English|中文
This warehouse contains a variety of classification samples for users' reference. The directory structure and specific instructions are as follows.
-
googlenet series sample
Sample name Sample description Characteristic analysis support chip googlenet_imagenet_picture Picture classification example Input and output are pictures Ascend310 googlenet_imagenet_multi_batch Picture classification example The input is a bin file, and the output is the result of inference, using the feature of multiple batches Ascend310 googlenet_imagenet_dynamic_batch Picture classification example The input is a bin file, and the output is the result of inference, using the characteristics of dynamic batch性 Ascend310 googlenet_imagenet_video Video classification example Input mp4 file, output as presenter interface display Ascend310 -
resnet50 series sample
Sample name Sample description Characteristic analysis support chip resnet50_imagenet_classification Picture classification example Realize the function of image classification (synchronous reasoning) based on Caffe ResNet-50 network (single input, single batch) Ascend310,Ascend310P,Ascend910 resnet50_async_imagenet_classification Picture classification example Realize the function of image classification (asynchronous reasoning) based on Caffe ResNet-50 network (single input, single batch) Ascend310,Ascend310P,Ascend910 vdec_resnet50_classification Picture classification example Image classification based on Caffe ResNet-50 network (video decoding + synchronous reasoning) Ascend310,Ascend310P,Ascend910 vpc_jpeg_resnet50_imagenet_classification Picture classification example Realize image classification based on Caffe ResNet-50 network (image decoding + matting zoom + image encoding + synchronous reasoning) Ascend310,Ascend310P,Ascend910 vpc_resnet50_imagenet_classification Picture classification example Image classification based on Caffe ResNet-50 network (picture decoding + scaling + synchronous reasoning) Ascend310,Ascend310P,Ascend910