Converts the original large-v3 Whisper weights into CTranslate2 format that faster-whisper loads, writing float16 weights into whisper-large-v3-ct2. The converter ships with the ctranslate2 package. Load the result by passing the output directory as the model path in faster-whisper.
Looking for more? Search all 7,657 commands — works offline, in English or Spanish, and fixes typos.