and performs inference.

What is the primary advantage of using model parallelism in large language model inference?

• Reducing the inference time by processing multiple inputs in parallel.

• Distributing the model's computations across multiple GPUs to handle large models that cannot fit into the memory of a single GPU.

• Increasing the precision of the model's predictions.

• Improving the training speed of the model.

What is the purpose of the torch.no_grad() context in the inference of large language models?

• To enable gradient computation for backpropagation.

• To reduce memory usage by disabling gradient computation.

• To speed up the training process.

• To initialize model weights.

When using a pre-trained large language model for inference, what is the recommended way to handle tokenization?

• Implement a custom tokenization algorithm.

• Use the tokenization method provided by the model's pre-trained package.

• Use a generic tokenization method.

• Use a generic tokenization method.

How can you generate more diverse outputs from a large language model during inference?

• By setting a higher batch size.

• By lowering the learning rate.

• By using techniques like top-k sampling or nucleus sampling.

• By using gradient clipping.

Which of the following methods is used to reduce the computational load during the inference of very large language models?

• Increasing the sequence length.

• Using model quantization techniques.

• Increasing the number of layers in the model.

• Disabling layer normalization.

What is the impact of setting the temperature parameter to 0 during the inference of a language model?

• It generates completely random outputs.

• It makes the model outputs more deterministic and less diverse.

• It increases the training speed.

• It reduces the size of the output sequence.