Collaborative Production Cycle
for Deep Learning


Gather all your Machines & GPUs,
Keep your Data and Models at Home

Open Source Server
Open Source platform
Private data

Easy Setup

Simple start and API, ready for

# Setup platform directory
export DD_PLATFORM=$HOME/deepdetect
export ARCH=cpu # gpu also available

# Install platform repository
curl -fsSL \
     -o && sh

# Go to directory
cd ${DD_PLATFORM}/code/cpu/

# Start platform docker containers
CURRENT_UID=$(id -u):$(id -g) MUID=$(id -u) \
 docker-compose up -d

# Go to http://localhost:1912
# Platform update
cd $DD_PLATFORM/code/${ARCH}

# Update containers
Jupyter Notebooks

Build and test datasets from Jupyter notebooks. Auto-previsualization for all tasks, from tabular data to text and images. Dataset validation is automated and best hyperparameter presets are provided for a variety of models. Fine-grained control remains available for experts.

Train Models for Images, Text,
Audio & Tabular Data

50+ pre-trained models for very quick transfer training convergence

Distributed Training & Monitoring

Train and monitor live metrics. Distribute jobs on one or more GPUs. Archive all experiments, compare results and publish best models to production safely.

Model Archives & Metrics

Keep track of all experiments, publish final models, visualize metrics and model outputs for images, audio, text, tabular data and time-series.

Quickly Validate & Integrate Models
into Your Applications

Visualize metrics, resources and model outputs

Deep Learning Models Library

Quickly put trained models into production. Monitor GPU usage, add & delete model services as required for your applications.

Fast & Secure Application Integration

Test & Verify all model predictions, test sets and service output. Easy integration via Copy & Paste of API code snippets. Shell, Python and Javascript ready-to-use pre-parametrized code samples.

Deep Learning in Production from Cloud to Edge

Export models for Cloud, Desktop and Embedded devices alike. Tuned performances and models for each architecture.

Uncompromised Performances on GPU and CPU

From Cloud to Desktop and Edge, the best in-production performances. Full C++-11 Open Source server eases deployment with top performances from virtual to bare-metal.