How to Use NVIDIA Canary-1B-v2 for ASR, Translation, and Automatic SRT Subtitle Export in Python

The article presents a tutorial on utilizing NVIDIA's Canary-1B-v2 model for automatic speech recognition (ASR) and translation tasks in Python. It details the process of preparing audio data at 16 kHz mono, performing ASR in English, translating into multiple languages (French, German, Spanish, Italian), and exporting subtitles in SRT format, while benchmarking inference speed for performance evaluation. This resource is significant for practitioners looking to implement multilingual ASR and translation capabilities efficiently using state-of-the-art models.