sinapsis-object-detection 0.2.0

Mono repo with packages for inference and training with object detection models

Sinapsis Object Detection

Mono repo with packages for training and inference with various models for advanced object detection tasks.

🐍 Installation • 📦 Packages • 🌐 Webapp • 📙 Documentation • 🔍 License

🐍 Installation

[!IMPORTANT] Sinapsis projects requires Python 3.10 or higher.

This repo includes packages for performing object detection using different models:

• sinapsis-dfine
• sinapsis-rfdetr

Install using your package manager of choice. We strongly encourage the use of uv. If you need to install uv please see the official documentation.

Example with uv:

  uv pip install sinapsis-dfine --extra-index-url https://pypi.sinapsis.tech

or with raw pip:

  pip install sinapsis-dfine --extra-index-url https://pypi.sinapsis.tech

Change the name of the package for the one you want to install.

[!IMPORTANT] Templates in each package may require extra dependencies. For development, we recommend installing the package with all the optional dependencies:

with uv:

  uv pip install sinapsis-dfine[all] --extra-index-url https://pypi.sinapsis.tech

or with raw pip:

  pip install sinapsis-dfine[all] --extra-index-url https://pypi.sinapsis.tech

Change the name of the package accordingly.

[!TIP] You can also install all the packages within this project:

  uv pip install sinapsis-object-detection[all] --extra-index-url https://pypi.sinapsis.tech

📦 Packages

This repository is organized into modular packages, each built for integration with different object detection models. These packages offer ready-to-use templates for training and performing inference with advanced models. Below is an overview of the available packages:

Sinapsis D-FINE

The package provides templates for training and inference with the D-FINE model, enabling advanced object detection tasks. It includes:

• DFINETraining: A template that implements the training pipeline for the D-FINE model, including logic for initializing configurations, downloading weights, and setting up the training solver.
• DFINEInference: A template designed for performing inference on a set of images using the different D-FINE architectures available.

For specific instructions and further details, see the README.md.

Sinapsis RF-DETR

The package provides templates for training, inference, and export with the RF-DETR model, enabling advanced object detection tasks. It includes:

• RFDETRExport and RFDETRLargeExport: Templates for exporting the RFDETRBase and RFDETRLarge models to ONNX format.
• RFDETRInference and RFDETRLargeInference: Templates designed to perform inference on a set of images using the RFDETRBase and RFDETRLarge models.
• RFDETRTrain and RFDETRLargeTrain: Templates for training the RFDETRBase and RFDETRLarge models.

For specific instructions and further details, see the README.md.

🌐 Webapp

The webapps included in this project demonstrate the modularity of the templates, showcasing the capabilities of various object detection models for different tasks.

[!IMPORTANT] To run the app, you first need to clone this repository:

git clone [e-mail]:Sinapsis-ai/sinapsis-object-detection.git
cd sinapsis-object-detection

[!NOTE] If you'd like to enable external app sharing in Gradio, export GRADIO_SHARE_APP=True

[!NOTE] Agent configuration can be changed through the AGENT_CONFIG_PATH env var. You can check the available configurations in each package configs folder.

[!NOTE] When running the app with the D-FINE model, it defaults to a confidence threshold of 0.5, uses CUDA for acceleration, and employs the nano-sized D-FINE model trained on the COCO dataset. These settings can be customized by modifying the demo.yml file inside the configs directory of the sinapsis-dfine package and restarting the webapp.

🐳 Docker

IMPORTANT This docker image depends on the sinapsis-nvidia:base image. Please refer to the official sinapsis instructions to Build with Docker.

1. Build the sinapsis-object-detection image:

docker compose -f docker/compose.yaml build

2. Start the app container:

docker compose -f docker/compose_apps.yaml up sinapsis-dfine-gradio -d

NOTE: You can also deploy the service for the RF-DETR package using

docker compose -f docker/compose_apps.yaml up sinapsis-rfdetr-gradio -d

3. Check the status:

docker logs -f sinapsis-dfine-gradio

NOTE: If using the RF-DETR package, please change the name of the service accordingly

4. The logs will display the URL to access the webapp, e.g.:

Running on local URL:  http://127.0.0.1:7860

NOTE: The url can be different, check the output of logs.

5. To stop the app:

docker compose -f docker/compose_apps.yaml down

💻 UV

To run the webapp using the uv package manager, please:

1. Create the virtual environment and sync the dependencies:

uv sync --frozen

2. Install the sinapsis-object-detection package:

uv pip install sinapsis-object-detection[all] --extra-index-url https://pypi.sinapsis.tech

3. Run the webapp:

uv run webapps/detection_demo.py

NOTE: To use the RF-DETR model, specify the correct configuration file before running the app

export AGENT_CONFIG_PATH=packages/sinapsis-rfdetr/src/sinapsis_rfdetr/configs/rfdetr_demo.yml

4. The terminal will display the URL to access the webapp, e.g.:

Running on local URL:  http://127.0.0.1:7860

NOTE: The url can be different, check the output of the terminal.

📙 Documentation

Documentation for this and other sinapsis packages is available on the sinapsis website

Tutorials for different projects within sinapsis are available at sinapsis tutorials page

🔍 License

This project is licensed under the AGPLv3 license, which encourages open collaboration and sharing. For more details, please refer to the LICENSE file.

For commercial use, please refer to our official Sinapsis website for information on obtaining a commercial license.