The Data-Centric AI
Modulos accelerates your AI delivery for data science and machine learning teams at leading enterprises. Data-Centric AI identifies the sources of noise and bias in your data which limit the accuracy and fairness of your models.
How does Modulos work?
- 1. Upload your dataset
Start by uploading your dirty datasets, which can have missing or incorrect values.
- 2. Find and Train ML Solutions
Use Modulos AutoML to train the best Solutions.
- 3. Select ML Solution
Choose the best Solution you wish to improve iteratively.
- 4. Assess Performance
Evaluate scores like Accuracy or Fairness and benchmark your Solution.
- 5. Improve Data Quality
Get prioritised recommendations on which samples to clean for reaching your objective.
- 6. Retrain ML Solution
Retrain your Solution using the improved data.
- 7. Deploy ML Solution
Download a deployment-ready solution and scale it.
What our customers think
Deployment & Requirements
We know that data security and deployment flexibility is essential. That’s why you can install Modulos anywhere you need: whether in the public cloud, virtual private cloud, or in your data center.
In order to run an instance of Modulos, you need:
- OS: Recent Linux OS (recmd. Ubuntu 20.04)
- CPU: 4 cores or more (recmd. at least i5 or equiv.)
- Memory: 16 GB or more
- Storage: 500 Gb or more (recmd. SSD)
- Docker CE v19.03 (API v1.40) or v20.10 (API v1.41)
- Python v3.8, v3.9 or v3.10
- Internet connection
Additionally, for a GPU instance (if applicable):
- NVIDIA CUDA 11.2
- NVIDIA GPUs with a compatible driver with CUDA 11.2