Multilanguage PR

whispering/cli.py

@@ -13,12 +13,18 @@ import torch
 from whisper import available_models
 from whisper.audio import N_FRAMES, SAMPLE_RATE
 from whisper.tokenizer import LANGUAGES, TO_LANGUAGE_CODE
-
+"""
 from whispering.pbar import ProgressBar
 from whispering.schema import Context, WhisperConfig
 from whispering.serve import serve_with_websocket
 from whispering.transcriber import WhisperStreamingTranscriber
 from whispering.websocket_client import run_websocket_client
+"""
+from pbar import ProgressBar
+from schema import Context, WhisperConfig
+from serve import serve_with_websocket
+from transcriber import WhisperStreamingTranscriber
+from websocket_client import run_websocket_client
 
 logger = getLogger(__name__)
 
@@ -105,7 +111,7 @@ def get_opts() -> argparse.Namespace:
         "--language",
         type=str,
         default=None,
-        choices=sorted(LANGUAGES.keys())
+        choices=["multilanguage"] + sorted(LANGUAGES.keys())
         + sorted([k.title() for k in TO_LANGUAGE_CODE.keys()]),
     )
     group_model.add_argument(

whispering/serve.py

@@ -9,7 +9,7 @@ import numpy as np
 import websockets
 from websockets.exceptions import ConnectionClosedOK
 
-from whispering.transcriber import Context, WhisperStreamingTranscriber
+from transcriber import Context, WhisperStreamingTranscriber
 
 logger = getLogger(__name__)
 
@@ -18,7 +18,6 @@ async def serve_with_websocket_main(websocket):
     global g_wsp
     idx: int = 0
     ctx: Optional[Context] = None
-
     while True:
         logger.debug(f"Audio #: {idx}")
         try:
@@ -78,7 +77,7 @@ async def serve_with_websocket(
             serve_with_websocket_main,
             host=host,
             port=port,
-            max_size=999999999,
+            max_size=999999999
         ):
             await asyncio.Future()
     except KeyboardInterrupt:

whispering/transcriber.py

@@ -18,8 +18,8 @@ from whisper.decoding import DecodingOptions, DecodingResult
 from whisper.tokenizer import get_tokenizer
 from whisper.utils import exact_div
 
-from whispering.schema import Context, ParsedChunk, WhisperConfig
-from whispering.vad import VAD
+from schema import Context, ParsedChunk, WhisperConfig
+from vad import VAD
 
 logger = getLogger(__name__)
 
@@ -41,11 +41,19 @@ class WhisperStreamingTranscriber:
     def __init__(self, *, config: WhisperConfig):
         self.config: Final[WhisperConfig] = config
         self.model: Final[Whisper] = load_model(config.model_name, device=config.device)
-        self.tokenizer = get_tokenizer(
-            self.model.is_multilingual,
-            language=config.language,
-            task="transcribe",
-        )
+        # language specified
+        if config.language != "multilanguage":
+            self.tokenizer = get_tokenizer(
+                self.model.is_multilingual,
+                language=config.language,
+                task="transcribe",
+            )
+        # Mulilanguage transcripts
+        else: 
+            self.tokenizer = get_tokenizer(
+                self.model.is_multilingual,
+                task="transcribe",
+            )
         self._set_dtype(config.fp16)
         self.input_stride: Final[int] = exact_div(
             N_FRAMES, self.model.dims.n_audio_ctx
@@ -65,23 +73,41 @@ class WhisperStreamingTranscriber:
         patience: Optional[float],
         best_of: Optional[int],
     ) -> DecodingOptions:
-        return DecodingOptions(
-            task="transcribe",
-            language=self.config.language,
-            temperature=t,
-            sample_len=None,
-            best_of=best_of,
-            beam_size=beam_size,
-            patience=patience,
-            length_penalty=None,
-            prompt=prompt,
-            prefix=None,
-            suppress_blank=True,
-            suppress_tokens="-1",
-            without_timestamps=False,
-            max_initial_timestamp=1.0,
-            fp16=self.fp16,
-        )
+        if self.config.language != "multilanguage":
+            return DecodingOptions(
+                task="transcribe",
+                language=self.config.language,
+                temperature=t,
+                sample_len=None,
+                best_of=best_of,
+                beam_size=beam_size,
+                patience=patience,
+                length_penalty=None,
+                prompt=prompt,
+                prefix=None,
+                suppress_blank=True,
+                suppress_tokens="-1",
+                without_timestamps=False,
+                max_initial_timestamp=1.0,
+                fp16=self.fp16,
+            )
+        else:
+            return DecodingOptions(
+                task="transcribe",
+                temperature=t,
+                sample_len=None,
+                best_of=best_of,
+                beam_size=beam_size,
+                patience=patience,
+                length_penalty=None,
+                prompt=prompt,
+                prefix=None,
+                suppress_blank=True,
+                suppress_tokens="-1",
+                without_timestamps=False,
+                max_initial_timestamp=1.0,
+                fp16=self.fp16,
+            )
 
     def _decode_with_fallback(
         self,
@@ -106,7 +132,6 @@ class WhisperStreamingTranscriber:
                 _decode_options,
             )  # type: ignore
             assert decode_result is not None
-
             needs_fallback: bool = False
             if (
                 ctx.compression_ratio_threshold is not None

whispering/vad.py

@@ -6,7 +6,7 @@ import numpy as np
 import torch
 from whisper.audio import N_FRAMES, SAMPLE_RATE
 
-from whispering.schema import SpeechSegment
+from schema import SpeechSegment
 
 
 class VAD:

whispering/websocket_client.py

@@ -9,8 +9,8 @@ import sounddevice as sd
 import websockets
 from whisper.audio import N_FRAMES, SAMPLE_RATE
 
-from whispering.schema import ParsedChunk
-from whispering.transcriber import Context
+from schema import ParsedChunk
+from transcriber import Context
 
 logger = getLogger(__name__)