See our collection for all versions of Moonshine.

Run Moonshine with Keras 3: JAX, PyTorch, or TensorFlow

kerasformers/moonshine_tiny

Paper: Moonshine: Speech Recognition for Live Transcription and Voice Commands (arXiv:2410.15608) · HF Papers

Moonshine is an English ASR encoder-decoder built for short / live audio: the encoder sees the raw waveform length you pass in (no Whisper-style 30 s pad), so short commands stay cheap. Output is cased and punctuated.

For more details on the model, please go to the upstream model card.

Pure-Keras 3 conversion of UsefulSensors/moonshine-tiny for kerasformers. One implementation runs unmodified on TensorFlow / Torch / JAX.

This is an ASR checkpoint (MoonshineSpeechToText).

✨ Quick start

import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

import soundfile as sf
from kerasformers.models.moonshine import (
    MoonshineProcessor,
    MoonshineSpeechToText,
)

model = MoonshineSpeechToText.from_weights("kerasformers/moonshine_tiny")
processor = MoonshineProcessor.from_weights("kerasformers/moonshine_tiny")

audio, sr = sf.read("your_audio.wav", dtype="float32")  # 16 kHz mono
# Cost scales with clip length: no fixed 30 s pad like Whisper.
text = model.generate(audio, processor)
print(repr(text[0]))

Load any Moonshine variant the same way with from_weights("kerasformers/<variant>"):

Variant	Hub
`moonshine_tiny`	`kerasformers/moonshine_tiny`
`moonshine_base`	`kerasformers/moonshine_base`

Tips

Set KERAS_BACKEND before importing Keras / kerasformers.
Prefer MoonshineProcessor.from_weights(...) so feature extraction matches.
English-only; pass a list of waveforms to batch.
See Moonshine docs and Loading Weights.
Community / upstream safetensors still work via the hf: prefix, e.g. MoonshineSpeechToText.from_weights("hf:UsefulSensors/moonshine-tiny").