remove log_prob_low_threshold (#1160)

faster_whisper/transcribe.py

@@ -77,7 +77,6 @@ class TranscriptionOptions:
     repetition_penalty: float
     no_repeat_ngram_size: int
     log_prob_threshold: Optional[float]
-    log_prob_low_threshold: Optional[float]
     no_speech_threshold: Optional[float]
     compression_ratio_threshold: Optional[float]
     condition_on_previous_text: bool
@@ -275,7 +274,6 @@ class BatchedInferencePipeline:
         ],
         compression_ratio_threshold: Optional[float] = 2.4,
         log_prob_threshold: Optional[float] = -1.0,
-        log_prob_low_threshold: Optional[float] = None,
         no_speech_threshold: Optional[float] = 0.6,
         condition_on_previous_text: bool = True,
         prompt_reset_on_temperature: float = 0.5,
@@ -356,9 +354,6 @@ class BatchedInferencePipeline:
                 treat as failed.
             log_prob_threshold: If the average log probability over sampled tokens is
                 below this value, treat as failed.
-            log_prob_low_threshold: This parameter alone is sufficient to skip an output text,
-                whereas log_prob_threshold also looks for appropriate no_speech_threshold value.
-                This value should be less than log_prob_threshold.
             no_speech_threshold: If the no_speech probability is higher than this value AND
                 the average log probability over sampled tokens is below `log_prob_threshold`,
                 consider the segment as silent.
@@ -490,7 +485,6 @@ class BatchedInferencePipeline:
             repetition_penalty=repetition_penalty,
             no_repeat_ngram_size=no_repeat_ngram_size,
             log_prob_threshold=log_prob_threshold,
-            log_prob_low_threshold=log_prob_low_threshold,
             no_speech_threshold=no_speech_threshold,
             compression_ratio_threshold=compression_ratio_threshold,
             temperatures=(
@@ -636,12 +630,10 @@ class WhisperModel:
                 local_files_only=local_files_only,
                 cache_dir=download_root,
             )
-        self.device = device
-        # set the random seed to make sure consistency across runs
-        ctranslate2.set_random_seed(42)
+
         self.model = ctranslate2.models.Whisper(
             model_path,
-            device=self.device,
+            device=device,
             device_index=device_index,
             compute_type=compute_type,
             intra_threads=cpu_threads,
@@ -719,7 +711,6 @@ class WhisperModel:
         ],
         compression_ratio_threshold: Optional[float] = 2.4,
         log_prob_threshold: Optional[float] = -1.0,
-        log_prob_low_threshold: Optional[float] = None,
         no_speech_threshold: Optional[float] = 0.6,
         condition_on_previous_text: bool = True,
         prompt_reset_on_temperature: float = 0.5,
@@ -766,9 +757,6 @@ class WhisperModel:
             treat as failed.
           log_prob_threshold: If the average log probability over sampled tokens is
             below this value, treat as failed.
-          log_prob_low_threshold: This parameter alone is sufficient to skip an output text,
-          wheras log_prob_threshold also looks for appropriate no_speech_threshold value.
-          This value should be less than log_prob_threshold.
           no_speech_threshold: If the no_speech probability is higher than this value AND
             the average log probability over sampled tokens is below `log_prob_threshold`,
             consider the segment as silent.
@@ -820,7 +808,6 @@ class WhisperModel:
             - a generator over transcribed segments
             - an instance of TranscriptionInfo
         """
-
         sampling_rate = self.feature_extractor.sampling_rate
 
         if multilingual and not self.model.is_multilingual:
@@ -933,7 +920,6 @@ class WhisperModel:
             repetition_penalty=repetition_penalty,
             no_repeat_ngram_size=no_repeat_ngram_size,
             log_prob_threshold=log_prob_threshold,
-            log_prob_low_threshold=log_prob_low_threshold,
             no_speech_threshold=no_speech_threshold,
             compression_ratio_threshold=compression_ratio_threshold,
             condition_on_previous_text=condition_on_previous_text,
@@ -977,6 +963,7 @@ class WhisperModel:
             vad_options=vad_parameters,
             all_language_probs=all_language_probs,
         )
+
         return segments, info
 
     def _split_segments_by_timestamps(
@@ -1188,18 +1175,6 @@ class WhisperModel:
                         options.no_speech_threshold,
                     )
 
-                # Skip if the logprob is very low (below the threshold value),
-                # despite no_speech_prob being low (ex: Too ambiguous outputs)
-                if options.log_prob_low_threshold:
-                    if avg_logprob < options.log_prob_low_threshold:
-                        should_skip = True
-                        self.logger.debug(
-                            "log prob low threshold is met (%f > %f)",
-                            avg_logprob,
-                            options.log_prob_low_threshold,
-                        )
-
-                if should_skip:
                     # fast-forward to the next segment boundary
                     seek += segment_size
                     continue