Question 1: Short answer
hardAI-checkedDeep Learning Frameworksimport jax
import jax.numpy as jnp
calls = []
@jax.jit
def f(x):
calls.append(1) # a Python side effect
return 2 * x
for v in [1.0, 2.0, 3.0]:
f(jnp.array(v))
f(jnp.arange(3.0))
f(jnp.arange(4.0))
f(jnp.arange(4.0) + 1)
print(len(calls))
What does the last line print?