Skip to content
This repository has been archived by the owner on Dec 15, 2023. It is now read-only.

Latest commit

 

History

History
43 lines (34 loc) · 3.76 KB

README.md

File metadata and controls

43 lines (34 loc) · 3.76 KB
Raw

Whisper Timestamped | Worker

This repository contains the Whisper Timestamped Worker for RunPod. The Whisper Timestamped Worker is designed to process audio files using various Whisper models, with options for transcription formatting, language translation, and more. It's part of the RunPod Workers collection aimed at providing diverse functionality for endpoint processing.

Docker Image

Model Inputs

Input Type Description
audio Path Audio file
model str Choose a Whisper model. Choices: "tiny", "base", "small", "medium", "large-v1", "large-v2". Default: "base"
transcription str Choose the format for the transcription. Choices: "plain text", "srt", "vtt". Default: "plain text"
translate bool Translate the text to English when set to True. Default: False
language str Language spoken in the audio, specify None to perform language detection. Default: None
temperature float Temperature to use for sampling. Default: 0
best_of int Number of candidates when sampling with non-zero temperature. Default: 5
beam_size int Number of beams in beam search, only applicable when temperature is zero. Default: 5
patience float Optional patience value to use in beam decoding. Default: None
length_penalty float Optional token length penalty coefficient (alpha). Default: None
suppress_tokens str Comma-separated list of token ids to suppress during sampling. Default: "-1"
initial_prompt str Optional text to provide as a prompt for the first window. Default: None
condition_on_previous_text bool If True, provide the previous output of the model as a prompt for the next window. Default: True
temperature_increment_on_fallback float Temperature to increase when falling back when the decoding fails. Default: 0.2
compression_ratio_threshold float If the gzip compression ratio is higher than this value, treat the decoding as failed. Default: 2.4
logprob_threshold float If the average log probability is lower than this value, treat the decoding as failed. Default: -1.0
no_speech_threshold float If the probability of the token is higher than this value, consider the segment as silence. Default: 0.6

Test Inputs

The following inputs can be used for testing the model:

{
    "input": {
        "audio": "https://github.com/runpod-workers/sample-inputs/raw/main/audio/gettysburg.wav"
    }
}