Install Faster Whisper Multilingual STT
Budget: ₹1,500 – ₹12,500 INR
I need Faster-Whisper installed and fully tested both on my Linux-based VPS and on my personal workstation at the same time. The goal is to obtain fast, accurate speech-to-text that handles every major Indian language—Hindi, Tamil, Bengali and the rest of the country’s linguistic landscape—so I can integrate it into my educational content pipeline.
Here is what I am after:
• A clean-room installation of Faster-Whisper (Python, pip, ffmpeg, any CUDA or CPU-only extras) on the VPS and on my local machine, with identical environments where practical.
• Proof that the model correctly transcribes sample audio in multiple Indian languages, delivered as short demo clips or command-line outputs.
• A lightweight script or notebook that accepts an audio file (and, if possible, live microphone input) and returns text.
• Step-by-step documentation so I can repeat the setup or move it to another server later.
• A quick walk-through call or recorded screen share showing everything running end-to-end.
You’ll get full SSH access to the VPS and, via AnyDesk/TeamViewer or similar, to my local computer so we can configure both simultaneously. No changes should break existing services on the server.
If you have prior experience compiling Whisper variants, optimising for low-resource CPUs or leveraging GPUs for faster inference, please mention it when you respond—speed and stability are equally important for this project.
Here is what I am after:
• A clean-room installation of Faster-Whisper (Python, pip, ffmpeg, any CUDA or CPU-only extras) on the VPS and on my local machine, with identical environments where practical.
• Proof that the model correctly transcribes sample audio in multiple Indian languages, delivered as short demo clips or command-line outputs.
• A lightweight script or notebook that accepts an audio file (and, if possible, live microphone input) and returns text.
• Step-by-step documentation so I can repeat the setup or move it to another server later.
• A quick walk-through call or recorded screen share showing everything running end-to-end.
You’ll get full SSH access to the VPS and, via AnyDesk/TeamViewer or similar, to my local computer so we can configure both simultaneously. No changes should break existing services on the server.
If you have prior experience compiling Whisper variants, optimising for low-resource CPUs or leveraging GPUs for faster inference, please mention it when you respond—speed and stability are equally important for this project.