TIDL network import tool and graph visualizer tool Support for Additional layers - Sigmoid layer, Standalone Pad layer, Color conversion (YUV420SP to RGB) Layer, Depth to Space (Pixel Shuffle) LayerĬustom / User defined layer Call-back to plugin user specific layers Supported layers: Convolution, deconvolution, pooling, ReLU, Elt-wise, Inner product, Soft-max, and many more, refer TIDL user guide ( LINK) Supports Caffe, ONXX, TFlite network exchange formats
MMALIB for accelerating convolution, deconvolution layers on MMA (used by TIDL internally) TIDL deep learning inference engine library running on C7x/MMA NEO-AI-DLR support - TVM model compilation for heterogeneous execution on ARM+C7x_MMA ONNX Runtime with Execution Provider support for heterogeneous execution on ARM+C7x_MMA TensorFlow Lite Runtime with Delegate API support for heterogeneous execution on ARM+C7x_MMA
TI Deep learning Product (TIDL) and MMALIB ¶
Timer utility 4.1.4 free#
Profiling, logging and other utility APIsĪll included MCU tasks are split between MCU2_0 (ETHFW, VPAC HWA) and MCU2_1 (DMPAC HWA), leaving MCU3_0 and MCU3_1 free on main islandĤ.1.1.2.4. Lidar, radar, camera SFM nodes for PTK demos Image post-processing and visualization nodes for deep learning, DOF, Stereo demos Image pre-processing (YUV 2 RGB) nodes for DL demos Single and multi camera to display (CSI2RX + VISS + LDC + MSC + Display) demoīlock based processing example kernel on C66x using UDMA/DRUĬ7x algorithm offload demo (with DMA acceleration)Ĭamera based SFM and occupancy grid mapping using Dense optical flow HW engineĬamera SFM, radar, lidar fusion for occupancy grid mapping Semantic segmentation, Parking spot detect and Vehicle detect algorithmsģ channel, 3 algo mode - combined network for parking spot and vehicle detect (AVP2 demo)ģ channel, 5 algo mode - combined network for parking spot and vehicle detect + Three task network for Semantic Segmentation, Motion Segmentation and Depth Estimation (AVP3 demo)ĭKAZE-based Visual Localization applicationĭebug support for the ability to debug intermediate layer information from TIDL nodeģD surround view using GPU on 4x 2MP 30fps live camera inputģD Surround view + AVP2 demo running simultaneouslyģD surround view using GPU on 4x2MP 30fps live camera input
Standalone Semantic Segmentation application Integrates all major HW components like CSI2 camera, eDP/HDMI display, UART, I2C, ethernet, SD card Integrates all major PSDK RTOS and PSDK LINUX SW components like TIDL, MMALIB, PDK, OpenVX, ETHFW, OpenGL, video codec, C6x algorithms, imaging/sensors, IPC, Linux, TI-RTOS, SPL/uboot. TI-RTOS on C7x, C6x, R5F and Linux on A72 OpenVX based demos for ADAS, Vision, deep learning applications
0 kommentar(er)
