Add moonshine

model.py

@@ -516,6 +516,61 @@ def _get_russian_pre_trained_model(
     return recognizer
 
 
+@lru_cache(maxsize=10)
+def _get_moonshine_model(
+    repo_id: str, decoding_method: str, num_active_paths: int
+) -> sherpa_onnx.OfflineRecognizer:
+    assert repo_id in ("moonshine-tiny", "moonshine-base"), repo_id
+
+    if repo_id == "moonshine-tiny":
+        full_repo_id = "csukuangfj/sherpa-onnx-moonshine-tiny-en-int8"
+    elif repo_id == "moonshine-base":
+        full_repo_id = "csukuangfj/sherpa-onnx-moonshine-base-en-int8"
+    else:
+        raise ValueError(f"Unknown repo_id: {repo_id}")
+
+    preprocessor = _get_nn_model_filename(
+        repo_id=full_repo_id,
+        filename=f"preprocess.onnx",
+        subfolder=".",
+    )
+
+    encoder = _get_nn_model_filename(
+        repo_id=full_repo_id,
+        filename=f"encode.int8.onnx",
+        subfolder=".",
+    )
+
+    uncached_decoder = _get_nn_model_filename(
+        repo_id=full_repo_id,
+        filename=f"uncached_decode.int8.onnx",
+        subfolder=".",
+    )
+
+    cached_decoder = _get_nn_model_filename(
+        repo_id=full_repo_id,
+        filename=f"cached_decode.int8.onnx",
+        subfolder=".",
+    )
+
+    tokens = _get_token_filename(
+        repo_id=full_repo_id,
+        subfolder=".",
+        filename="tokens.txt",
+    )
+
+    recognizer = sherpa_onnx.OfflineRecognizer.from_moonshine(
+        preprocessor=preprocessor,
+        encoder=encoder,
+        uncached_decoder=uncached_decoder,
+        cached_decoder=cached_decoder,
+        tokens=tokens,
+        num_threads=2,
+    )
+
+    return recognizer
+
+
 @lru_cache(maxsize=10)
 def _get_whisper_model(
     repo_id: str, decoding_method: str, num_active_paths: int
@@ -1618,6 +1673,8 @@ english_models = {
     "whisper-tiny.en": _get_whisper_model,
     "whisper-base.en": _get_whisper_model,
     "whisper-small.en": _get_whisper_model,
+    "moonshine-tiny": _get_moonshine_model,
+    "moonshine-base": _get_moonshine_model,
     "csukuangfj/sherpa-onnx-nemo-parakeet_tdt_ctc_110m-en-36000": _get_sherpa_onnx_nemo_ctc_models,
     "csukuangfj/sherpa-onnx-nemo-parakeet_tdt_transducer_110m-en-36000": _get_sherpa_onnx_nemo_transducer_models,
     #  "whisper-medium.en": _get_whisper_model,