Mirroristas/whisper
2
0
Fork 0
mirror of https://github.com/openai/whisper.git synced 2025-11-24 06:26:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge abf6778935308f636c7d5c6bc2861f67cc699d53 into c0d2f624c09dc18e709e37c2ad90c039a4eb72a2

Browse Source
This commit is contained in:
Millan Kumar 2025-06-27 02:28:02 +00:00 committed by GitHub
commit 085ed74035
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 16 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
whisper/transcribe.py
View File

@ -2,7 +2,7 @@ import argparse
import os import os
import traceback import traceback
import warnings import warnings
from typing import TYPE_CHECKING, List, Optional, Tuple, Union from typing import TYPE_CHECKING, List, Optional, Tuple, Union, Callable
import numpy as np import numpy as np
import torch import torch
@ -52,6 +52,7 @@ def transcribe(
append_punctuations: str = "\"'.。,，!！?？:：”)]}、", append_punctuations: str = "\"'.。,，!！?？:：”)]}、",
clip_timestamps: Union[str, List[float]] = "0", clip_timestamps: Union[str, List[float]] = "0",
hallucination_silence_threshold: Optional[float] = None, hallucination_silence_threshold: Optional[float] = None,
callback: Optional[Callable[[int, int, float], None]] = None,
**decode_options, **decode_options,
): ):
""" """
@ -119,6 +120,10 @@ def transcribe(
When word_timestamps is True, skip silent periods longer than this threshold (in seconds) When word_timestamps is True, skip silent periods longer than this threshold (in seconds)
when a possible hallucination is detected when a possible hallucination is detected
callback: Optional[Callable[int, int, float]] = None,
After each step in the transcription process, call the callback function with
the arguments current posistion, total frames, estimated time to finish in seconds
Returns Returns
------- -------
A dictionary containing the resulting text ("text") and segment-level details ("segments"), and A dictionary containing the resulting text ("text") and segment-level details ("segments"), and
@ -504,8 +509,17 @@ def transcribe(
# do not feed the prompt tokens if a high temperature was used # do not feed the prompt tokens if a high temperature was used
prompt_reset_since = len(all_tokens) prompt_reset_since = len(all_tokens)
total_position = min(content_frames, seek)
increase = total_position - previous_seek
if callback is not None:
rate = pbar.format_dict["rate"]
remaining = (pbar.total - pbar.n) / rate if rate and pbar.total else 0
callback(total_position, content_frames, remaining)
# update progress bar # update progress bar
pbar.update(min(content_frames, seek) - previous_seek) pbar.update(increase)
return dict( return dict(
text=tokenizer.decode(all_tokens[len(initial_prompt_tokens) :]), text=tokenizer.decode(all_tokens[len(initial_prompt_tokens) :]),