Library to transcribe YouTube videos using Whisper model
Project description
Transcribe YouTube videos using Whisper models
Adopts faster_whisperer, a cTransformer's based model for faster transcription.
Usage
from youtescribe import transcribe
transcript = transcribe(url="https://www.youtube.com/watch?v=9bZkp7q19f0")
transcript.text()
Prompting
By default, the video title and description are used as prompts to the transcription model. But you can also specify your own prompt:
transcript = transcribe(
url="https://www.youtube.com/watch?v=9bZkp7q19f0",
prompt="Enter prompt here"
)
You can also choose not to include prompt by setting prompt=False
.
transcript = transcribe(
url="https://www.youtube.com/watch?v=9bZkp7q19f0",
prompt=False
)
Working with WhisperTranscript
objects
The transcribe()
function, if executed successfully, will return a WhisperTranscript
object. You can view the transcript as plain text, SRT-formatted text, or a Python dictionary.
transcript = transcribe(
url="https://www.youtube.com/watch?v=9bZkp7q19f0",
prompt=False
)
transcript.text()
transcript.srt()
transcript.json()
transcript.segment
Customise Whisper model
In the transcribe function, you can pass your own custom Whisper model:
from youtescribe import WhisperTranscriber
from youtescribe import models
custom_transcriber = WhisperTranscriber(model_size = models.TINY_EN, cpu_threads=6, device="auto")
transcript = transcribe(
url="https://www.youtube.com/watch?v=9bZkp7q19f0",
transcriber=custom_transcriber
)
transcript.text()
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
youscribe-0.0.2.tar.gz
(7.4 kB
view hashes)
Built Distribution
Close
Hashes for youscribe-0.0.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | b5959477203e5e2848cd876094f852f8cc8f271326937db980385425b6314114 |
|
MD5 | 5b4c7a7d44f25dff18c4a7727d7c633e |
|
BLAKE2b-256 | d3fc43f2d5fa9847f0dbadfeb7e4e43ff9f3edd3cf2d60e3319610c0fb910da0 |