Just built a real-time audio analysis app using Whisper-Large-v3 + DeepSeek-R1 — powered by SambaNova’s blazing-fast SN40L chip!
Upload audio or speak live → get instant transcripts & AI insights via Gradio UI.
Hit a sequence length error (804 > 448) during Whisper testing — working around it.
Big thanks to Durgesh from SambaNova for the support!
Demo + tech breakdown: