MLComp

MLComp is a distributed DAG (Directed acyclic graph) framework for machine learning with UI.

The goal of MLComp is to provide tools for training, inference, creating complex pipelines (especially for computer vision) in a rapid, well manageable, way.

MLComp is compatible with: Python 3.6+, Unix operation system.

Features

• Amazing UI

• Catalyst support

• Distributed training

• Supervisor that controls computational resources

• Synchronization of both code and data

• Resource monitoring

• Full-functionally pause and continue on UI

• Auto control of the requirements

• Code dumping(with syntax highlight on UI)

• Kaggle integration

• Hierarchical logging

• Grid search

• Experiments comparison

• Customizing layouts

Overview:

• MLComp
• Installation
• Usage
• Layout
• Grid search
• FileSync