1*4882a593SmuzhiyunThe directory structure of examples is as follows: 2*4882a593Smuzhiyun. 3*4882a593Smuzhiyun├── caffe 4*4882a593Smuzhiyun│ ├── mobilenet_v2 # mobilenet_v2 float model 5*4882a593Smuzhiyun│ └── vgg-ssd # vgg-ssd float model 6*4882a593Smuzhiyun├── onnx 7*4882a593Smuzhiyun│ ├── resnet50v2 # resnet50v2 float model 8*4882a593Smuzhiyun│ └── yolov5 # yolov5 float model 9*4882a593Smuzhiyun├── pytorch 10*4882a593Smuzhiyun│ ├── resnet18 # resnet18 float model 11*4882a593Smuzhiyun│ ├── resnet18_qat # resnet18 QAT model 12*4882a593Smuzhiyun│ └── resnet18_export_onnx # how to export onnx model from pytorch 13*4882a593Smuzhiyun├── tensorflow 14*4882a593Smuzhiyun│ ├── ssd_mobilenet_v1 # ssd_mobilenet_v1 float model 15*4882a593Smuzhiyun│ └── inception_v3_qat # inception_v3 QAT model 16*4882a593Smuzhiyun├── tflite 17*4882a593Smuzhiyun│ ├── mobilenet_v1 # mobilenet_v1 float model 18*4882a593Smuzhiyun│ └── mobilenet_v1_qat # mobilenet_v1 QAT model 19*4882a593Smuzhiyun├── darknet 20*4882a593Smuzhiyun│ └── yolov3_416x416 # yolov3 float model 21*4882a593Smuzhiyun└── functions 22*4882a593Smuzhiyun ├── accuracy_analysis # how to use accuracy-analysis function 23*4882a593Smuzhiyun ├── batch_size # how to expand batch for use multi-batch function 24*4882a593Smuzhiyun ├── multi_input_test # multi-input float model 25*4882a593Smuzhiyun ├── hybrid_quant # how to use hybrid-quantization function 26*4882a593Smuzhiyun ├── mmse # how to use mmse function 27*4882a593Smuzhiyun ├── model_pruning # how to use model_pruning function 28*4882a593Smuzhiyun ├── dynamic_input # how to use dynamic_input function 29*4882a593Smuzhiyun └── board_test # how to connect the board for debugging 30