Quiz · 4 questions

⚡ MLX Core — Installation Through nn.Module

Arrays, lazy evaluation, function transforms, and a working tiny network

Level 0Curious

0 XP0/51 lessons0/15 achievements

0/100 XP to next level100 XP to go0% complete

Quiz

01What's the most reliable single command to confirm your Python interpreter is genuinely running as native ARM (not under Rosetta x86 emulation)?

Hint

The check has to run inside the Python you're worried about.

02Which two divergences between MLX and NumPy are flagged in lesson 3 as the only ones worth pre-checking?

Hint

One is about when things compute; the other is about argument shape.

03In an MLX training loop, why is mx.eval(model.parameters(), optimizer.state) called at the end of every step?

Hint

Think about what happens to the lazy graph if you never force evaluation.

04Lesson 7 measures mx.compile giving roughly a 3x speedup on a 20-op tanh chain. On which kind of function would you expect mx.compile to give the LEAST benefit?

Hint

What does amortize require?

← Previousmx.compile — Kernel Fusion You Can Actually Measure 🤖 mlx-lm — Running Language Models on Your Mac →mlx-lm in Five Minutes — Load, Prompt, Done

Spotted a bug or have feedback on this page?Report an Issue

Comments 0

🔔 Reply notifications (sign in)

No comments yet — be the first.