Instructions to use UsefulSensors/moonshine-streaming-small with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use UsefulSensors/moonshine-streaming-small with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="UsefulSensors/moonshine-streaming-small")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("UsefulSensors/moonshine-streaming-small", dtype="auto") - Notebooks
- Google Colab
- Kaggle
GGUF + pure-C++ runtime in CrispASR — Moonshine streaming small
#2
by cstr - opened
The streaming Moonshine variants run through CrispASR's moonshine-streaming backend (src/moonshine_streaming.cpp) — a separate runtime from offline Moonshine because the encoder is sliding-window + raw-waveform.
Auto-detected from GGUF metadata; same backend handles tiny (34M), small (110M), and medium (~245M). KV-cached autoregressive decoder, companion tokenizer fetched alongside the GGUF.
Pre-quantised GGUFs (MIT): cstr/moonshine-streaming-small-GGUF
./build/bin/crispasr --backend moonshine-streaming \
-m moonshine-streaming-small-q4_k.gguf --mic
Sibling sizes: -tiny, -medium. Offline siblings at cstr/moonshine-{tiny,base}-GGUF + 6 language variants.