HaiLabel
Smart HPC

Overview

HaiLabel-Detector provides a convenient solution for AI annotation tasks in object detection and classification development. The solution significantly reduces the human workload on data annotation and AI training. Users can easily create custom AI models with image or video data and have these models directly available via our Model Server API to use in your applications. Additionally, users can implement a continuous training pipeline to ensure continual improvement with new data.

Benefits

Annotation Speed-Up

AI Annotation Speed-Up

HailLabe improves the efficacy of dataset preparation during the AI model development cycle, providing up to x10 improvement for the dataset annotation workflows.

Model Re-Training

Our AiON cloud supports model training on demand where users can easily manage resources, scheduling and scaling.

Automatic Retraining

Our Automatic Retraining pipelines allow for stress free model development. Users can set frequency and data thresholds to trigger AI training automatically when new data is received.

Model Management

AI Model Management

Model Management serves as the AI repository and provides features for AI model version management and control to EDGE devices.

AI Model Selection

HailLabel currently support TensorFlow and PyTorch models. Our team is consistently working towards expanding our model repository!

AI Server

Users can manage their AI models and use them via our REST-API, reducing the resource constraints to serve a variety of models architectures while allowing for flexibility for versioning and application deployment.

Products

HPC

Provides high-speed CPU/GPU resources, an incredibly-quick InfiniBand Network, and Lustre shared storage with high capacity for any application.