2022年9月21日 · Whisper is an automatic speech recognition (ASR) system trained on 680,000 hours of multilingual and multitask supervised data collected from the web. We show that the use of such a large and diverse dataset leads to improved robustness to accents, background noise and technical language.

2024年4月24日 · Developers can now use our open-source Whisper large-v2 model in the API with much faster and cost-effective results. GPT-3.5 API users can expect continuous model improvements and the option to choose dedicated capacity for deeper control over the models.

Explore developer resources, tutorials, API docs, and dynamic examples to get the most out of OpenAI's platform.

We call our approach Whisper2. We demonstrate models trained at this scale transfer well to existing datasets zero-shot, removing the need for any dataset-specific fine-tuning to achieve high-quality results. In addition to scale, our work also focuses on broaden-ing the scope of weakly supervised pre-training beyond

OpenAI's Whisper models have the potential to be used in a wide range of applications, from transcription services to voice assistants and more. You can get started building with the Whisper API using our speech to text developer guide .

2024年10月27日 · The short answer is yes, the open-source Whisper model downloaded and run locally from the GitHub repository is safe in the sense that your audio data is not sent to OpenAI. You are running the model entirely on your own hardware (in this case, Google Colab’s servers), and you control the entire pipeline.

2023年11月14日 · At the moment, it is only possible to get timecodes within subtitle files (srt, vtt). If you want word alignment and timestamps, you would need to combine Whisper with some other alignment solutions - and as these models are built for each language separately, it complicates the integration a bit.

Whisper Audio API FAQ General questions about the Whisper, speech to text, Audio API

2023年7月8日 · A popular method is to combine the two and use time stamps to sync up the accurate whisper word detection with the other systems ability to detect who sad it and when.

2023年10月28日 · Update: this is definitely an API-related bug, as I just tried using the Github version of Whisper in my web app and it worked perfectly. Hopefully this gets resolved soon as I’d much rather use the official API!