Researchers have developed a new approach to approximating self-attention, a key component of transformer models, using the Nyström method. This allow...
The U.S. National AI Research Resource has released an interim report outlining its goals and objectives for developing a national research infrastruc...
A new technique for training large language models has been developed by OpenAI. The method, called 'Efficient Training of Language Models to Fill in ...
Hugging Face has released an update to their dataset platform, adding support for audio and visual documentation. This includes features such as image...
Researchers have found a way to speed up text generation using TensorFlow's XLA compiler. By applying XLA's just-in-time compilation, they achieved si...
OpenAI has published a framework for analyzing potential hazards in large language models that synthesize code. The framework, designed to identify an...
Hugging Face has released a tutorial on deploying TensorFlow vision models using TF Serving. This allows users to serve and deploy pre-trained vision ...
Deep reinforcement learning algorithm A2C is a variant of policy gradient methods that uses an actor-critic approach to balance exploration and exploi...
OpenAI has announced that its text-to-image model DALL·E is now available in beta to a select group of users. One million people from the company's wa...
OpenAI has introduced a new method to reduce bias in its DALL·E 2 image generation model. The technique aims to produce people with characteristics th...
Researchers have developed a method for training machine learning models on dynamic, adversarial data. This approach involves creating artificial exam...
OpenAI's DALL·E 2 is a text-to-image model that has been adopted by over 3,000 artists from more than 118 countries. These artists have used the tool ...