Fix to save the rest of mel

whisper_streaming/transcriber.py

@@ -4,6 +4,7 @@ from typing import Iterator, List, Optional, Union
 
 import numpy as np
 import torch
+from transformers.utils.dummy_pt_objects import OPT_PRETRAINED_MODEL_ARCHIVE_LIST
 from whisper import Whisper, load_model
 from whisper.audio import (
     HOP_LENGTH,
@@ -38,6 +39,7 @@ class WhisperStreamingTranscriber:
         )  # time per output token: 0.02 (seconds)
 
         self.buffer_tokens = []
+        self.buffer_mel = None
 
     def _get_decoding_options(
         self,
@@ -202,27 +204,52 @@ class WhisperStreamingTranscriber:
         *,
         segment: np.ndarray,
     ) -> Iterator[ParsedChunk]:
-        log_spec = log_mel_spectrogram(audio=segment).unsqueeze(0)
+        new_mel = log_mel_spectrogram(audio=segment).unsqueeze(0)
-        segment = (
+        if self.buffer_mel is None:
-            pad_or_trim(log_spec, N_FRAMES)
+            mel = new_mel
-            .to(self.model.device)  # type:ignore
+        else:
-            .to(self.dtype)  # type:ignore
+            mel = torch.cat([self.buffer_mel, new_mel], dim=-1)
-        )
+            self.buffer_mel = None
-        results = self._decode_with_fallback(
-            segment=segment,
-        )
-        result = results[0]
 
-        if self.config.no_speech_threshold is not None:
+        seek: int = 0
-            if (result.no_speech_prob > self.config.no_speech_threshold) and not (
+        rest_start: Optional[int] = None
-                self.config.logprob_threshold is not None
+        while seek < mel.shape[-1]:
-                and result.avg_logprob > self.config.logprob_threshold
+            segment = (
+                pad_or_trim(mel[:, :, seek:], N_FRAMES)
+                .to(self.model.device)  # type: ignore
+                .to(self.dtype)
+            )
+
+            results = self._decode_with_fallback(
+                segment=segment,
+            )
+            result = results[0]
+
+            if self.config.no_speech_threshold is not None:
+                if (result.no_speech_prob > self.config.no_speech_threshold) and not (
+                    self.config.logprob_threshold is not None
+                    and result.avg_logprob > self.config.logprob_threshold
+                ):
+                    return
+
+            segment_duration = segment.shape[-1] * HOP_LENGTH / SAMPLE_RATE
+            last_timestamp_position: Optional[int] = None
+            for v in self._deal_timestamp(
+                result=result, segment_duration=segment_duration
             ):
-                return
+                if isinstance(v, int):
-        segment_duration = segment.shape[-1] * HOP_LENGTH / SAMPLE_RATE
+                    last_timestamp_position = v
-        for v in self._deal_timestamp(result=result, segment_duration=segment_duration):
+                else:
-            if isinstance(v, int):
+                    yield v
-                # FIXME: save log_spec
+            if last_timestamp_position is None:
-                pass
+                seek += segment.shape[-1]
+                rest_start = None
             else:
-                yield v
+                seek += last_timestamp_position
+                rest_start = seek
+
+        if rest_start is None:
+            return
+
+        self.buffer_mel = mel[:, :, rest_start:]
+        del mel