Moonshine Streaming for anira
Moonshine AI's streaming speech-to-text models, tiny and small, English and German, in ONNX
Runtime's format (.ort), mirrored unchanged for real-time inference with anira. The fast engine
of anira foundation models (speech to text): words appear while someone speaks.
| Folder | Model | Upstream |
|---|---|---|
tiny-en/ |
Moonshine Streaming Tiny, English | moonshine-ai/moonshine-voice-assets @ 0bf2f2e, model/tiny-streaming-en/quantized_26_08_21/ |
small-en/ |
Moonshine Streaming Small, English | same, model/small-streaming-en/quantized_26_08_21/ |
tiny-de/ |
Moonshine Streaming Tiny, German | download.moonshine.ai/model/tiny-streaming-de/quantized_26_08_24/ |
small-de/ |
Moonshine Streaming Small, German | download.moonshine.ai/model/small-streaming-de/quantized_26_08_24/ |
Each folder holds the encoder, adapter, cross-attention and decoder graphs, the audio frontend,
streaming_config.json and the tokenizer.
Files
| File | Size | SHA256 |
|---|---|---|
tiny-en/adapter.ort |
1,319,664 | 22ecc949e146c49667fda28d102d4e30749a107dc88a396292aa8f277ef1347c |
tiny-en/cross_kv.ort |
1,287,544 | 143a36667b8d05fd9d04e8c337b7ee121f37ef299aea6b3d82bdb3d3401950b4 |
tiny-en/decoder_kv.ort |
32,583,720 | 8852553f312adb6c9aa4d17418015049b30f412209ee569d336548c0044627de |
tiny-en/encoder.ort |
7,675,440 | a8414e1a5dedf9f2093d7680601dd8a9b0433e7020260eafe0e370ead91134ca |
tiny-en/frontend.model.ort |
23,344 | 5121b561417b638afce0c6c31b760e37c93cf97f80d9b0031aad1fe7b6f25d61 |
tiny-en/frontend.weights.ort |
2,093,464 | 217da24ac6f522ebf02da8ef288e77d1ac68d50d4a6821433182e4fbf4204bbd |
tiny-en/streaming_config.json |
509 | 74fe5ddebd63b17caf59e8a3b18c17547ff7bce1642050edbb1c3962674f8950 |
tiny-en/tokenizer.bin |
249,974 | 6884b35fd6377d4c4d32336a0bc152f36b64d1e45b6503683cdc238250a8472d |
tiny-de/adapter.ort |
1,318,472 | 23b91142a35cfbc533a7f03ad8387d0b6d32856b465c6fe7a14ede44a254a345 |
tiny-de/cross_kv.ort |
1,288,120 | 25e6d9d4f16569c2ede7e23a381fe7c60cc4b0f141adc88580557d4530634f17 |
tiny-de/decoder_kv.ort |
19,717,336 | 768b1464074fc4fb7327e8b863fc13dcd12cb7bbe8d8e9428fa7292f7fd9ca6e |
tiny-de/encoder.ort |
7,772,792 | 74610b818058e396f4b9d9ffd730cf8e36772122464d989dd3011dfa76bd6baf |
tiny-de/frontend.model.ort |
23,176 | 4445bb4c2f41b0dd993999e06dc9367a0dde0d42a2e777116e7d93be102dc43d |
tiny-de/frontend.weights.ort |
2,093,280 | 233fda61d583403bbf13ea71b6ba2335fdf57eac71be85a97dcbb90a6c2f4186 |
tiny-de/streaming_config.json |
509 | d31afd7c2dd5ee9ae3fa232312c5b42cb3e4fe1c6e9c53d885f2bea63023790a |
tiny-de/tokenizer.bin |
103,319 | 5725ad3771a1339802f8cebeda0fa8f7723f5282b1ee726c2544774d7399b9a7 |
small-en/adapter.ort |
2,870,368 | c665f742364febad597cc9ac1e0b341ffbee0e24a1466e2f3bde95e6e4771762 |
small-en/cross_kv.ort |
5,356,536 | e2d3417144e9514055ebfefe8dcc4c0a55a55adcb8530435844c75c53e352bf6 |
small-en/decoder_kv.ort |
81,878,600 | 1a05465b1dd955858dfcbee039c0020fb5dd982b0f5094c34e61735d518d771b |
small-en/encoder.ort |
44,148,576 | 2d4d973e91e8aca08c51e7e7efa28a46ab265b63d809d5294d18b86bcd85b993 |
small-en/frontend.model.ort |
26,944 | 09b1210ae30dc5f0f3e45f0ebab914c254741323114f53fbbe5ae62cca35058f |
small-en/frontend.weights.ort |
7,769,464 | 7ef97521bd4bad3928f5bb6808586f4fcc6e92bd5990394112eed7d4052ec338 |
small-en/streaming_config.json |
512 | 26f02b6afb22d60871a5efd85c3d38e569cc0ddb6c5eb6e93d3260152ae8a47a |
small-en/tokenizer.bin |
249,974 | 6884b35fd6377d4c4d32336a0bc152f36b64d1e45b6503683cdc238250a8472d |
small-de/adapter.ort |
2,869,296 | be438a3eb12ba3697a9e474d20f899e52d8e27b43a6fc4c4c809d2a099646744 |
small-de/cross_kv.ort |
5,358,752 | 0c329cbf001bebbb9e34835e95676b9eb29e010405da46f65af511b8da16620e |
small-de/decoder_kv.ort |
61,314,512 | 7da8ac9a1da2a486158ccde88c756fd876bf7f11ef1c2ba14b80e8edcd30028c |
small-de/encoder.ort |
44,358,376 | f31ef482f940e23af87906aeec986bf2e0a4862cad577bbc745264caab5a1858 |
small-de/frontend.model.ort |
26,776 | a93453223d93aa41d21691568cde3a01125fe00710a5f7004a825e1a322d3025 |
small-de/frontend.weights.ort |
7,769,280 | 8b417e2e0ee762e34d7163d24383823bb8bd1d037fb6966c32b3f86a3de4961f |
small-de/streaming_config.json |
512 | 12d16c7f5ea6734d197b79baf47914ea7e996d6fca8d303a19aca10d0617cecc |
small-de/tokenizer.bin |
103,319 | 5725ad3771a1339802f8cebeda0fa8f7723f5282b1ee726c2544774d7399b9a7 |
Use with anira
Pin a revision and check each file's SHA256 (listed above) when you download:
https://huggingface.co/anira-project/moonshine-streaming/resolve/<revision>/tiny-en/encoder.ort
Licence
MIT. Moonshine's licence states that its models are released under the MIT License by default, in
every language and at every size, including all streaming speech-to-text models; only legacy
non-streaming models for languages other than English are under the Moonshine Community License, and
none of those are here. LICENSE is Moonshine's licence file, copied from
moonshine-ai/moonshine @ 234f60f.