warmup base whisper when using simulstreaming

whisperlivekit/core.py

@@ -2,7 +2,7 @@ try:
     from whisperlivekit.whisper_streaming_custom.whisper_online import backend_factory
     from whisperlivekit.whisper_streaming_custom.online_asr import VACOnlineASRProcessor, OnlineASRProcessor
 except ImportError:
-    from .whisper_streaming_custom.whisper_online import backend_factory, warmup_asr
+    from .whisper_streaming_custom.whisper_online import backend_factory
     from .whisper_streaming_custom.online_asr import VACOnlineASRProcessor, OnlineASRProcessor
 from whisperlivekit.warmup import warmup_asr, warmup_online
 from argparse import Namespace
@@ -109,7 +109,7 @@ class TranscriptionEngine:
 
             else:
                 self.asr, self.tokenizer = backend_factory(self.args)
-                warmup_asr(self.asr, self.args.warmup_file) #for simulstreaming, warmup should be done in the online class not here
+            warmup_asr(self.asr, self.args.warmup_file) #for simulstreaming, warmup should be done in the online class not here
 
         if self.args.diarization:
             from whisperlivekit.diarization.diarization_online import DiartDiarization
@@ -130,7 +130,7 @@ def online_factory(args, asr, tokenizer, logfile=sys.stderr):
             asr,
             logfile=logfile,
         )
-        warmup_online(online, args.warmup_file)
+        # warmup_online(online, args.warmup_file)
     elif args.vac:
         online = VACOnlineASRProcessor(
             args.min_chunk_size,

whisperlivekit/simul_whisper/backend.py

@@ -125,8 +125,6 @@ class SimulStreamingOnlineProcessor:
     def warmup(self, audio, init_prompt=""):
         """Warmup the SimulStreaming model."""
         try:
-            if isinstance(audio, np.ndarray):
-                audio = torch.from_numpy(audio).float()
             self.model.insert_audio(audio)
             self.model.infer(True)
             self.model.refresh_segment(complete=True)
@@ -217,3 +215,9 @@ class SimulStreamingASR():
             num_languages=self.model.model.num_languages,
             task="translate"
         )
+
+    def transcribe(self, audio):
+        """
+        Only used for warmup. It's a direct whisper call, not a simulstreaming call
+        """
+        self.whisper_model.transcribe(audio, language=self.original_language)

whisperlivekit/warmup.py

@@ -53,7 +53,7 @@ def warmup_asr(asr, warmup_file=None, timeout=5):
     Warmup the ASR model by transcribing a short audio file.
     """
     audio = load_file(warmup_file=None, timeout=5)
-    asr.warmup(audio)
+    asr.transcribe(audio)
     logger.info("ASR model is warmed up")
     
 def warmup_online(online, warmup_file=None, timeout=5):