Amazon Transcribe カスタム言語モデルで「お食事券」と「汚職事件」を聞き分ける
https://qiita.com/mksamba/items/6b38ac0d49eb7d060fee?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items

#qiita #AWS #Transcribe #AmazonTranscribe

Amazon Transcribe カスタム言語モデルで「お食事券」と「汚職事件」を聞き分ける - Qiita

1. 背景 業務で少しTranscribeを使っている。 Transcribeの音声認識の精度を上げる方法として、「カスタムボキャブラリ」と「カスタム言語モデル」がある。 「カスタムボキャブラリ」は、標準のTranscribeでは認識されにくい専門用語などを、個別に事前...

Qiita

#RealtimeTranscription for #PyConIT 🎙️

Whisper hallucinated on real conference audio: a problem I skipped for time: everything else to build.

🔮 Spoiler: #AmazonTranscribe doesn't generate text, it decodes it.

In the article: the choices I made, and the stories you only find when plugging things in (Linux audio is a zoo) 😄

https://alessandra.bilardi.net/diary/articles/2026-04/realtime-transcription-choices-and-stories-for-pycon-it.en

#DiaryOfALazyDeveloper #aws #transcribe #docker #fastapi

Are you feeding #AWS your data for it's AI? If you use any of these, AWS is probably training on your data:

Amazon #CodeGuru Profiler, Amazon #CodeWhisperer Individual, Amazon #Comprehend, Amazon #Lex, Amazon #Polly, Amazon #Rekognition, Amazon #Textract, #AmazonTranscribe, and #AmazonTranslate.

https://www.lastweekinaws.com/blog/How-to-Stop-Feeding-AWSs-AI-With-Your-Data/

How to Stop Feeding AWS's AI With Your Data

AWS may be using your data to train its AI models, and you may have unwittingly consented to it. Prepare to jump through a series of complex hoops to stop it.

Last Week in AWS