patches/batch-transcribe.patch

--- a/faster_whisper/transcribe.py
+++ b/faster_whisper/transcribe.py
@@ -219,6 +219,10 @@
             for i, language_token in enumerate(language_tokens):
                 prompts[i][language_token_index] = language_token

+        max_initial_timestamp_index = int(
+            round(options.max_initial_timestamp / self.model.time_precision)
+        )
+
         results = self.model.model.generate(
             encoder_output,
             prompts,
@@ -228,6 +232,7 @@
             max_length=max_length,
             suppress_blank=options.suppress_blank,
             suppress_tokens=options.suppress_tokens,
+            max_initial_timestamp_index=max_initial_timestamp_index,
             return_scores=True,
             return_no_speech_prob=True,
             sampling_temperature=options.temperatures[0],
@@ -280,7 +285,7 @@
         prefix: Optional[str] = None,
         suppress_blank: bool = True,
         suppress_tokens: Optional[List[int]] = [-1],
-        without_timestamps: bool = True,
+        without_timestamps: bool = False,
         max_initial_timestamp: float = 1.0,
         word_timestamps: bool = False,
         prepend_punctuations: str = "\"'"¿([{-",
@@ -321,6 +326,7 @@
             suppress_tokens: List of token IDs to suppress. -1 will suppress a default set
                 of symbols as defined in `tokenizer.non_speech_tokens()`.
             without_timestamps: Only sample text tokens.
+            max_initial_timestamp: The initial timestamp cannot be later than this.
             word_timestamps: Extract word-level timestamps using the cross-attention pattern
                 and dynamic time warping, and include the timestamps for each word in each segment.
                 Set as False.
@@ -363,7 +369,6 @@
             prompt_reset_on_temperature: Resets prompt if temperature is above this value.
                 Arg has effect only if condition_on_previous_text is True. Set at 0.5
             prefix: Optional text to provide as a prefix at the beginning of each window.
-            max_initial_timestamp: The initial timestamp cannot be later than this, set at 0.0.
             hallucination_silence_threshold: Optional[float]
                 When word_timestamps is True, skip silent periods longer than this threshold
                 (in seconds) when a possible hallucination is detected. set as None.
@@ -549,7 +554,7 @@
             prompt_reset_on_temperature=0.5,
             multilingual=multilingual,
             without_timestamps=without_timestamps,
-            max_initial_timestamp=0.0,
+            max_initial_timestamp=max_initial_timestamp,
         )

         info = TranscriptionInfo(