Blog

Building Powerful Agents with Adept

August 23, 2024 — Adept Product Team

Adept Workflow Language (AWL) is an expressive, custom language that allows users to easily compose powerful multimodal web interactions on top of Adept’s models.

Announcements
Research

Adept Fuyu-Heavy: A new multimodal model

January 24, 2024 — Adept Team

Adept Fuyu-Heavy is a new multimodal model designed specifically for digital agents.

Announcements
Research

Fuyu-8B: A Multimodal Architecture for AI Agents

October 17, 2023 — Rohan Bavishi, Erich Elsen, Curtis Hawthorne, Maxwell Nye, Augustus Odena, Arushi Somani, Sağnak Taşırlar

We’re open-sourcing Fuyu-8B - a small version of the multimodal model that powers our product.

Announcements
Research

The Adventure of the Errant Hardware

September 19, 2023 — Erich Elsen, Curtis Hawthorne, Arushi Somani

A tale of mystery, intrigue and derring-do. We recount our investigation into curious errors occuring during our large training runs–clues found, causes deciphered and solutions implemented.

Research

Releasing Persimmon-8B

September 7, 2023 — Erich Elsen, Augustus Odena, Maxwell Nye, Sağnak Taşırlar, Tri Dao, Curtis Hawthorne, Deepak Moparthi, Arushi Somani

We’re open-sourcing Persimmon-8B, the most powerful fully permissively-licensed language model with <10 billion parameters.

Announcements
Research

FlashAttention: Fast Transformer training with long sequences

January 17, 2023 — Tri Dao

‍Transformers have grown deeper and wider, but training them on long sequences remains difficult. The attention layer at their heart is the compute and memory bottleneck: doubling the sequence length would quadruple the runtime and memory requirements.

Research