jolibrain/deepdetect

Deep Learning Server and CLI for Torch and TensorRT

What it solves

DeepDetect provides a unified interface for managing the full lifecycle of deep learning models, from training to inference. It simplifies the process of deploying models as REST services or running them via a command-line tool, removing the need for complex infrastructure or database dependencies for model storage.

How it works

DeepDetect operates as a deep learning runtime that can be embedded into Python environments via a wheel or run as a standalone REST server. It uses Torch as its primary backend for training and inference, with TensorRT support for optimized inference. The system organizes model repositories on the filesystem and allows users to define repeatable workflows using YAML configuration files.

Who it’s for

It is designed for developers and ML engineers who need a practical, repeatable way to train and serve models across various data modalities (images, text, tabular data, and time series) without managing a separate database for model storage.

Highlights

  • Unified API: Supports image classification, object detection, semantic segmentation, and language models through a single surface.
  • Unified CLI: A command-line tool for repeatable training and inference workflows.
  • Flexible Deployment: Can be deployed as an embedded Python library, a standalone REST server, or within Docker containers.
  • Broad Model Support: Compatible with a wide range of architectures including YOLOX, SegFormer, BERT, GPT-2, and various TorchVision classifiers.
  • Filesystem-based Storage: Manages model repositories directly on the disk without requiring a database.

Related

  • Project
    pytorch-tabular/pytorch_tabular

    A high-level PyTorch framework that provides a unified interface for applying various deep learning architectures to tabular data.

  • Project
    open-mmlab/mmengine

    A foundational PyTorch library that serves as a training engine for deep learning models, providing standardized components for training loops, configuration, and monitoring.

  • Project
    texttron/tevatron

    Tevatron is a unified toolkit for training and evaluating neural retrievers and rerankers. It supports large language models on GPUs and TPUs, offers LoRA-based efficient tuning, and includes ready-to-use datasets and evaluation workflows for text, image, and multilingual retrieval.

  • Project
    pyg-team/pytorch-frame

    A modular deep learning extension for PyTorch designed to build neural network models on heterogeneous tabular data containing mixed column types like text, images, and numerical values.

  • Project
    tile-ai/TileRT

    A tile-based runtime engine for ultra-low-latency LLM inference that optimizes hardware utilization to achieve millisecond-level time per output token.