Financial Coconut Transcripts
End-to-end ML project
Links
Summary
Built an end-to-end ML project from model training to prompt engineering to front end
- Trained a range of OpenAI Whisper (text-to-speech) models on 800GB of Singaporean-accented English
- Used finetuned model to transcribe podcast
- Used Large Language Models(LLMs) to summarise transcripts
- Self-learnt React (along with JSX, HTML and CSS) in 1.5 weeks to build front-end
- ML Ops to automate pipeline of downloading video, transcribing, summarising and updating website
Landing Page
Individual Page
