Solutions

DL-C16-C-001

Subtract the maximum value:

max([3, 10, 4]) = 10

DL-C16-C-002

3 - 10 = -7

DL-C16-C-003

0.004 + 0.001 = 0.005

DL-C16-C-004

The answer is 1. exp(1000) is an overflow warning in ordinary floating-point computation.

DL-C16-C-005

The answer is 1. NaN should trigger debugging.

DL-C16-C-006

The answer is 1. Dtype affects storage, range, precision, memory, and computation.

DL-C16-C-007

The answer is 1. Mixed precision uses lower precision where useful while keeping stability-sensitive work safe enough.