SyNAP Manual

• Introduction
  • Online Model Conversion
  • Offline Model Conversion
  • NPU Hardware Capabilities
• Getting Started
  • synap_cli_ic application
  • synap_cli_od application
  • synap_cli_ip application
  • synap_cli_ic2 application
  • synap_cli application
  • synap_init application
  • Troubleshooting
• Using Online Inference
  • Online Inference with NNAPI
  • Benchmarking Models with NNAPI
  • NNAPI Compilation Caching
  • Disabling NPU Usage from NNAPI
  • Online Inference with TimVx delegate
  • Benchmarking Models with TimVx delegate
• Reference Models
  • Timings
    • Inference timings on VS680 and SL1680
    • Inference timings on VS640 and SL1640
  • Super Resolution
  • Format Conversion
• Statistics and Usage
  • inference_count
  • inference_time
  • networks
  • network_profile
  • Clearing statistics
  • Using /sysfs information
• Working with Models
  • Model Conversion
  • Installing Docker
    • Linux/Ubuntu
    • macOS - Docker
    • macOS - Colima
    • Windows
  • Installing SyNAP Tools
  • Running SyNAP Tools
  • Installing pip
  • Running pip Console Scripts
  • Conversion Metafile
  • Preprocessing
    • type^(*)
    • size
    • crop
  • Model Quantization
    • Quantization Images Resize
    • Data Normalizaton
    • Quantization and Accuracy
    • Per-Channel Quantization
    • Mixed Quantization
  • Heterogeneous Inference
  • Model Conversion Tutorial
  • Model Profiling
  • Compatibility with SyNAP 2.x
  • Working with PyTorch Models
    • Importing YOLO PyTorch Models
• Framework API
  • Basic Usage
    • Network Class
      • Network
        • Network::load_model()
        • Network::load_model()
        • Network::predict()
        • Network::inputs
        • Network::outputs
    • Using a Network
  • Advanced Topics
    • Tensors
      • Tensor
        • Tensor::name()
        • Tensor::shape()
        • Tensor::dimensions()
        • Tensor::layout()
        • Tensor::format()
        • Tensor::data_type()
        • Tensor::security()
        • Tensor::size()
        • Tensor::item_count()
        • Tensor::is_scalar()
        • Tensor::assign()
        • Tensor::assign()
        • Tensor::assign()
        • Tensor::assign()
        • Tensor::assign()
        • Tensor::assign()
        • Tensor::data()
        • Tensor::data()
        • Tensor::as_float()
        • Tensor::buffer()
        • Tensor::set_buffer()
        • Tensor::Private
      • DataType
        • DataType::invalid
        • DataType::byte
        • DataType::int8
        • DataType::uint8
        • DataType::int16
        • DataType::uint16
        • DataType::int32
        • DataType::uint32
        • DataType::float16
        • DataType::float32
    • Buffers
      • Buffer
        • Buffer::Buffer()
        • Buffer::Buffer()
        • Buffer::Buffer()
        • Buffer::Buffer()
        • Buffer::Buffer()
        • Buffer::Buffer()
        • Buffer::operator=()
        • Buffer::resize()
        • Buffer::assign()
        • Buffer::size()
        • Buffer::data()
        • Buffer::allow_cpu_access()
        • Buffer::set_allocator()
    • Allocators
      • std_allocator()
      • cma_allocator()
  • Advanced Examples
    • Accessing Tensor Data
    • Setting Buffers
    • Settings Default Buffer Properties
    • Buffer Sharing
    • Recycling Buffers
    • Using BufferCache
    • Copying and Moving
  • NPU Locking
    • NPU Locking
    • NNAPI Locking
    • Description
      • Npu
        • Npu::available()
        • Npu::lock()
        • Npu::unlock()
        • Npu::is_locked()
        • Npu::Private
    • Sample Usage
  • Preprocessing and Postprocessing
    • InputData Class
    • Preprocessor Class
    • ImagePostprocessor Class
    • Classifier Class
      • Classifier
        • Classifier::Classifier()
        • Classifier::process()
        • Classifier::Result
          • Classifier::Result::success
          • Classifier::Result::items
          • Classifier::Result::Item
            • Classifier::Result::Item::class_index
            • Classifier::Result::Item::confidence
    • Detector Class
      • Detector
        • Detector::Detector()
        • Detector::init()
        • Detector::process()
        • Detector::Impl
        • Detector::Result
          • Detector::Result::success
          • Detector::Result::items
          • Detector::Result::Item
            • Detector::Result::Item::class_index
            • Detector::Result::Item::confidence
            • Detector::Result::Item::bounding_box
            • Detector::Result::Item::landmarks
            • Detector::Result::Item::mask
  • Building Sample Code
• Neural Network Processing Unit Operator Support
  • Basic Operations
  • Activation Operations
  • Elementwise Operations
  • Normalization Operations
  • Reshape Operations
  • RNN Operations
  • Pooling Operations
  • Miscellaneous Operations
• Direct access in Android Applications