Hugging Face has recently launched a groundbreaking new tool called the Transformers Agent. This tool is set to revolutionize how we manage over 100,000 HF models. The system supports both OpenAI modes and open-source alternatives from BigCode and OpenAssistant. The Transformers Agent provides a natural language API on top of transformers with a set of curated tools. Hugging Face has also announced its partnership with ServiceNow to develop a new open-source language model for codes called StarCoder.
Users can utilize an agent that interprets natural language and employs these tools. This cutting-edge tool aims to facilitate the management of a wide range of tasks by enabling communication with the Transformers and Diffusers interface.
Here’s everything you need to know about the Transformers Agent:
Two APIs Available
The Transformers Agent offers two APIs, making it easy for users to choose the one that best suits their needs. The first API is the single execution (run) API, which empowers users to execute single or multiple tasks within a single instruction. The second is the chat-based execution (chat) API, ideal for maintaining state across different instructions.
One of the most exciting features of the Hugging Face’s Transformers Agent is its set of curated tools. These tools consist of a single function with a designated name and description, which prompt an agent to perform a given task. The Transformers Agent offers many tools to meet your needs, from document-based question answering to text summarization. Other tools include unconditional image captioning, image segmentation, speech-to-text, text-to-speech, and zero-shot text classification.
In addition to its curated tools, the Transformers Agent allows users to create custom tools. Some of the custom tools available on the platform include text downloader, text-to-image, image transformation, and text-to-video.
Partnership with ServiceNow
Last week, Hugging Face announced its partnership with ServiceNow to develop a new open-source language model for codes called StarCoder. This model represents an enhanced version of the StarCoderBase model, which underwent training on 35 billion Python tokens.
The StarCoder has been tested on various benchmarks, including the HumanEval benchmark for Python. It outperformed larger models like PaLM, LaMDA, and LLaMA. It has also proven to be on par with or even better than closed models like OpenAI’s code-Cushman-001.
In conclusion, the Transformers Agent by Hugging Face is a game-changing tool offering many features and capabilities. With its natural language API, curated tools, and custom tools, users can easily manage various tasks, from document-based question answering to text summarization. Moreover, its partnership with ServiceNow and the development of the open-source language model for codes called StarCoder will further enhance its capabilities. So, if you’re looking for an all-in-one platform to help you manage your AI models, look no further than the Transformers Agent by Hugging Face.