OpenAI, the research lab behind ChatGPT/GPT-3 and DALL-E, has launched a new developer platform called Foundry, according to Techcrunch. Foundry allows customers to run OpenAI’s latest AI models on dedicated compute resources with full control over the model configuration and performance profile.
Foundry is designed for cutting-edge customers who need to run larger workloads with high throughput and low latency. Customers can choose from a range of models, such as GPT-3.5, CLIP, DALL-E, Codex, and more. They can also customize the model parameters, such as temperature, top-k, top-p, frequency penalty, presence penalty, etc.
Tools to Boost Model Performance
Foundry also provides tools for monitoring and optimizing the model performance. Customers can view metrics such as inference time, memory usage, cost per token, etc. They can also use Foundry’s API to integrate the models with their own applications and workflows.
Foundry appears to be currently in private beta, requiring an invitation to access. Customers who are interested in using Foundry can request access through OpenAI’s website. The pricing for Foundry is not yet disclosed but it is expected to be based on the amount of compute resources used by each customer.
Tip of the day: Do you know that Windows 11 / Windows 10 allows creating PDFs from basically any app with printing support? In our tutorial, we show you how this works via Microsoft Print to PDF and Bullzip PDF Printer to save a PDF from any app, even with advanced options like adjusted quality, multi-page printing, and password protection.
