Daily Productive Sharing 845 - OpenAI's Updates

Daily Productive Sharing 845 - OpenAI's Updates
Photo by Vitalii Mazur / Unsplash

One helpful tip per day:)

On November 6, OpenAI hosted the Open Dev Conference to introduce their new progress. These advancements are very inspiring:

  1. GPT-4 will be updated to GPT-4 Turbo, which can process text data up to 300 pages in a single request and can handle single requests containing multiple commands;
  2. GPT-4 is also more developer-friendly: it is cheaper, can return data in JSON format, and if a seed parameter is included in the request, the results will remain consistent, which is convenient for debugging;
  3. The new assistant API is actually a more practical tool that allows developers to call other APIs, such as the Retrieval API, without the need to write their own web crawlers;
  4. Other updates include an API integrated with image processing capabilities, a TTS API that can convert text to speech, and the release of two models, Whisper large-v3 and Consistency Decoder.

If you enjoy today's sharing, why not subscribe

Need a superb CV, please try our CV Consultation

已经有超过五千位朋友通过各种渠道订阅我们的内容,你还在犹豫什么呢?不如直接支持我们 :)

OpenAI 于11月6日召开了 Open Dev 大会,介绍了他们的新进展。这些进展非常鼓舞人心:

  1. GPT-4 将更新至 GPT-4 Turbo,不仅可以在单次请求中可以处理长达300页的文本数据,而且可以处理包含多个命令的单次请求;
  2. GPT-4 对开发者也更加友好:价格更便宜,可以返回 Json 格式的数据,在请求中包含 seed 参数的话,返回结果将保持一致,方便 debug;
  3. 新的 assistant API 其实是更实用的一个工具,通过这一 API,开发者可以调用其他 API,比如调用 Retrival API 的话,就不需要自己再写爬虫;
  4. 其他的更新包括,集成了图片处理能力的 API,能够讲文本转换为语音的 TTS API,开源了 Whisper large-v3 和 Consistency Decoder 两个模型。

需要更丝滑的移民体验,不妨试试 AwesomeVisa