Ora Computing
Making AI models smaller and faster for any hardware
- Category
- AI & Machine Learning
- Launched
- January 1, 2024
- Stage
- Seed
- Pricing
- enterprise
- Founder
- Stefan Sack
- Website
- oracomputing.com
Ora Computing provides an automated AI model compression platform that reduces the size and inference cost of large language models (LLMs) and other foundation models. Their technology combines pruning, quantization, and information theoretic methods to shrink memory footprint by up to 70% and increase throughput by up to 4x, with minimal accuracy loss. The compressed models are compatible with popular runtimes like vLLM and llama.cpp, and can be deployed on edge devices, on prem servers, or cloud GPUs.
Discovered via