fix model download and vad fail

whisperkit/pipelines.py

@@ -13,6 +13,7 @@
 from typing import Optional
 
 from argmaxtools.utils import _maybe_git_clone, get_logger
+from huggingface_hub import snapshot_download
 
 from whisperkit import _constants
 
@@ -132,8 +133,29 @@ def clone_models(self):
         """ Download WhisperKit model files from Hugging Face Hub
         (only the files needed for `self.whisper_version`)
         """
-        self.models_dir = os.path.join(self.repo_dir, "models")  # dummy
-        self.results_dir = os.path.join(self.repo_dir, "results")
+        self.models_dir = os.path.join(
+            self.repo_dir, "Models", self.whisper_version.replace("/", "_"))
+
+        os.makedirs(self.models_dir, exist_ok=True)
+
+        snapshot_download(
+            repo_id=_constants.MODEL_REPO_ID,
+            allow_patterns=f"{self.whisper_version.replace('/', '_')}/*",
+            revision=self.model_commit_hash,
+            local_dir=os.path.dirname(self.models_dir),
+            local_dir_use_symlinks=True
+        )
+
+        if self.model_commit_hash is None:
+            self.model_commit_hash = subprocess.run(
+                f"git ls-remote git@hf.co:{_constants.MODEL_REPO_ID}",
+                shell=True, stdout=subprocess.PIPE
+            ).stdout.decode("utf-8").rsplit("\n")[0].rsplit("\t")[0]
+            logger.info(
+                "--model-commit-hash not specified, "
+                f"imputing with HEAD={self.model_commit_hash}")
+
+        self.results_dir = os.path.join(self.models_dir, "results")
         os.makedirs(self.results_dir, exist_ok=True)
 
     def transcribe(self, audio_file_path: str, forced_language: Optional[str] = None) -> str:
@@ -143,11 +165,10 @@ def transcribe(self, audio_file_path: str, forced_language: Optional[str] = None
             self.cli_path,
             "transcribe",
             "--audio-path", audio_file_path,
-            "--model-prefix", self.whisper_version.rsplit("/")[0],
-            "--model", self.whisper_version.rsplit("/")[1],
+            "--model-path", self.models_dir,
             "--text-decoder-compute-units", self._text_decoder_compute_units,
             "--audio-encoder-compute-units", self._audio_encoder_compute_units,
-            "--chunking-strategy", "vad",
+            # "--chunking-strategy", "vad",
             "--report-path", self.results_dir, "--report",
             "--word-timestamps" if self._word_timestamps else "",
             "" if forced_language is None else f"--use-prefill-prompt --language {forced_language}",
@@ -182,8 +203,7 @@ def transcribe_folder(self, audio_folder_path: str, forced_language: Optional[st
             self.cli_path,
             "transcribe",
             "--audio-folder", audio_folder_path,
-            "--model-prefix", self.whisper_version.rsplit("/")[0],
-            "--model", self.whisper_version.rsplit("/")[1],
+            "--model-path", self.models_dir,
             "--text-decoder-compute-units", self._text_decoder_compute_units,
             "--audio-encoder-compute-units", self._audio_encoder_compute_units,
             "--report-path", self.results_dir, "--report",