- ...
- I Load net complete...
- I Load data...
- I Fitting image with scale.
- I Reorder channels.
- I Channel mean value [104.04, 113.985, 119.85, 71.0]
- I Quantization start...
- I Init validate tensor provider.
- I Enqueue samples 5000
- I Init provider with 5000 samples.
- I Load quantization tensor table
- D Optimizing network with qnt_hybrid_insert_layer
- D set up a quantize net
- D Process data_0 ...
- D RKNN output shape(input): (100 512 512 3)
- D Real output shape: (100, 512, 512, 3)
- D Process conv1_1 ...
- D RKNN output shape(convolution): (100 512 512 16)
- D Real output shape: (100, 512, 512, 16)
- D Process relu1_4 ...
- D RKNN output shape(relu): (100 512 512 16)
- D Real output shape: (100, 512, 512, 16)
- D Process conv2_5 ...
- D RKNN output shape(convolution): (100 512 512 16)
- D Real output shape: (100, 512, 512, 16)
- D Process relu2_8 ...
- D RKNN output shape(relu): (100 512 512 16)
- D Real output shape: (100, 512, 512, 16)
- D Process conv3_9 ...
- D RKNN output shape(convolution): (100 256 256 32)
- D Real output shape: (100, 256, 256, 32)
- D Process relu3_12 ...
- D RKNN output shape(relu): (100 256 256 32)
- D Real output shape: (100, 256, 256, 32)
- D Process conv5_17 ...
- D RKNN output shape(convolution): (100 128 128 64)
- D Real output shape: (100, 128, 128, 64)
- D Process relu4_20 ...
- D RKNN output shape(relu): (100 128 128 64)
- D Real output shape: (100, 128, 128, 64)
- ...
复制代码| 欢迎光临 Toybrick (https://t.rock-chips.com/) | Powered by Discuz! X3.3 |