Latency Estimate

Problem

Estimate per-token latency in milliseconds: total_ms / num_tokens.

The signature is latency_ms_per_token(total_ms, num_tokens).

Hint

Per-token latency is what matters for streaming generation.

case 1

in: [100,50]

expected: 2

case 2

in: [1,8]

expected: 0.125

case 3

in: [250,1000]

expected: 0.25

case 4

in: [5,4]

expected: 1.25

Starter codePython

def latency_ms_per_token(total_ms, num_tokens):
    # Your code here
    pass

4 test cases · run your code against them in the editor