Repolode
Submit
jolibrain

jolibrain/deepdetect

A deep learning server, Python wheel and CLI for training and running models on images, text, tabular and time-series data via a REST API.

View on GitHubWebsite
Stars
2,551
Forks
544
Open issues
99
Last commit
8 days ago

What is it?

DeepDetect is a deep learning runtime that can be used three ways: as a Python wheel that embeds the runtime in the current environment with a `deepdetect` CLI, as a long-running REST server, or through Docker-based serving. It covers creating model services backed by local model repositories, running asynchronous training jobs, monitoring their status, and running predictions.

The project supports images, text, CSV/tabular data, time series, sparse data, object-detection boxes and segmentation masks through a single API. The main backend is Torch, with TensorRT available for optimized inference. Built-in templates include YOLOX and SegFormer CLI profiles, plus TorchVision-style classifiers, Faster R-CNN, RetinaNet, BERT, GPT-2 and several time-series models.

Model storage is filesystem-based, so no database is required. Configuration is driven by YAML files for repeatable runs, and the CLI can emit structured events for automation.

Key features

  • 01Train and run inference from a CLI or a REST API.
  • 02Create model services backed by local model repositories without a database.
  • 03Run asynchronous training jobs and check their status.
  • 04Supports image, text, tabular, time-series and sparse data through one API.
  • 05Ships YAML-configured CLI profiles for YOLOX and SegFormer.
  • 06Uses Torch as the primary backend with optional TensorRT inference.

Use cases

  • Running repeatable object detection training and inference from the command line.
  • Serving deep learning models behind a REST API in a dedicated process or container.
  • Adding model training and prediction to automation pipelines via structured CLI output.
  • Deploying time-series or tabular models with a filesystem-based model store.

Quick start

  1. 1

    Install the CPU wheel into a Python environment

    python -m pip install --extra-index-url https://www.deepdetect.com/download/wheels/simple deepdetect-cpu
  2. 2

    List the packaged CLI profiles and options

    deepdetect inspect models
  3. 3

    Prepare the YOLOX quickstart dataset and model repository

    python bindings/python/scripts/prepare_cli_yolox_quickstart.py --output /tmp/deepdetect-yolox-quickstart --force
  4. 4

    Train a very small YOLOX run

    deepdetect train yolox --config /tmp/deepdetect-yolox-quickstart/yolox-quickstart.yaml --terminal live

Things to know

  • Licensed under the LGPL v3, a copyleft license.
  • The packaged CLI covers only YOLOX and SegFormer workflows; the broader API requires the server.
  • The CPU and GPU wheels cannot be installed in the same Python environment.

In collections

Run and train AI on GPUsComputer vision libraries & tools

Topics

#deep-learning#machine-learning#rest-api#object-detection#image-classification#tensorrt#pytorch#inference-server#caffe#gpu

FAQ

What is deepdetect?
A deep learning server, Python wheel and CLI for training and running models on images, text, tabular and time-series data via a REST API. DeepDetect is a deep learning runtime that can be used three ways: as a Python wheel that embeds the runtime in the current environment with a `deepdetect` CLI, as a long-running REST server, or through Docker-based serving. It covers creating model services backed by local model repositories, running asynchronous training jobs, monitoring their status, and running predictions.
Is deepdetect free to use? What license does it have?
deepdetect is open source under the Other license. Check the license terms before using it in commercial products.
Is deepdetect still maintained?
Actively maintained. The last commit to deepdetect was on Oct 30, 2026.
What can I use deepdetect for?
Running repeatable object detection training and inference from the command line. Serving deep learning models behind a REST API in a dedicated process or container. Adding model training and prediction to automation pipelines via structured CLI output. Deploying time-series or tabular models with a filesystem-based model store.
How do I install deepdetect?
Install the CPU wheel into a Python environment: python -m pip install --extra-index-url https://www.deepdetect.com/download/wheels/simple deepdetect-cpu List the packaged CLI profiles and options: deepdetect inspect models Prepare the YOLOX quickstart dataset and model repository: python bindings/python/scripts/prepare_cli_yolox_quickstart.py --output /tmp/deepdetect-yolox-quickstart --force Train a very small YOLOX run: deepdetect train yolox --config /tmp/deepdetect-yolox-quickstart/yolox-quickstart.yaml --terminal live
Where can I get deepdetect?
The source code and installation instructions are on GitHub: https://github.com/jolibrain/deepdetect

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects