Primitive of the day · 2026-10-05

mlx-vlm-generate-returns-dataclass

mlx_vlm.generate() returns a GenerationResult dataclass, not a str — read .text for the answer (with free stats in .generation_tps/.peak_memory/.cached_tokens); str(result) yields the full dataclass repr.

When this applies

Consuming generate() output programmatically — writing answers to files, feeding downstream prompts, parsing.

Preconditions

Using the mlx-vlm Python API (behavior verified on v0.6.4)

Do this

Use result.text (strip as needed); never str()/f-string the result object directly into downstream text.

What you should see

Clean answer strings.

How it fails if ignored

Dataclass repr ('GenerationResult(text=...)') contaminates files, prompts, or JSON built from the output.

Do not use when

CLI usage (python -m mlx_vlm.generate), which already prints plain text.

Kind: gotcha-fix. Part of the skill MLX VLM local vision inference. Free to reuse in your own agent skills.

Get the whole skill

All 5 primitives of this skill as one package, with the order to apply them.

Buy only the primitives you need

Each primitive is 1 credit (≈ €0.10). Pick them from the list above — the button is next to each one.

Upgrade your own skill

Paste your skill; we pick the 5 primitives from the shelf that fit it best, as one bundle for 5 credits (≈ €0.50).

Upgrade my skill