Reproducible Research Platform Player Bundle for Deep learning, ML and LLM Player Bundle

RRP Deep learning, ML and LLM

Reproducible Research Platform (RRP) project for deep learning or machine learning applications.

Overview

This RRP project repository demonstrates how to train a deep learning model for cell segmentation of microscopy data mounted from the research data management system (RDMS) openBIS ELN-LIMS. Microscopy data is managed in the RDMS and annotated with respective metadata. Using their permID the data is directly available in the RRP project for analysis and can be integrated in reproducible workflows.

The main outcome of this example is:

Training a DL model for image segmentation.

Visualization of the model predictions.

Example of LLM assisting analysis.

Features

Integration with RRP and openBIS ELN--LIMS: mounts data (microscopy) from the research data management system openBIS in RRP for direct use.

Reproducible workflow: Notebooks document interactions.

Visualization: Generates visualizations of the model predictions.

Installation

Prerequisites

Local Docker installation.

Steps

Download and extract the .zip file

Run the play script for your operating system

Linux / macOS: play

Windows: play.bat or play.ps1

Usage

In RRP open and run the Jupyter notebook run_model_training.ipynb and after the training the run_LLM_assistance.ipynb

Workflow Overview

Training: Train a DL model for image segmentation

Visualization: Visualize model predictions

LLM assistance: Example how LLM can assist analysis.

Directory Structure

├── .binder/ # Environment specifications, Python version, packages (qute) etc.
├── .rrp/ # Specification of openBIS server and required data from RDMS openBIS ELN-LIMS (Microscopy data)
├── cell_segmentation_demo_unet.py # Python script for segmentation model training
├── cell_segmentation_demo_unet_config.ini # Config file for segmentation model training
├── prepare_data.sh # Prepares image data for deep learning
├── LICENSE # License
├── README.md # This file
├── run_LLM_assistance.ipynb # Notebook demonstrating LLM help for image analysis
└── run_model_training.ipynb # Notebook to train the DL model

Contributing

We welcome contributions to improve this repository! Please follow these steps:

Fork the repository.

Create a new branch for your feature or bug fix.

Submit a pull request with a clear description of your changes.

Authors and acknowledgment

Andreas P. Cuny, ETH Zurich initial implementation

License

This project is licensed under the Apache 2.0 License. See LICENSE file for details.

Copyright © 2020-2025 ETH Zurich, Andreas P. Cuny, D-BSSE, CSB Group