tracel-ai/models

Models and examples built with Burn

What it solves

This repository provides a centralized collection of deep learning models implemented using the Burn deep learning framework. It allows developers to access a variety of pre-built model architectures for different AI tasks without having to implement them from scratch.

How it works

The project organizes a set of official and community-contributed models. These implementations are built on top of the Burn framework, enabling the use of Rust-based deep learning. It covers a wide range of modalities including text, image, and audio.

Who it’s for

Developers and researchers who are using the Burn framework and need ready-to-use model architectures for tasks like language modeling, image classification, and speech recognition.

Highlights

  • Official implementations of models like Llama, ALBERT, and ResNet.
  • Community contributions including Stable Diffusion, Whisper, and RWKV v7.
  • Support for multiple modalities: text (LLMs, encoders), vision (classification, object detection), and audio (speech recognition, music note detection).
  • Distributed under MIT and Apache 2.0 licenses.

Related

  • Project
    tracel-ai/burn

    Burn is a unified tensor library and deep learning framework written in Rust that allows the same code to be used for both training and inference across diverse hardware.

  • Project
    microsoft/unilm

    A comprehensive collection of foundation models and architectures from Microsoft focusing on large-scale self-supervised pre-training across text, vision, speech, and multimodal tasks.

  • Project
    jolibrain/deepdetect

    A deep learning runtime and REST server that provides a unified API for training and inference across images, text, and tabular data.

  • Project
    ZHZisZZ/dllm

    A library that unifies the training, inference, and evaluation of diffusion language models, providing scalable pipelines and recipes for converting AR models into diffusion models.

  • Project
    huggingface/transformers

    A model-definition framework providing a unified API to access, train, and deploy state-of-the-art pretrained models across text, vision, audio, and multimodal domains.