Latest Posts

Search
2024-11-28
Tech
Peking University and ByteDance Launched FAN, an Improvement on Transformers, Focusing on Cyclical Features and Patterns

It seamlessly replaces traditional MLPs, reducing FLOPS and params.

Peking University and ByteDance Launched FAN, an Improvement on Transformers, Focusing on Cyclical Features and Patterns
2024-11-20
Tech
A Better Performer than CLIP, without Text Encoders? Meet SuperClass, a Pioneering Model by the Doubao (Seed) Team

Its research results have been selected for NeurIPS 2024.

A Better Performer than CLIP, without Text Encoders? Meet SuperClass, a Pioneering Model by the Doubao (Seed) Team
2024-11-11
Tech
Edit Images With a Single Phrase! ByteDance Announces Image Editing Model SeedEdit, Available for Test Use

Highly aesthetic editing with precise response to any command. Now open to testing!

Edit Images With a Single Phrase! ByteDance Announces Image Editing Model SeedEdit, Available for Test Use
2024-11-08
Tech
Can a Video Generative Model "Understand" the Laws of Physics? Doubao Big Model Team Announces

Research done by a post-95s and a post-00s generation

Can a Video Generative Model "Understand" the Laws of Physics? Doubao Big Model Team Announces
2024-11-01
Tech
Up to 20 times higher throughput! Doubao (Seed) team releases a brand new RLHF framework, now open-sourced!

Decoupling control flow and computing flow for both flexibility and efficiency.

Up to 20 times higher throughput! Doubao (Seed) team releases a brand new RLHF framework, now open-sourced!
2024-10-11
Tech
ByteDance and Tsinghua AIR Establish Joint Research Center to Advance Industry-Academia-Research Collaboration on Large Models

SIA Lab aims to achieve breakthroughs in foundational large model technologies and build industrial applications through effective industry-academia-research collaboration.

ByteDance and Tsinghua AIR Establish Joint Research Center to Advance Industry-Academia-Research Collaboration on Large Models
2024-09-24
Tech
Doubao releases two video generative models! Multiple vertical models are newly launched, and the research and development of basic models is in full swing!

Comprehensive efforts have been made in the research and development of the basic models, which have unique advantages in various business scenarios.

Doubao releases two video generative models! Multiple vertical models are newly launched, and the research and development of basic models is in full swing!
2024-09-18
Tech
Seed-Music: Music Large Model Officially Released! Exceling in both music generation and editing, covering ten types of creative tasks to meet diverse needs.

With a unified and flexible framework, it supports score-to-song conversion, controllable generation, music and lyrics editing, and low-threshold voice cloning, etc.

Seed-Music: Music Large Model Officially Released! Exceling in both music generation and editing, covering ten types of creative tasks to meet diverse needs.
2024-08-22
Tech
Doubao's "Listening" skills unboxing live! See how Seed-ASR overcomes the speech recognition bottleneck

"Pragmatic" and "technology-focused" are the team's fundamental principles.

Doubao's "Listening" skills unboxing live! See how Seed-ASR overcomes the speech recognition bottleneck
2024-08-08
Tech
Optimized by up to 529 times! Doubao's Large Model Teams Up with The University of Hong Kong (HKU) to Launch a New Checkpoint System to Optimize Training Efficiency

ByteCheckpoint improves performance by up to 529.22 times in saving and 3.51 times in loading speeds.

Optimized by up to 529 times! Doubao's Large Model Teams Up with The University of Hong Kong (HKU) to Launch a New Checkpoint System to Optimize Training Efficiency
2024-08-01
Tech
Doubao LLM's Visual Model and Voice Model Upgraded! Text-to-Image has a better understanding of "Chinese style", and TTS can "capture" emotions

Doubao large model's daily token usage exceeds 50 bn.

Doubao LLM's Visual Model and Voice Model Upgraded! Text-to-Image has a better understanding of "Chinese style", and TTS can "capture" emotions
2024-07-19
Tech
Selected for ICML Oral!A DOUBAO Team Intern Proposes IR-QLoRA, Introducing Information Theory to LLMs

This paper proposes a novel IR-QLoRA for pushing quantized LLMs with LoRA to be highly accurate through information retention. This is the first time the perspective of information theory has been introduced and theories related to information entropy are utilized to examine and measure the quantization of large models. The paper has been selected for an oral presentation at ICML 2024. The first author of the paper is an intern with the speech group of ByteDance's Doubao Team (Seed) and a candidate in the ByteDance Scholars Program. Teachers and students from the State Key Laboratory of Complex & Critical Software Environment at Beihang University also contributed to the research.

Selected for ICML Oral!A DOUBAO Team Intern Proposes IR-QLoRA, Introducing Information Theory to LLMs
2024-07-10
Tech
Uncover the Story behind Depth Anything - Selected by Apple CoreML Model Library and Boasts 8K Star on Github

The first author of the paper is a team intern.

Uncover the Story behind Depth Anything - Selected by Apple CoreML Model Library and Boasts 8K Star on Github
2024-07-04
Education
ByteDance Doubao LLM Offers Trillions of Tokens for Free Use

Volcano Engine unleashes the power of large models to support academic frontier exploration

ByteDance Doubao LLM Offers Trillions of Tokens for Free Use
2024-06-25
Team Story
A Peek into the Technologies behind Seed-TTS - the Generated Sound That is Too "Real" to Be True

The relevant technology has been applied for a while and has been recognized by users in real-life scenarios.

A Peek into the Technologies behind Seed-TTS - the Generated Sound That is Too "Real" to Be True
2024-06-24
Events
ByteDance CVPR 2024 Offline Event Review: Hundreds of Researchers Gathered to Shed Light on Generative AI Trends

Generative AI further enables the integration between the industry and the academic and research fields.

ByteDance CVPR 2024 Offline Event Review: Hundreds of Researchers Gathered to Shed Light on Generative AI Trends
1-16 of 16
Page 1 of 1