Corerain
-
Technologies
- CAISA? 3.0
- CAISA Architecture
- RainBuilder?
- RainBuilder Toolchain
-
Products
- Chips Based on
CAISA Architecture - Streaming AI Chip
- Nebula Accelerator X3
- Nebula Accelerator NA-100c
- Rainman Accelerator
- Chips Based on
- Cases
- Institute
- Community
- Join Us
-
- 中
The head background
The head background
Specially designed for the CAISA architecture, RainBuilder provides rapid deployment for
deep learning algorithms in just two steps. It supports TensorFlow, Caffe, PyTorch, ONNX
and most deep learning frameworks with a versatile and user-friendly developing environment.
RainBuilder
Compiler
- Compatible with mainstream deep learning frameworks
- Enhanced automatic compiling network models
- Automatic quantitative compression, node fusion, memory optimization
RainBuilder
Runtime
- Simplifies development process
- Provides C ++ / Python API for expansion
- Provides performance model analysis, accuracy verification, automatic thread scheduling, etc.
RainBuilder
Driver
- Support CAISA hardware driver, software-hardware scheduling, and streaming graph analysis
- Straightforward, automatic performance optimization and scheduling
Technical Application
RainBuilder provides technical support for rapid landing of AI applications
