NVIDIA TensorRT 10.0 offers new features like weight streaming, weight-stripped engines, INT4 quantization, and improved memory allocation. It also includes Model Optimizer for model optimizations.
•6m read time• From developer.nvidia.com
Sort:
NVIDIA TensorRT 10.0 offers new features like weight streaming, weight-stripped engines, INT4 quantization, and improved memory allocation. It also includes Model Optimizer for model optimizations.
Sort: