MERA Compiler and Software Framework
Supported Frameworks & Applications
MERA Compiler and Software Framework
EdgeCortix® MERA is a compiler and software framework providing a robust platform for deploying the latest neural network models in a framework agnostic manner. MERA enables deep neural network graph compilation and AI inference using the Dynamic Neural Accelerator (DNA) and provides the necessary tools, APIs, code-generator, and runtime needed to deploy a pre-trained deep neural network.
MERA is designed to handle the most challenging Generative AI applications at the edge, enabling designers to create new content for use in Vision, Language, Audio, and many other applications.
With MERA’s built-in heterogeneous support for leading processors, including AMD, Intel, Arm, and RISC-V, it is fast and easy to integrate the EdgeCortix AI platform into existing systems. Pre-defined and optimized models can be directly sourced from Hugging Face or EdgeCortix Model Library, followed by calibration and quantization. MERA offers software developers and data scientists familiar workflows for developing these models without dealing with chip-level hardware architecture. MERA leverages Apache TVM and MLIR functionality, and its front end is open sourced for easy distribution to project teams.
Integrated in familiar AI inference workflows
MERA provides the entire stack for edge AI inferencing from modeling to deployment with familiar neural network model workflows and supports easy integration with existing systems, reducing time-to-market. MERA is optimized for the latest vision and Generative AI models with market-leading energy efficiency and low latency in combination with SAKURA-II.
MERA is a compiler and software framework enabling deep neural network graph compilation and AI inference using the Dynamic Neural Accelerator (DNA). MERA provides the necessary tools, APIs, code-generator and runtime needed to deploy a pre-trained deep neural network after a simple calibration and quantization step.
MERA Key Benefits
- Source pre-defined and optimized models from Hugging Face and the EdgeCortix Model Library
- Generative AI Models - Llama-2, Stable Diffusion, Whisper, DETR, DistillBert, DINO and ViT
- Supports Python and C++ interfaces for workflow integration and customization
- Native support for ML frameworks PyTorch, TensorFlow Lite, and ONNX
- Post-training calibration and quantization of user-defined models
- Stand-alone estimation of inference latency and throughput
Everything needed for AI inference development and deployment
- Runtime configuration of DNA
- Simulation and profiling for stand-alone use
- Low-level IR (Intermediate Representation)
- Host processor code generation
- High-level graph partitioning
- APIs for popular machine learning frameworks
- Model calibration and quantization
- Pre-trained applications in the Model Library
MERA Software Supports Diverse Neural Networks from Convolutions to the Latest Generative AI Models
Transformer Models | Convolutional Models | ||
DETR DINO Whisper Encoder / Decoder DistillBERT DistilBert - SST2 Nano-GPT GPT-2 - 150M Distil-GPT-2 (HF) GPT-2 (HF) - 117M GPT-2 (HF) - medium / large GPT-2 - XL (HF) - 1.5B |
TinyLama (HF) - 1.1B Phi-2 (HF) - 3B Open-Llama2 (HF) - 7B CodeLlama (HF) - 7B Mistral-v0.2 (HF) - 7B Llama3 - 8B ViT (HF) / CLIP / Mobile-ViT ConvNextV1/V2 (HF) SegFormer Roberta-Emotion StableDiffusion V1.5 |
ResNet 18 ResNet 50/101 Big YoloV3 TinyYolo V3 Yolo V5/V6/V8 YoloX EfficientNet-Lite EfficientNet-V2 SFA3D |
MonoDepth - MiDaS U-Net MoveNet DeepLab MobileNet V1-V2 MobileNetV2-SSD GladNet ABPN SCI |
Transformer Models |
DETR DINO Whisper Encoder / Decoder DistillBERT DistilBert - SST2 Nano-GPT GPT-2 - 150M Distil-GPT-2 (HF) GPT-2 (HF) - 117M GPT-2 (HF) - medium / large GPT-2 - XL (HF) - 1.5B TinyLama (HF) - 1.1B Phi-2 (HF) - 3B Open-Llama2 (HF) - 7B CodeLlama (HF) - 7B Mistral-v0.2 (HF) - 7B Llama3 - 8B ViT (HF) / CLIP / Mobile-ViT ConvNextV1/V2 (HF) SegFormer Roberta-Emotion StableDiffusion V1.5 |
Convolutional Models |
ResNet 18 ResNet 50/101 Big YoloV3 TinyYolo V3 Yolo V5/V6/V8 YoloX EfficientNet-Lite EfficientNet-V2 SFA3D MonoDepth - MiDaS U-Net MoveNet DeepLab MobileNet V1-V2 MobileNetV2-SSD GladNet ABPN SCI |
SAKURA-II M.2 Modules and PCIe Cards
EdgeCortix SAKURA-II can be easily integrated into a host system for software development and AI model inference tasks.
Pre-Order an M.2 Module or a PCIe Card and get started today!
EdgeCortix Edge AI Platform
Proprietary Architecture
Efficient Hardware
Deployable Systems
Given the tectonic shift in information processing at the edge, companies are now seeking near cloud level performance where data curation and AI driven decision making can happen together. Due to this shift, the market opportunity for the EdgeCortix solutions set is massive, driven by the practical business need across multiple sectors which require both low power and cost-efficient intelligent solutions. Given the exponential global growth in both data and devices, I am eager to support EdgeCortix in their endeavor to transform the edge AI market with an industry-leading IP portfolio that can deliver performance with orders of magnitude better energy efficiency and a lower total cost of ownership than existing solutions."
Improving the performance and the energy efficiency of our network infrastructure is a major challenge for the future. Our expectation of EdgeCortix is to be a partner who can provide both the IP and expertise that is needed to tackle these challenges simultaneously."
With the unprecedented growth of AI/Machine learning workloads across industries, the solution we're delivering with leading IP provider EdgeCortix complements BittWare's Intel Agilex FPGA-based product portfolio. Our customers have been searching for this level of AI inferencing solution to increase performance while lowering risk and cost across a multitude of business needs both today and in the future."
EdgeCortix is in a truly unique market position. Beyond simply taking advantage of the massive need and growth opportunity in leveraging AI across many business key sectors, it’s the business strategy with respect to how they develop their solutions for their go-to-market that will be the great differentiator. In my experience, most technology companies focus very myopically, on delivering great code or perhaps semiconductor design. EdgeCortix’s secret sauce is in how they’ve co-developed their IP, applying equal importance to both the software IP and the chip design, creating a symbiotic software-centric hardware ecosystem, this sets EdgeCortix apart in the marketplace.”
We recognized immediately the value of adding the MERA compiler and associated tool set to the RZ/V MPU series, as we expect many of our customers to implement application software including AI technology. As we drive innovation to meet our customer's needs, we are collaborating with EdgeCortix to rapidly provide our customers with robust, high-performance and flexible AI-inference solutions. The EdgeCortix team has been terrific, and we are excited by the future opportunities and possibilities for this ongoing relationship."