Jannah Theme License is not validated, Go to the theme options page to validate the license, You need a single license for each domain name.
Tech

OpenAI President Shares First Image Generated by GPT-4o

Join us as we return to New York on June 5 to collaborate with leaders to explore comprehensive methods for auditing AI models for bias, performance, and ethical compliance in diverse organizations. Find out how you can attend here.


Greg Brockman, President of OpenAI posted from his X account what appears to be the first public image generated using the company’s newest GPT-4o model.

As you’ll see in the image below, it’s quite convincingly photorealistic, showing a person wearing a black T-shirt with an OpenAI logo writing text in chalk on a blackboard that says ” Transfer between modalities. » Suppose we directly model P (text, pixels, sound) with a large autoregressive transformer. What are the advantages and disadvantages?”

The new GPT-4o model, which debuted Monday, improves on the previous GPT-4 family of models (GPT-4, GPT-4 Vision and GPT-4 Turbo) by being faster, cheaper and retaining more energy. ‘entry information. such as audio and vision.

It is able to do this because OpenAI has taken a different approach from its previous GPT-4 class LLMs. While these chained together several different models and converted other media such as audio and visuals to text and back, the new GPT-4o was trained on media tokens from the start, allowing it to analyze and to directly interpret vision and audio without prior conversion. in text.

VB event

The AI ​​Impact Tour: The AI ​​Audit

Join us when we return to New York on June 5 to engage with top leaders and dig deeper into strategies for auditing AI models to ensure fairness, peak performance, and ethical compliance across diverse organizations. Ensure your participation in this exclusive, invitation-only event.

Request an invitation

Based on the image above, the new approach is a notable improvement over OpenAI’s latest image generation model, DALL-E 3, which debuted in September 2023. I ran a similar prompt via DALL-E 3 in ChatGPT and this is the result.

As you can see, the image Brockman shared created with GPT-4o improves significantly in terms of quality, photorealism, and text generation accuracy.

However, GPT-4o’s native image generation capabilities are not yet publicly available. As Brockman mentioned in his X article saying “The team is working hard to bring them into the world.”



News Source : venturebeat.com
Gn tech

Back to top button