🎙️ Next-Gen Speech Recognition
Built with NVIDIA Parakeet & Custom Fine-Tuning
This model was fine-tuned offline to achieve a highly competitive 0.29 Word Error Rate on a rigorous test dataset.
1. Input Audio
2. Transcription Result
System Specs: Parakeet-tdt-0.6b-v2 Base | Custom LoRA Adapter | Greedy Decoding