root@7552a77f8bdb:/srv/keras_compressor/example/mnist# python train.py
Using TensorFlow backend.
Downloading data from https://s3.amazonaws.com/img-datasets/mnist.npz
11493376/11490434 [==============================] - 3s 0us/step
11501568/11490434 [==============================] - 3s 0us/step
Train on 60000 samples, validate on 10000 samples
Epoch 1/10
2019-02-26 12:21:58.576138: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
60000/60000 [==============================] - 113s 2ms/step - loss: 0.2448 - acc: 0.9252 - val_loss: 0.0556 - val_acc: 0.9822
Epoch 2/10
60000/60000 [==============================] - 114s 2ms/step - loss: 0.0871 - acc: 0.9740 - val_loss: 0.0396 - val_acc: 0.9864
Epoch 3/10
60000/60000 [==============================] - 117s 2ms/step - loss: 0.0661 - acc: 0.9807 - val_loss: 0.0361 - val_acc: 0.9880
Epoch 4/10
60000/60000 [==============================] - 124s 2ms/step - loss: 0.0594 - acc: 0.9821 - val_loss: 0.0345 - val_acc: 0.9889
Epoch 5/10
60000/60000 [==============================] - 118s 2ms/step - loss: 0.0536 - acc: 0.9843 - val_loss: 0.0353 - val_acc: 0.9892
Epoch 6/10
60000/60000 [==============================] - 119s 2ms/step - loss: 0.0514 - acc: 0.9854 - val_loss: 0.0337 - val_acc: 0.9888
Epoch 7/10
60000/60000 [==============================] - 118s 2ms/step - loss: 0.0535 - acc: 0.9842 - val_loss: 0.0349 - val_acc: 0.9890
Epoch 8/10
60000/60000 [==============================] - 118s 2ms/step - loss: 0.0530 - acc: 0.9850 - val_loss: 0.0318 - val_acc: 0.9895
Epoch 9/10
60000/60000 [==============================] - 118s 2ms/step - loss: 0.0543 - acc: 0.9844 - val_loss: 0.0341 - val_acc: 0.9901
Epoch 10/10
60000/60000 [==============================] - 118s 2ms/step - loss: 0.0551 - acc: 0.9846 - val_loss: 0.0355 - val_acc: 0.9909
10000/10000 [==============================] - 5s 473us/step
('test accuracy: ', 0.9909)
エラーだった。TypeError: 'NoneType' object is not callable
呼び出すそうだ。backend.clear_session()
root@7552a77f8bdb:/srv/keras_compressor/example/mnist# python3 finetune.py
Using TensorFlow backend.
2019-02-26 13:19:58.359630: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
Train on 60000 samples, validate on 10000 samples
Epoch 1/10
60000/60000 [==============================] - 52s 867us/step - loss: 0.3731 - acc: 0.8887 - val_loss: 0.1206 - val_acc: 0.9643
Epoch 2/10
60000/60000 [==============================] - 52s 864us/step - loss: 0.1599 - acc: 0.9554 - val_loss: 0.0921 - val_acc: 0.9722
Epoch 3/10
60000/60000 [==============================] - 52s 862us/step - loss: 0.1333 - acc: 0.9626 - val_loss: 0.0773 - val_acc: 0.9769
Epoch 4/10
60000/60000 [==============================] - 52s 864us/step - loss: 0.1178 - acc: 0.9670 - val_loss: 0.0797 - val_acc: 0.9769
Epoch 5/10
60000/60000 [==============================] - 52s 867us/step - loss: 0.1074 - acc: 0.9703 - val_loss: 0.0770 - val_acc: 0.9766
Epoch 6/10
60000/60000 [==============================] - 52s 863us/step - loss: 0.0986 - acc: 0.9725 - val_loss: 0.0622 - val_acc: 0.9810
Epoch 7/10
60000/60000 [==============================] - 52s 869us/step - loss: 0.0958 - acc: 0.9729 - val_loss: 0.0650 - val_acc: 0.9813
Epoch 8/10
60000/60000 [==============================] - 52s 865us/step - loss: 0.0930 - acc: 0.9740 - val_loss: 0.0644 - val_acc: 0.9792
Epoch 9/10
60000/60000 [==============================] - 52s 865us/step - loss: 0.0875 - acc: 0.9756 - val_loss: 0.0607 - val_acc: 0.9829
Epoch 10/10
60000/60000 [==============================] - 52s 864us/step - loss: 0.0835 - acc: 0.9770 - val_loss: 0.0741 - val_acc: 0.
10000/10000 [==============================] - 2s 202us/step
test accuracy: 0.9793
FROM gw000/keras
RUN apt-get update && apt-get install -y vim
RUN apt-get install -y python3-pip
RUN pip3 install keras
RUN pip3 install tensorflow
RUN pip3 install scikit-learn
docker build -t keras_compressor .
docker run -ti --rm --name keras_compressorc compressor /bin/bash
を追加する。PYTHONPATH に追加するわけだ。export PYTHONPATH="$PYTHONPATH:/srv/keras_compressor/"
root@f3527f089871:/workspace/distiller/examples/classifier_compression# python3 compress_classifier.py -a resnet20_cifar ../../../data.cifar10 --resume ../ssl/checkpoints/checkpoint_trained_dense.pth.tar --quantize-eval --evaluate
Log file for this run: /workspace/distiller/examples/classifier_compression/logs/2019.02.24-195538/2019.02.24-195538.log
==> using cifar10 dataset
=> creating resnet20_cifar model for CIFAR10
--------------------------------------------------------
Logging to TensorBoard - remember to execute the server:
> tensorboard --logdir='./logs'
=> loading checkpoint ../ssl/checkpoints/checkpoint_trained_dense.pth.tar
best top@1: 92.540
Loaded compression schedule from checkpoint (epoch 179)
=> loaded checkpoint '../ssl/checkpoints/checkpoint_trained_dense.pth.tar' (epoch 179)
Optimizer Type: <class 'torch.optim.sgd.SGD'>
Optimizer Args: {'lr': 0.1, 'momentum': 0.9, 'dampening': 0, 'weight_decay': 0.0001, 'nesterov': False}
Files already downloaded and verified
Files already downloaded and verified
Dataset sizes:
training=45000
validation=5000
test=10000
Preparing model for quantization using PostTrainLinearQuantizer
Parameter 'conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.downsample.0.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.downsample.0.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'fc.wrapped_module.weight' will be quantized to 8 bits
Quantized model:
ResNetCifar(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=67.0449, w_zero_point=0.0000
(wrapped_module): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace)
(layer1): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=107.7180, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=144.9845, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=180.5892, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=183.1618, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=122.4423, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=152.7345, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(layer2): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=181.2943, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=238.1681, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
(downsample): Sequential(
(0): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=88.9728, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 32, kernel_size=(1, 1), stride=(2, 2), bias=False)
)
(1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=206.7310, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=266.5453, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=152.4554, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=240.0073, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(layer3): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=215.9280, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=235.3784, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
(downsample): Sequential(
(0): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=225.9492, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 64, kernel_size=(1, 1), stride=(2, 2), bias=False)
)
(1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=261.7885, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=295.0370, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=212.4505, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=450.7900, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(avgpool): AvgPool2d(kernel_size=8, stride=1, padding=0)
(fc): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=83.8628, w_zero_point=0.0000
base_b_scale=261.5564, base_b_zero_point=0.0000
(wrapped_module): Linear(in_features=64, out_features=10, bias=True)
)
)
--- test ---------------------
10000 samples (256 per mini-batch)
Test: [ 10/ 39] Loss 0.350214 Top1 92.187500 Top5 99.687500
Test: [ 20/ 39] Loss 0.340852 Top1 92.167969 Top5 99.707031
Test: [ 30/ 39] Loss 0.351019 Top1 91.953125 Top5 99.661458
Test: [ 40/ 39] Loss 0.354121 Top1 91.770000 Top5 99.610000
==> Top1: 91.770 Top5: 99.610 Loss: 0.354
Saving checkpoint to: logs/2019.02.24-195538/quantized_checkpoint.pth.tar
Log file for this run: /workspace/distiller/examples/classifier_compression/logs/2019.02.24-195538/2019.02.24-195538.log
root@f3527f089871:/workspace/distiller/examples/classifier_compression#
root@f3527f089871:/workspace/distiller/examples/classifier_compression# python3 compress_classifier.py --resume=resnet20_cifar/quantized_checkpoint.pth.tar -a=resnet20_cifar ../../../data.cifar10 --summary=sparsity
Log file for this run: /workspace/distiller/examples/classifier_compression/logs/2019.02.24-200303/2019.02.24-200303.log
==> using cifar10 dataset
=> creating resnet20_cifar model for CIFAR10
--------------------------------------------------------
Logging to TensorBoard - remember to execute the server:
> tensorboard --logdir='./logs'
=> loading checkpoint resnet20_cifar/quantized_checkpoint.pth.tar
best top@1: 91.770
Loaded compression schedule from checkpoint (epoch 0)
Loaded quantizer metadata from the checkpoint
Preparing model for quantization using PostTrainLinearQuantizer
Parameter 'conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.downsample.0.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.downsample.0.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'fc.wrapped_module.weight' will be quantized to 8 bits
Quantized model:
ResNetCifar(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=350.8756, w_zero_point=0.0000
(wrapped_module): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace)
(layer1): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=312.6898, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=271.3224, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=231.3102, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=301.7677, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=305.3219, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=282.7444, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(layer2): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=390.4178, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=364.4760, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
(downsample): Sequential(
(0): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=167.9384, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 32, kernel_size=(1, 1), stride=(2, 2), bias=False)
)
(1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=366.3052, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=382.9035, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=362.4750, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=391.2636, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(layer3): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=537.1642, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=531.2303, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
(downsample): Sequential(
(0): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=193.8010, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 64, kernel_size=(1, 1), stride=(2, 2), bias=False)
)
(1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=444.0079, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=544.6738, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=524.7949, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=514.5536, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(avgpool): AvgPool2d(kernel_size=8, stride=1, padding=0)
(fc): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=1020.4681, w_zero_point=0.0000
base_b_scale=1211.1595, base_b_zero_point=0.0000
(wrapped_module): Linear(in_features=64, out_features=10, bias=True)
)
)
=> loaded checkpoint 'resnet20_cifar/quantized_checkpoint.pth.tar' (epoch 0)
Optimizer Type: <class 'torch.optim.sgd.SGD'>
Optimizer Args: {'lr': 0.1, 'momentum': 0.9, 'dampening': 0, 'weight_decay': 0.0001, 'nesterov': False}
Parameters:
+----+---------------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+----------+----------+------------+
| | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
|----+---------------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+----------+----------+------------|
| 0 | conv1.wrapped_module.weight | (16, 3, 3, 3) | 432 | 423 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.08333 | 28.75588 | -0.63426 | 20.47685 |
| 1 | layer1.0.conv1.wrapped_module.weight | (16, 16, 3, 3) | 2304 | 2202 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 4.42708 | 16.98713 | -1.12153 | 12.02691 |
| 2 | layer1.0.conv2.wrapped_module.weight | (16, 16, 3, 3) | 2304 | 2260 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.90972 | 22.50954 | -0.29210 | 16.26259 |
| 3 | layer1.1.conv1.wrapped_module.weight | (16, 16, 3, 3) | 2304 | 2255 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.12674 | 24.41840 | -2.58550 | 18.58637 |
| 4 | layer1.1.conv2.wrapped_module.weight | (16, 16, 3, 3) | 2304 | 2260 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.90972 | 23.33428 | -0.94010 | 17.67795 |
| 5 | layer1.2.conv1.wrapped_module.weight | (16, 16, 3, 3) | 2304 | 2247 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.47396 | 21.48796 | -1.00347 | 15.57726 |
| 6 | layer1.2.conv2.wrapped_module.weight | (16, 16, 3, 3) | 2304 | 2248 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.43056 | 21.59341 | 0.07943 | 16.16710 |
| 7 | layer2.0.conv1.wrapped_module.weight | (32, 16, 3, 3) | 4608 | 4528 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.73611 | 25.68778 | -0.13411 | 19.85069 |
| 8 | layer2.0.conv2.wrapped_module.weight | (32, 32, 3, 3) | 9216 | 9076 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.51910 | 28.72585 | -1.76801 | 22.54427 |
| 9 | layer2.0.downsample.0.wrapped_module.weight | (32, 16, 1, 1) | 512 | 502 | 0.00000 | 0.00000 | 0.00000 | 1.95312 | 0.00000 | 1.95312 | 23.43659 | -0.99414 | 16.49414 |
| 10 | layer2.1.conv1.wrapped_module.weight | (32, 32, 3, 3) | 9216 | 9013 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.20269 | 21.76668 | -2.07650 | 17.18414 |
| 11 | layer2.1.conv2.wrapped_module.weight | (32, 32, 3, 3) | 9216 | 9042 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.88802 | 24.62638 | -1.73253 | 19.44021 |
| 12 | layer2.2.conv1.wrapped_module.weight | (32, 32, 3, 3) | 9216 | 8967 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.70182 | 15.97449 | -1.79373 | 12.49447 |
| 13 | layer2.2.conv2.wrapped_module.weight | (32, 32, 3, 3) | 9216 | 9031 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.00738 | 20.47896 | -0.12066 | 16.13780 |
| 14 | layer3.0.conv1.wrapped_module.weight | (64, 32, 3, 3) | 18432 | 18093 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.83919 | 22.38574 | -1.24680 | 17.73107 |
| 15 | layer3.0.conv2.wrapped_module.weight | (64, 64, 3, 3) | 36864 | 36218 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.75239 | 22.71294 | -1.07943 | 18.02051 |
| 16 | layer3.0.downsample.0.wrapped_module.weight | (64, 32, 1, 1) | 2048 | 2020 | 0.00000 | 0.00000 | 0.00000 | 1.36719 | 0.00000 | 1.36719 | 33.22110 | -2.75146 | 26.28564 |
| 17 | layer3.1.conv1.wrapped_module.weight | (64, 64, 3, 3) | 36864 | 36239 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.69542 | 24.10124 | -1.99618 | 19.11982 |
| 18 | layer3.1.conv2.wrapped_module.weight | (64, 64, 3, 3) | 36864 | 36213 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.76595 | 24.02485 | -2.30968 | 18.97195 |
| 19 | layer3.2.conv1.wrapped_module.weight | (64, 64, 3, 3) | 36864 | 35969 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.42784 | 17.23314 | -1.77412 | 13.64255 |
| 20 | layer3.2.conv2.wrapped_module.weight | (64, 64, 3, 3) | 36864 | 36093 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 2.09147 | 22.90270 | -0.40156 | 17.74607 |
| 21 | fc.wrapped_module.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 48.12321 | -0.01875 | 41.37500 |
| 22 | Total sparsity: | - | 270896 | 265539 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 1.97751 | 0.00000 | 0.00000 | 0.00000 |
+----+---------------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+----------+----------+------------+
Total sparsity: 1.98
Log file for this run: /workspace/distiller/examples/classifier_compression/logs/2019.02.24-200303/2019.02.24-200303.log
root@f3527f089871:/workspace/distiller/examples/classifier_compression#
root@f3527f089871:/workspace/distiller/examples/classifier_compression# python3 compress_classifier.py --resume=resnet20_cifar/quantized_checkpoint.pth.tar -a=resnet20_cifar ../../../data.cifar10 --summary=modules
Log file for this run: /workspace/distiller/examples/classifier_compression/logs/2019.02.24-200528/2019.02.24-200528.log
==> using cifar10 dataset
=> creating resnet20_cifar model for CIFAR10
--------------------------------------------------------
Logging to TensorBoard - remember to execute the server:
> tensorboard --logdir='./logs'
=> loading checkpoint resnet20_cifar/quantized_checkpoint.pth.tar
best top@1: 91.770
Loaded compression schedule from checkpoint (epoch 0)
Loaded quantizer metadata from the checkpoint
Preparing model for quantization using PostTrainLinearQuantizer
Parameter 'conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer1.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.0.downsample.0.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer2.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.0.downsample.0.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.1.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.1.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.2.conv1.wrapped_module.weight' will be quantized to 8 bits
Parameter 'layer3.2.conv2.wrapped_module.weight' will be quantized to 8 bits
Parameter 'fc.wrapped_module.weight' will be quantized to 8 bits
Quantized model:
ResNetCifar(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=343.7104, w_zero_point=0.0000
(wrapped_module): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace)
(layer1): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=299.4469, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=316.5102, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=333.9775, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=306.0066, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=304.4463, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=246.7246, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(layer2): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=401.2170, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=405.1727, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
(downsample): Sequential(
(0): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=174.5428, w_zero_point=0.0000
(wrapped_module): Conv2d(16, 32, kernel_size=(1, 1), stride=(2, 2), bias=False)
)
(1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=346.6644, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=407.8266, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=337.9731, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=378.4581, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(layer3): Sequential(
(0): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=504.6201, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=513.4936, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
(downsample): Sequential(
(0): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=212.4980, w_zero_point=0.0000
(wrapped_module): Conv2d(32, 64, kernel_size=(1, 1), stride=(2, 2), bias=False)
)
(1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=526.9995, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=486.5478, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
(2): BasicBlock(
(conv1): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=472.2161, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu1): ReLU()
(conv2): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=459.6506, w_zero_point=0.0000
(wrapped_module): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu2): ReLU()
)
)
(avgpool): AvgPool2d(kernel_size=8, stride=1, padding=0)
(fc): RangeLinearQuantParamLayerWrapper(
mode=SYMMETRIC, num_bits_acts=8, num_bits_params=8, num_bits_accum=32, clip_acts=False, per_channel_wts=False
preset_activation_stats=False
w_scale=1017.0168, w_zero_point=0.0000
base_b_scale=1095.1624, base_b_zero_point=0.0000
(wrapped_module): Linear(in_features=64, out_features=10, bias=True)
)
)
=> loaded checkpoint 'resnet20_cifar/quantized_checkpoint.pth.tar' (epoch 0)
Optimizer Type: <class 'torch.optim.sgd.SGD'>
Optimizer Args: {'lr': 0.1, 'momentum': 0.9, 'dampening': 0, 'weight_decay': 0.0001, 'nesterov': False}
Name Type
------------------------------------ -----------
conv1.wrapped_module Conv2d
bn1 BatchNorm2d
relu ReLU
layer1.0.conv1.wrapped_module Conv2d
layer1.0.bn1 BatchNorm2d
layer1.0.relu1 ReLU
layer1.0.conv2.wrapped_module Conv2d
layer1.0.bn2 BatchNorm2d
layer1.0.relu2 ReLU
layer1.1.conv1.wrapped_module Conv2d
layer1.1.bn1 BatchNorm2d
layer1.1.relu1 ReLU
layer1.1.conv2.wrapped_module Conv2d
layer1.1.bn2 BatchNorm2d
layer1.1.relu2 ReLU
layer1.2.conv1.wrapped_module Conv2d
layer1.2.bn1 BatchNorm2d
layer1.2.relu1 ReLU
layer1.2.conv2.wrapped_module Conv2d
layer1.2.bn2 BatchNorm2d
layer1.2.relu2 ReLU
layer2.0.conv1.wrapped_module Conv2d
layer2.0.bn1 BatchNorm2d
layer2.0.relu1 ReLU
layer2.0.conv2.wrapped_module Conv2d
layer2.0.bn2 BatchNorm2d
layer2.0.relu2 ReLU
layer2.0.downsample.0.wrapped_module Conv2d
layer2.0.downsample.1 BatchNorm2d
layer2.1.conv1.wrapped_module Conv2d
layer2.1.bn1 BatchNorm2d
layer2.1.relu1 ReLU
layer2.1.conv2.wrapped_module Conv2d
layer2.1.bn2 BatchNorm2d
layer2.1.relu2 ReLU
layer2.2.conv1.wrapped_module Conv2d
layer2.2.bn1 BatchNorm2d
layer2.2.relu1 ReLU
layer2.2.conv2.wrapped_module Conv2d
layer2.2.bn2 BatchNorm2d
layer2.2.relu2 ReLU
layer3.0.conv1.wrapped_module Conv2d
layer3.0.bn1 BatchNorm2d
layer3.0.relu1 ReLU
layer3.0.conv2.wrapped_module Conv2d
layer3.0.bn2 BatchNorm2d
layer3.0.relu2 ReLU
layer3.0.downsample.0.wrapped_module Conv2d
layer3.0.downsample.1 BatchNorm2d
layer3.1.conv1.wrapped_module Conv2d
layer3.1.bn1 BatchNorm2d
layer3.1.relu1 ReLU
layer3.1.conv2.wrapped_module Conv2d
layer3.1.bn2 BatchNorm2d
layer3.1.relu2 ReLU
layer3.2.conv1.wrapped_module Conv2d
layer3.2.bn1 BatchNorm2d
layer3.2.relu1 ReLU
layer3.2.conv2.wrapped_module Conv2d
layer3.2.bn2 BatchNorm2d
layer3.2.relu2 ReLU
avgpool AvgPool2d
fc.wrapped_module Linear
Log file for this run: /workspace/distiller/examples/classifier_compression/logs/2019.02.24-200528/2019.02.24-200528.log
root@f3527f089871:/workspace/distiller/examples/classifier_compression#
version: 1
pruners:
conv1_pruner:
class: 'AutomatedGradualPruner'
initial_sparsity : 0.15
final_sparsity: 0.3
weights: ['conv1.weight']
conv2_pruner:
class: 'AutomatedGradualPruner'
initial_sparsity : 0.15
final_sparsity: 0.5
weights: ['conv2.weight']
fc_pruner:
class: 'AutomatedGradualPruner'
initial_sparsity : 0.15
final_sparsity: 0.80
weights: ['fc1.weight', 'fc2.weight','fc3.weight']
lr_schedulers:
pruning_lr:
class: StepLR
step_size: 30
gamma: 0.10
policies:
- pruner:
instance_name : 'conv1_pruner'
starting_epoch: 1
ending_epoch: 100
frequency: 2
- pruner:
instance_name : 'conv2_pruner'
starting_epoch: 1
ending_epoch: 100
frequency: 2
- pruner:
instance_name : 'fc_pruner'
starting_epoch: 1
ending_epoch: 100
frequency: 2
- lr_scheduler:
instance_name: pruning_lr
starting_epoch: 1
ending_epoch: 100
frequency: 1
docker run -ti --rm --name distiller -p 6006:6006 -p 8888:8888 -v /home/masaaki/Docker/distiller/workspace/:/workspace/ pytorch0.4cuda9 /bin/bash
だった。OSError: [Errno 99] Cannot assign requested address
だそうだ。Writing default config to: /root/.jupyter/jupyter_notebook_config.py
の下に、## The IP address the notebook server will listen on.
#c.NotebookApp.ip = 'localhost'
を追加した。c.NotebookApp.ip = '0.0.0.0'
と言われたので、Running as root is not recommended. Use --allow-root to bypass.
root@175abda804b7:/workspace/distiller/examples/classifier_compression# python compress_classifier.py -h
usage: compress_classifier.py [-h] [--arch ARCH] [-j N] [--epochs N] [-b N]
[--lr LR] [--momentum M] [--weight-decay W]
[--print-freq N] [--resume PATH] [-e]
[--pretrained]
[--activation-stats PHASE [PHASE ...]]
[--masks-sparsity] [--param-hist]
[--summary {sparsity,compute,model,modules,png,png_w_params,onnx}]
[--compress [COMPRESS]]
[--sense {element,filter,channel}]
[--sense-range SENSITIVITY_RANGE SENSITIVITY_RANGE SENSITIVITY_RANGE]
[--extras EXTRAS] [--deterministic]
[--gpus DEV_ID] [--cpu] [--name NAME]
[--out-dir OUTPUT_DIR]
[--validation-split VALIDATION_SPLIT]
[--effective-train-size EFFECTIVE_TRAIN_SIZE]
[--effective-valid-size EFFECTIVE_VALID_SIZE]
[--effective-test-size EFFECTIVE_TEST_SIZE]
[--confusion]
[--earlyexit_lossweights [EARLYEXIT_LOSSWEIGHTS [EARLYEXIT_LOSSWEIGHTS ...]]]
[--earlyexit_thresholds [EARLYEXIT_THRESHOLDS [EARLYEXIT_THRESHOLDS ...]]]
[--num-best-scores NUM_BEST_SCORES]
[--load-serialized] [--thinnify]
[--kd-teacher ARCH] [--kd-pretrained]
[--kd-resume PATH] [--kd-temperature TEMP]
[--kd-distill-wt WEIGHT]
[--kd-student-wt WEIGHT]
[--kd-teacher-wt WEIGHT]
[--kd-start-epoch EPOCH_NUM]
[--quantize-eval | --qe-calibration PORTION_OF_TEST_SET]
[--qe-mode QE_MODE] [--qe-bits-acts NUM_BITS]
[--qe-bits-wts NUM_BITS]
[--qe-bits-accum NUM_BITS] [--qe-clip-acts]
[--qe-no-clip-layers LAYER_NAME [LAYER_NAME ...]]
[--qe-per-channel] [--qe-stats-file PATH]
[--qe-config-file PATH] [--greedy]
[--greedy-ft-epochs GREEDY_FT_EPOCHS]
[--greedy-target-density GREEDY_TARGET_DENSITY]
[--greedy-pruning-step GREEDY_PRUNING_STEP]
[--greedy-finetuning-policy {constant,linear-grow}]
[--amc]
[--amc-protocol {mac-constrained,param-constrained,accuracy-guaranteed,mac-constrained-experimental}]
[--amc-ft-epochs AMC_FT_EPOCHS]
[--amc-save-chkpts]
[--amc-action-range AMC_ACTION_RANGE AMC_ACTION_RANGE]
[--amc-heatup-epochs AMC_HEATUP_EPOCHS]
[--amc-training-epochs AMC_TRAINING_EPOCHS]
[--amc-reward-frequency AMC_REWARD_FREQUENCY]
[--amc-target-density AMC_TARGET_DENSITY]
[--amc-agent-algo {ClippedPPO-continuous,ClippedPPO-discrete,DDPG,Random-policy}]
DIR
Distiller image classification model compression
positional arguments:
DIR path to dataset
optional arguments:
-h, --help show this help message and exit
--arch ARCH, -a ARCH model architecture: alexnet | alexnet_bn | densenet121
| densenet161 | densenet169 | densenet201 |
inception_v3 | mobilenet | mobilenet_025 |
mobilenet_050 | mobilenet_075 | plain20_cifar |
preact_resnet101 | preact_resnet110_cifar |
preact_resnet110_cifar_conv_ds | preact_resnet152 |
preact_resnet18 | preact_resnet20_cifar |
preact_resnet20_cifar_conv_ds | preact_resnet32_cifar
| preact_resnet32_cifar_conv_ds | preact_resnet34 |
preact_resnet44_cifar | preact_resnet44_cifar_conv_ds
| preact_resnet50 | preact_resnet56_cifar |
preact_resnet56_cifar_conv_ds | resnet101 |
resnet101_earlyexit | resnet110_cifar_earlyexit |
resnet1202_cifar_earlyexit | resnet152 |
resnet152_earlyexit | resnet18 | resnet18_earlyexit |
resnet20_cifar | resnet20_cifar_earlyexit |
resnet32_cifar | resnet32_cifar_earlyexit | resnet34 |
resnet34_earlyexit | resnet44_cifar |
resnet44_cifar_earlyexit | resnet50 |
resnet50_earlyexit | resnet56_cifar |
resnet56_cifar_earlyexit | simplenet_cifar |
squeezenet1_0 | squeezenet1_1 | vgg11 | vgg11_bn |
vgg11_bn_cifar | vgg11_cifar | vgg13 | vgg13_bn |
vgg13_bn_cifar | vgg13_cifar | vgg16 | vgg16_bn |
vgg16_bn_cifar | vgg16_cifar | vgg19 | vgg19_bn |
vgg19_bn_cifar | vgg19_cifar (default: resnet18)
-j N, --workers N number of data loading workers (default: 4)
--epochs N number of total epochs to run
-b N, --batch-size N mini-batch size (default: 256)
--lr LR, --learning-rate LR
initial learning rate
--momentum M momentum
--weight-decay W, --wd W
weight decay (default: 1e-4)
--print-freq N, -p N print frequency (default: 10)
--resume PATH path to latest checkpoint (default: none)
-e, --evaluate evaluate model on validation set
--pretrained use pre-trained model
--activation-stats PHASE [PHASE ...], --act-stats PHASE [PHASE ...]
collect activation statistics on phases: train, valid,
and/or test (WARNING: this slows down training)
--masks-sparsity print masks sparsity table at end of each epoch
--param-hist log the parameter tensors histograms to file (WARNING:
this can use significant disk space)
--summary {sparsity,compute,model,modules,png,png_w_params,onnx}
print a summary of the model, and exit - options:
sparsity | compute | model | modules | png |
png_w_params | onnx
--compress [COMPRESS]
configuration file for pruning the model (default is
to use hard-coded schedule)
--sense {element,filter,channel}
test the sensitivity of layers to pruning
--sense-range SENSITIVITY_RANGE SENSITIVITY_RANGE SENSITIVITY_RANGE
an optional parameter for sensitivity testing
providing the range of sparsities to test. This is
equivalent to creating sensitivities =
np.arange(start, stop, step)
--extras EXTRAS file with extra configuration information
--deterministic, --det
Ensure deterministic execution for re-producible
results.
--gpus DEV_ID Comma-separated list of GPU device IDs to be used
(default is to use all available devices)
--cpu Use CPU only. Flag not set => uses GPUs according to
the --gpus flag value.Flag set => overrides the --gpus
flag
--name NAME, -n NAME Experiment name
--out-dir OUTPUT_DIR, -o OUTPUT_DIR
Path to dump logs and checkpoints
--validation-split VALIDATION_SPLIT, --valid-size VALIDATION_SPLIT, --vs VALIDATION_SPLIT
Portion of training dataset to set aside for
validation
--effective-train-size EFFECTIVE_TRAIN_SIZE, --etrs EFFECTIVE_TRAIN_SIZE
Portion of training dataset to be used in each epoch.
NOTE: If --validation-split is set, then the value of
this argument is applied AFTER the train-validation
split according to that argument
--effective-valid-size EFFECTIVE_VALID_SIZE, --evs EFFECTIVE_VALID_SIZE
Portion of validation dataset to be used in each
epoch. NOTE: If --validation-split is set, then the
value of this argument is applied AFTER the train-
validation split according to that argument
--effective-test-size EFFECTIVE_TEST_SIZE, --etes EFFECTIVE_TEST_SIZE
Portion of test dataset to be used in each epoch
--confusion Display the confusion matrix
--earlyexit_lossweights [EARLYEXIT_LOSSWEIGHTS [EARLYEXIT_LOSSWEIGHTS ...]]
List of loss weights for early exits (e.g.
--earlyexit_lossweights 0.1 0.3)
--earlyexit_thresholds [EARLYEXIT_THRESHOLDS [EARLYEXIT_THRESHOLDS ...]]
List of EarlyExit thresholds (e.g.
--earlyexit_thresholds 1.2 0.9)
--num-best-scores NUM_BEST_SCORES
number of best scores to track and report (default: 1)
--load-serialized Load a model without DataParallel wrapping it
--thinnify physically remove zero-filters and create a smaller
model
--greedy greedy filter pruning
--amc AutoML Compression
Knowledge Distillation Training Arguments:
--kd-teacher ARCH Model architecture for teacher model
--kd-pretrained Use pre-trained model for teacher
--kd-resume PATH Path to checkpoint from which to load teacher weights
--kd-temperature TEMP, --kd-temp TEMP
Knowledge distillation softmax temperature
--kd-distill-wt WEIGHT, --kd-dw WEIGHT
Weight for distillation loss (student vs. teacher soft
targets)
--kd-student-wt WEIGHT, --kd-sw WEIGHT
Weight for student vs. labels loss
--kd-teacher-wt WEIGHT, --kd-tw WEIGHT
Weight for teacher vs. labels loss
--kd-start-epoch EPOCH_NUM
Epoch from which to enable distillation
Arguments controlling quantization at evaluation time ("post-training quantization"):
--quantize-eval, --qe
Apply linear quantization to model before evaluation.
Applicable only if --evaluate is also set
--qe-calibration PORTION_OF_TEST_SET
Run the model in evaluation mode on the specified
portion of the test dataset and collect statistics.
Ignores all other 'qe--*' arguments
--qe-mode QE_MODE, --qem QE_MODE
Linear quantization mode. Choices: sym | asym_s |
asym_u
--qe-bits-acts NUM_BITS, --qeba NUM_BITS
Number of bits for quantization of activations
--qe-bits-wts NUM_BITS, --qebw NUM_BITS
Number of bits for quantization of weights
--qe-bits-accum NUM_BITS
Number of bits for quantization of the accumulator
--qe-clip-acts, --qeca
Enable clipping of activations using min/max values
averaging over batch
--qe-no-clip-layers LAYER_NAME [LAYER_NAME ...], --qencl LAYER_NAME [LAYER_NAME ...]
List of layer names for which not to clip activations.
Applicable only if --qe-clip-acts is also set
--qe-per-channel, --qepc
Enable per-channel quantization of weights (per output
channel)
--qe-stats-file PATH Path to YAML file with calibration stats. If not
given, dynamic quantization will be run (Note that not
all layer types are supported for dynamic
quantization)
--qe-config-file PATH
Path to YAML file containing configuration for
PostTrainLinearQuantizer (if present, all other --qe*
arguments are ignored)
Greedy Pruning:
--greedy-ft-epochs GREEDY_FT_EPOCHS
number of epochs to fine-tune each discovered network
--greedy-target-density GREEDY_TARGET_DENSITY
target density of the network we are seeking
--greedy-pruning-step GREEDY_PRUNING_STEP
size of each pruning step (as a fraction in [0..1])
--greedy-finetuning-policy {constant,linear-grow}
policy used for determining how long to fine-tune
AutoML Compression Arguments:
--amc-protocol {mac-constrained,param-constrained,accuracy-guaranteed,mac-constrained-experimental}
Compression-policy search protocol
--amc-ft-epochs AMC_FT_EPOCHS
The number of epochs to fine-tune each discovered
network
--amc-save-chkpts Save checkpoints of all discovered networks
--amc-action-range AMC_ACTION_RANGE AMC_ACTION_RANGE
Density action range (a_min, a_max)
--amc-heatup-epochs AMC_HEATUP_EPOCHS
The number of epochs for heatup/exploration
--amc-training-epochs AMC_TRAINING_EPOCHS
The number of epochs for training/exploitation
--amc-reward-frequency AMC_REWARD_FREQUENCY
Reward computation frequency (measured in agent steps)
--amc-target-density AMC_TARGET_DENSITY
Target density of the network we are seeking
--amc-agent-algo {ClippedPPO-continuous,ClippedPPO-discrete,DDPG,Random-policy}
The agent algorithm to use
root@175abda804b7:/workspace/distiller/examples/classifier_compression#
FROM pytorch/pytorch:0.4_cuda9_cudnn7
RUN apt-get update && apt-get install -y vim
docker build -t pytorch0.4cuda9 .
docker run -ti --rm --name distiller -p 6006:6006 -v /home/masaaki/Docker/distiller/workspace/:/workspace/ pytorch0.4cuda9 /bin/bash
だということだ。image.ub
BOOT.BIN
sdhci@ff160000: 0 (SD)Card did not respond to voltage select!
mmc_init: -95, time 27
*** Warning - bad CRC, using default environment
In: serial@ff010000
Out: serial@ff010000
Err: serial@ff010000
Board: Xilinx ZynqMP
Bootmode: SD_MODE
U-BOOT for cam_test_ov5642_182
Hit any key to stop autoboot: 0
Device: sdhci@ff160000
Manufacturer ID: 3
OEM: 5344
Name: SL08G
Tran Speed: 50000000
Rd Block Len: 512
SD version 3.0
High Capacity: Yes
Capacity: 7.4 GiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
reading image.ub
13407472 bytes read in 1046 ms (12.2 MiB/s)
## Loading kernel from FIT Image at 10000000 ...
Using 'conf@system-top.dtb' configuration
Trying 'kernel@1' kernel subimage
Description: Linux kernel
Type: Kernel Image
Compression: gzip compressed
Data Start: 0x10000104
Data Size: 6960376 Bytes = 6.6 MiB
Architecture: AArch64
OS: Linux
Load Address: 0x00080000
Entry Point: 0x00080000
Hash algo: sha1
Hash value: f7969295bb0009929ae004732e24833a06e81c68
Verifying Hash Integrity ... sha1+ OK
## Loading ramdisk from FIT Image at 10000000 ...
Using 'conf@system-top.dtb' configuration
Trying 'ramdisk@1' ramdisk subimage
Description: petalinux-user-image
Type: RAMDisk Image
Compression: gzip compressed
Data Start: 0x106ab1c0
Data Size: 6413725 Bytes = 6.1 MiB
Architecture: AArch64
OS: Linux
Load Address: unavailable
Entry Point: unavailable
Hash algo: sha1
Hash value: 6f3c1c7264cea1a5b391c6f6eefbeb190b3e9cf6
Verifying Hash Integrity ... sha1+ OK
## Loading fdt from FIT Image at 10000000 ...
Using 'conf@system-top.dtb' configuration
Trying 'fdt@system-top.dtb' fdt subimage
Description: Flattened Device Tree blob
Type: Flat Device Tree
Compression: uncompressed
Data Start: 0x106a3700
Data Size: 31225 Bytes = 30.5 KiB
Architecture: AArch64
Hash algo: sha1
Hash value: a6944372b09e37df9457fdbb9be433c8088ff6e3
Verifying Hash Integrity ... sha1+ OK
Booting using the fdt blob at 0x106a3700
Uncompressing Kernel Image ... OK
Loading Ramdisk to 079e2000, end 07fffd9d ... OK
Loading Device Tree to 00000000079d7000, end 00000000079e19f8 ... OK
Starting kernel ...
[ 0.000000] Booting Linux on physical CPU 0x0
[ 0.000000] Linux version 4.14.0-xilinx-v2018.2 (oe-user@oe-host) (gcc version 7.2.0 (GCC)) #5 SMP Mon Feb 18 12:13:05 UTC 2019
[ 0.000000] Boot CPU: AArch64 Processor [410fd034]
[ 0.000000] Machine model: xlnx,zynqmp
[ 0.000000] earlycon: cdns0 at MMIO 0x00000000ff010000 (options '115200n8')
[ 0.000000] bootconsole [cdns0] enabled
[ 0.000000] efi: Getting EFI parameters from FDT:
[ 0.000000] efi: UEFI not found.
[ 0.000000] cma: Reserved 256 MiB at 0x000000006fc00000
[ 0.000000] psci: probing for conduit method from DT.
[ 0.000000] psci: PSCIv1.1 detected in firmware.
[ 0.000000] psci: Using standard PSCI v0.2 function IDs
[ 0.000000] psci: MIGRATE_INFO_TYPE not supported.
[ 0.000000] percpu: Embedded 21 pages/cpu @ffffffc07fe79000 s46488 r8192 d31336 u86016
[ 0.000000] Detected VIPT I-cache on CPU0
[ 0.000000] CPU features: enabling workaround for ARM erratum 845719
[ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 516867
[ 0.000000] Kernel command line: earlycon clk_ignore_unused
[ 0.000000] PID hash table entries: 4096 (order: 3, 32768 bytes)
[ 0.000000] Dentry cache hash table entries: 262144 (order: 9, 2097152 bytes)
[ 0.000000] Inode-cache hash table entries: 131072 (order: 8, 1048576 bytes)
[ 0.000000] Memory: 1778840K/2096128K available (9980K kernel code, 644K rwdata, 3128K rodata, 512K init, 2168K bss, 55144K reserved, 262144K cma-reserved)
[ 0.000000] Virtual kernel memory layout:
[ 0.000000] modules : 0xffffff8000000000 - 0xffffff8008000000 ( 128 MB)
[ 0.000000] vmalloc : 0xffffff8008000000 - 0xffffffbebfff0000 ( 250 GB)
[ 0.000000] .text : 0xffffff8008080000 - 0xffffff8008a40000 ( 9984 KB)
[ 0.000000] .rodata : 0xffffff8008a40000 - 0xffffff8008d60000 ( 3200 KB)
[ 0.000000] .init : 0xffffff8008d60000 - 0xffffff8008de0000 ( 512 KB)
[ 0.000000] .data : 0xffffff8008de0000 - 0xffffff8008e81200 ( 645 KB)
[ 0.000000] .bss : 0xffffff8008e81200 - 0xffffff800909f2b0 ( 2169 KB)
[ 0.000000] fixed : 0xffffffbefe7fd000 - 0xffffffbefec00000 ( 4108 KB)
[ 0.000000] PCI I/O : 0xffffffbefee00000 - 0xffffffbeffe00000 ( 16 MB)
[ 0.000000] vmemmap : 0xffffffbf00000000 - 0xffffffc000000000 ( 4 GB maximum)
[ 0.000000] 0xffffffbf00000000 - 0xffffffbf01bfc800 ( 27 MB actual)
[ 0.000000] memory : 0xffffffc000000000 - 0xffffffc07ff00000 ( 2047 MB)
[ 0.000000] Hierarchical RCU implementation.
[ 0.000000] RCU event tracing is enabled.
[ 0.000000] RCU restricting CPUs from NR_CPUS=8 to nr_cpu_ids=4.
[ 0.000000] RCU: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=4
[ 0.000000] NR_IRQS: 64, nr_irqs: 64, preallocated irqs: 0
[ 0.000000] GIC: Adjusting CPU interface base to 0x00000000f902f000
[ 0.000000] GIC: Using split EOI/Deactivate mode
[ 0.000000] arch_timer: cp15 timer(s) running at 100.00MHz (phys).
[ 0.000000] clocksource: arch_sys_counter: mask: 0xffffffffffffff max_cycles: 0x171024ee1c, max_idle_ns: 440795203561 ns
[ 0.000004] sched_clock: 56 bits at 100MHz, resolution 10ns, wraps every 4398046511100ns
[ 0.008366] Console: colour dummy device 80x25
[ 0.012620] console [tty0] enabled
[ 0.015985] bootconsole [cdns0] disabled
[ 0.000000] Booting Linux on physical CPU 0x0
[ 0.000000] Linux version 4.14.0-xilinx-v2018.2 (oe-user@oe-host) (gcc version 7.2.0 (GCC)) #5 SMP Mon Feb 18 12:13:05 UTC 2019
[ 0.000000] Boot CPU: AArch64 Processor [410fd034]
[ 0.000000] Machine model: xlnx,zynqmp
[ 0.000000] earlycon: cdns0 at MMIO 0x00000000ff010000 (options '115200n8')
[ 0.000000] bootconsole [cdns0] enabled
[ 0.000000] efi: Getting EFI parameters from FDT:
[ 0.000000] efi: UEFI not found.
[ 0.000000] cma: Reserved 256 MiB at 0x000000006fc00000
[ 0.000000] psci: probing for conduit method from DT.
[ 0.000000] psci: PSCIv1.1 detected in firmware.
[ 0.000000] psci: Using standard PSCI v0.2 function IDs
[ 0.000000] psci: MIGRATE_INFO_TYPE not supported.
[ 0.000000] percpu: Embedded 21 pages/cpu @ffffffc07fe79000 s46488 r8192 d31336 u86016
[ 0.000000] Detected VIPT I-cache on CPU0
[ 0.000000] CPU features: enabling workaround for ARM erratum 845719
[ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 516867
[ 0.000000] Kernel command line: earlycon clk_ignore_unused
[ 0.000000] PID hash table entries: 4096 (order: 3, 32768 bytes)
[ 0.000000] Dentry cache hash table entries: 262144 (order: 9, 2097152 bytes)
[ 0.000000] Inode-cache hash table entries: 131072 (order: 8, 1048576 bytes)
[ 0.000000] Memory: 1778840K/2096128K available (9980K kernel code, 644K rwdata, 3128K rodata, 512K init, 2168K bss, 55144K reserved, 262144K cma-reserved)
[ 0.000000] Virtual kernel memory layout:
[ 0.000000] modules : 0xffffff8000000000 - 0xffffff8008000000 ( 128 MB)
[ 0.000000] vmalloc : 0xffffff8008000000 - 0xffffffbebfff0000 ( 250 GB)
[ 0.000000] .text : 0xffffff8008080000 - 0xffffff8008a40000 ( 9984 KB)
[ 0.000000] .rodata : 0xffffff8008a40000 - 0xffffff8008d60000 ( 3200 KB)
[ 0.000000] .init : 0xffffff8008d60000 - 0xffffff8008de0000 ( 512 KB)
[ 0.000000] .data : 0xffffff8008de0000 - 0xffffff8008e81200 ( 645 KB)
[ 0.000000] .bss : 0xffffff8008e81200 - 0xffffff800909f2b0 ( 2169 KB)
[ 0.000000] fixed : 0xffffffbefe7fd000 - 0xffffffbefec00000 ( 4108 KB)
[ 0.000000] PCI I/O : 0xffffffbefee00000 - 0xffffffbeffe00000 ( 16 MB)
[ 0.000000] vmemmap : 0xffffffbf00000000 - 0xffffffc000000000 ( 4 GB maximum)
[ 0.000000] 0xffffffbf00000000 - 0xffffffbf01bfc800 ( 27 MB actual)
[ 0.000000] memory : 0xffffffc000000000 - 0xffffffc07ff00000 ( 2047 MB)
[ 0.000000] Hierarchical RCU implementation.
[ 0.000000] RCU event tracing is enabled.
[ 0.000000] RCU restricting CPUs from NR_CPUS=8 to nr_cpu_ids=4.
[ 0.000000] RCU: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=4
[ 0.000000] NR_IRQS: 64, nr_irqs: 64, preallocated irqs: 0
[ 0.000000] GIC: Adjusting CPU interface base to 0x00000000f902f000
[ 0.000000] GIC: Using split EOI/Deactivate mode
[ 0.000000] arch_timer: cp15 timer(s) running at 100.00MHz (phys).
[ 0.000000] clocksource: arch_sys_counter: mask: 0xffffffffffffff max_cycles: 0x171024ee1c, max_idle_ns: 440795203561 ns
[ 0.000004] sched_clock: 56 bits at 100MHz, resolution 10ns, wraps every 4398046511100ns
[ 0.008366] Console: colour dummy device 80x25
[ 0.012620] console [tty0] enabled
[ 0.015985] bootconsole [cdns0] disabled
[ 0.019900] Calibrating delay loop (skipped), value calculated using timer frequency.. 200.00 BogoMIPS (lpj=400000)
[ 0.019914] pid_max: default: 32768 minimum: 301
[ 0.020034] Mount-cache hash table entries: 4096 (order: 3, 32768 bytes)
[ 0.020052] Mountpoint-cache hash table entries: 4096 (order: 3, 32768 bytes)
[ 0.020816] ASID allocator initialised with 65536 entries
[ 0.020874] Hierarchical SRCU implementation.
[ 0.021167] EFI services will not be available.
[ 0.021194] zynqmp_plat_init Platform Management API v1.0
[ 0.021203] zynqmp_plat_init Trustzone version v1.0
[ 0.021321] smp: Bringing up secondary CPUs ...
[ 0.021605] Detected VIPT I-cache on CPU1
[ 0.021645] CPU1: Booted secondary processor [410fd034]
[ 0.021950] Detected VIPT I-cache on CPU2
[ 0.021969] CPU2: Booted secondary processor [410fd034]
[ 0.022258] Detected VIPT I-cache on CPU3
[ 0.022276] CPU3: Booted secondary processor [410fd034]
[ 0.022318] smp: Brought up 1 node, 4 CPUs
[ 0.022351] SMP: Total of 4 processors activated.
[ 0.022360] CPU features: detected feature: 32-bit EL0 Support
[ 0.022372] CPU: All CPU(s) started at EL2
[ 0.022390] alternatives: patching kernel code
[ 0.023372] devtmpfs: initialized
[ 0.026698] random: get_random_u32 called from bucket_table_alloc+0x108/0x260 with crng_init=0
[ 0.026876] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns
[ 0.026900] futex hash table entries: 1024 (order: 5, 131072 bytes)
[ 0.033084] xor: measuring software checksum speed
[ 0.071993] 8regs : 2303.000 MB/sec
[ 0.112022] 8regs_prefetch: 2053.000 MB/sec
[ 0.152052] 32regs : 2830.000 MB/sec
[ 0.192082] 32regs_prefetch: 2379.000 MB/sec
[ 0.192090] xor: using function: 32regs (2830.000 MB/sec)
[ 0.192186] pinctrl core: initialized pinctrl subsystem
[ 0.192452] random: fast init done
[ 0.192909] NET: Registered protocol family 16
[ 0.193606] cpuidle: using governor menu
[ 0.194006] vdso: 2 pages (1 code @ ffffff8008a46000, 1 data @ ffffff8008de4000)
[ 0.194026] hw-breakpoint: found 6 breakpoint and 4 watchpoint registers.
[ 0.194670] DMA: preallocated 256 KiB pool for atomic allocations
[ 0.225155] reset_zynqmp reset-controller: Xilinx zynqmp reset driver probed
[ 0.225667] ARM CCI_400_r1 PMU driver probed
[ 0.230165] zynqmp-pinctrl ff180000.pinctrl: zynqmp pinctrl initialized
[ 0.238782] HugeTLB registered 2.00 MiB page size, pre-allocated 0 pages
[ 0.304374] raid6: int64x1 gen() 402 MB/s
[ 0.372301] raid6: int64x1 xor() 446 MB/s
[ 0.440423] raid6: int64x2 gen() 688 MB/s
[ 0.508430] raid6: int64x2 xor() 603 MB/s
[ 0.576452] raid6: int64x4 gen() 1042 MB/s
[ 0.644510] raid6: int64x4 xor() 742 MB/s
[ 0.712591] raid6: int64x8 gen() 980 MB/s
[ 0.780603] raid6: int64x8 xor() 745 MB/s
[ 0.848649] raid6: neonx1 gen() 726 MB/s
[ 0.916709] raid6: neonx1 xor() 853 MB/s
[ 0.984787] raid6: neonx2 gen() 1169 MB/s
[ 1.052826] raid6: neonx2 xor() 1207 MB/s
[ 1.120889] raid6: neonx4 gen() 1506 MB/s
[ 1.188926] raid6: neonx4 xor() 1442 MB/s
[ 1.256979] raid6: neonx8 gen() 1652 MB/s
[ 1.325041] raid6: neonx8 xor() 1534 MB/s
[ 1.325049] raid6: using algorithm neonx8 gen() 1652 MB/s
[ 1.325056] raid6: .... xor() 1534 MB/s, rmw enabled
[ 1.325065] raid6: using neon recovery algorithm
[ 1.326150] SCSI subsystem initialized
[ 1.326371] usbcore: registered new interface driver usbfs
[ 1.326412] usbcore: registered new interface driver hub
[ 1.326454] usbcore: registered new device driver usb
[ 1.326528] media: Linux media interface: v0.10
[ 1.326558] Linux video capture interface: v2.00
[ 1.326607] pps_core: LinuxPPS API ver. 1 registered
[ 1.326616] pps_core: Software ver. 5.3.6 - Copyright 2005-2007 Rodolfo Giometti <giometti@linux.it>
[ 1.326640] PTP clock support registered
[ 1.326678] EDAC MC: Ver: 3.0.0
[ 1.327071] zynqmp-ipi ff9905c0.mailbox: Probed ZynqMP IPI Mailbox driver.
[ 1.327233] FPGA manager framework
[ 1.327355] fpga-region fpga-full: FPGA Region probed
[ 1.327460] Advanced Linux Sound Architecture Driver Initialized.
[ 1.327764] Bluetooth: Core ver 2.22
[ 1.327799] NET: Registered protocol family 31
[ 1.327807] Bluetooth: HCI device and connection manager initialized
[ 1.327821] Bluetooth: HCI socket layer initialized
[ 1.327831] Bluetooth: L2CAP socket layer initialized
[ 1.327852] Bluetooth: SCO socket layer initialized
[ 1.328414] clocksource: Switched to clocksource arch_sys_counter
[ 1.328507] VFS: Disk quotas dquot_6.6.0
[ 1.328564] VFS: Dquot-cache hash table entries: 512 (order 0, 4096 bytes)
[ 1.332955] NET: Registered protocol family 2
[ 1.333324] TCP established hash table entries: 16384 (order: 5, 131072 bytes)
[ 1.333443] TCP bind hash table entries: 16384 (order: 6, 262144 bytes)
[ 1.333783] TCP: Hash tables configured (established 16384 bind 16384)
[ 1.333890] UDP hash table entries: 1024 (order: 3, 32768 bytes)
[ 1.333934] UDP-Lite hash table entries: 1024 (order: 3, 32768 bytes)
[ 1.334079] NET: Registered protocol family 1
[ 1.334300] RPC: Registered named UNIX socket transport module.
[ 1.334310] RPC: Registered udp transport module.
[ 1.334318] RPC: Registered tcp transport module.
[ 1.334325] RPC: Registered tcp NFSv4.1 backchannel transport module.
[ 1.334438] Trying to unpack rootfs image as initramfs...
[ 1.593990] Freeing initrd memory: 6260K
[ 1.594473] hw perfevents: no interrupt-affinity property for /pmu, guessing.
[ 1.594653] hw perfevents: enabled with armv8_pmuv3 PMU driver, 7 counters available
[ 1.595523] audit: initializing netlink subsys (disabled)
[ 1.595646] audit: type=2000 audit(1.579:1): state=initialized audit_enabled=0 res=1
[ 1.596018] workingset: timestamp_bits=62 max_order=19 bucket_order=0
[ 1.596843] NFS: Registering the id_resolver key type
[ 1.596871] Key type id_resolver registered
[ 1.596879] Key type id_legacy registered
[ 1.596893] nfs4filelayout_init: NFSv4 File Layout Driver Registering...
[ 1.596918] jffs2: version 2.2. (NAND) (SUMMARY) © 2001-2006 Red Hat, Inc.
[ 1.624254] Block layer SCSI generic (bsg) driver version 0.4 loaded (major 246)
[ 1.624284] io scheduler noop registered
[ 1.624293] io scheduler deadline registered
[ 1.624315] io scheduler cfq registered (default)
[ 1.624324] io scheduler mq-deadline registered
[ 1.624333] io scheduler kyber registered
[ 1.625770] xilinx-dpdma fd4c0000.dma: Xilinx DPDMA engine is probed
[ 1.626207] xilinx-zynqmp-dma fd500000.dma: ZynqMP DMA driver Probe success
[ 1.626359] xilinx-zynqmp-dma fd510000.dma: ZynqMP DMA driver Probe success
[ 1.626510] xilinx-zynqmp-dma fd520000.dma: ZynqMP DMA driver Probe success
[ 1.626667] xilinx-zynqmp-dma fd530000.dma: ZynqMP DMA driver Probe success
[ 1.626821] xilinx-zynqmp-dma fd540000.dma: ZynqMP DMA driver Probe success
[ 1.626977] xilinx-zynqmp-dma fd550000.dma: ZynqMP DMA driver Probe success
[ 1.627149] xilinx-zynqmp-dma fd560000.dma: ZynqMP DMA driver Probe success
[ 1.627306] xilinx-zynqmp-dma fd570000.dma: ZynqMP DMA driver Probe success
[ 1.627529] xilinx-zynqmp-dma ffa80000.dma: ZynqMP DMA driver Probe success
[ 1.627682] xilinx-zynqmp-dma ffa90000.dma: ZynqMP DMA driver Probe success
[ 1.627829] xilinx-zynqmp-dma ffaa0000.dma: ZynqMP DMA driver Probe success
[ 1.627983] xilinx-zynqmp-dma ffab0000.dma: ZynqMP DMA driver Probe success
[ 1.628134] xilinx-zynqmp-dma ffac0000.dma: ZynqMP DMA driver Probe success
[ 1.628286] xilinx-zynqmp-dma ffad0000.dma: ZynqMP DMA driver Probe success
[ 1.628464] xilinx-zynqmp-dma ffae0000.dma: ZynqMP DMA driver Probe success
[ 1.628620] xilinx-zynqmp-dma ffaf0000.dma: ZynqMP DMA driver Probe success
[ 1.655925] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
[ 1.659223] cacheinfo: Unable to detect cache hierarchy for CPU 0
[ 1.663995] brd: module loaded
[ 1.668278] loop: module loaded
[ 1.669240] mtdoops: mtd device (mtddev=name/number) must be supplied
[ 1.671407] libphy: Fixed MDIO Bus: probed
[ 1.672554] tun: Universal TUN/TAP device driver, 1.6
[ 1.672897] CAN device driver interface
[ 1.673756] usbcore: registered new interface driver asix
[ 1.673829] usbcore: registered new interface driver ax88179_178a
[ 1.673862] usbcore: registered new interface driver cdc_ether
[ 1.673895] usbcore: registered new interface driver net1080
[ 1.673927] usbcore: registered new interface driver cdc_subset
[ 1.673957] usbcore: registered new interface driver zaurus
[ 1.674001] usbcore: registered new interface driver cdc_ncm
[ 1.674389] xilinx-axipmon ffa00000.perf-monitor: Probed Xilinx APM
[ 1.674863] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[ 1.674873] ehci-pci: EHCI PCI platform driver
[ 1.675117] usbcore: registered new interface driver uas
[ 1.675159] usbcore: registered new interface driver usb-storage
[ 1.676094] rtc_zynqmp ffa60000.rtc: rtc core: registered ffa60000.rtc as rtc0
[ 1.676154] i2c /dev entries driver
[ 1.676909] cdns-i2c ff030000.i2c: 400 kHz mmio ff030000 irq 31
[ 1.677181] IR NEC protocol handler initialized
[ 1.677190] IR RC5(x/sz) protocol handler initialized
[ 1.677198] IR RC6 protocol handler initialized
[ 1.677206] IR JVC protocol handler initialized
[ 1.677213] IR Sony protocol handler initialized
[ 1.677220] IR SANYO protocol handler initialized
[ 1.677228] IR Sharp protocol handler initialized
[ 1.677235] IR MCE Keyboard/mouse protocol handler initialized
[ 1.677243] IR XMP protocol handler initialized
[ 1.678213] usbcore: registered new interface driver uvcvideo
[ 1.678222] USB Video Class driver (1.1.1)
[ 1.679547] cdns-wdt fd4d0000.watchdog: Xilinx Watchdog Timer at ffffff800913d000 with timeout 10s
[ 1.679843] Bluetooth: HCI UART driver ver 2.3
[ 1.679854] Bluetooth: HCI UART protocol H4 registered
[ 1.679863] Bluetooth: HCI UART protocol BCSP registered
[ 1.679891] Bluetooth: HCI UART protocol LL registered
[ 1.679900] Bluetooth: HCI UART protocol ATH3K registered
[ 1.679909] Bluetooth: HCI UART protocol Three-wire (H5) registered
[ 1.679953] Bluetooth: HCI UART protocol Intel registered
[ 1.679961] Bluetooth: HCI UART protocol QCA registered
[ 1.680002] usbcore: registered new interface driver bcm203x
[ 1.680038] usbcore: registered new interface driver bpa10x
[ 1.680072] usbcore: registered new interface driver bfusb
[ 1.680106] usbcore: registered new interface driver btusb
[ 1.680115] Bluetooth: Generic Bluetooth SDIO driver ver 0.1
[ 1.680169] usbcore: registered new interface driver ath3k
[ 1.680299] EDAC MC: ECC not enabled
[ 1.680512] EDAC DEVICE0: Giving out device to module zynqmp-ocm-edac controller zynqmp_ocm: DEV ff960000.memory-controller (INTERRUPT)
[ 1.681546] sdhci: Secure Digital Host Controller Interface driver
[ 1.681556] sdhci: Copyright(c) Pierre Ossman
[ 1.681564] sdhci-pltfm: SDHCI platform and OF driver helper
[ 1.724428] mmc0: SDHCI controller on ff160000.sdhci [ff160000.sdhci] using ADMA 64-bit
[ 1.776413] mmc1: SDHCI controller on ff170000.sdhci [ff170000.sdhci] using ADMA 64-bit
[ 1.783933] ledtrig-cpu: registered to indicate activity on CPUs
[ 1.784114] usbcore: registered new interface driver usbhid
[ 1.784122] usbhid: USB HID core driver
[ 1.787429] fpga_manager fpga0: Xilinx ZynqMP FPGA Manager registered
[ 1.788347] pktgen: Packet Generator for packet performance testing. Version: 2.75
[ 1.790615] Netfilter messages via NETLINK v0.30.
[ 1.790758] ip_tables: (C) 2000-2006 Netfilter Core Team
[ 1.790929] Initializing XFRM netlink socket
[ 1.791012] NET: Registered protocol family 10
[ 1.791444] Segment Routing with IPv6
[ 1.791495] ip6_tables: (C) 2000-2006 Netfilter Core Team
[ 1.791684] sit: IPv6, IPv4 and MPLS over IPv4 tunneling driver
[ 1.792050] NET: Registered protocol family 17
[ 1.792068] NET: Registered protocol family 15
[ 1.792092] bridge: filtering via arp/ip/ip6tables is no longer available by default. Update your scripts to load br_netfilter if you need this.
[ 1.792107] Ebtables v2.0 registered
[ 1.792531] can: controller area network core (rev 20170425 abi 9)
[ 1.792571] NET: Registered protocol family 29
[ 1.792581] can: raw protocol (rev 20170425)
[ 1.792589] can: broadcast manager protocol (rev 20170425 t)
[ 1.792602] can: netlink gateway (rev 20170425) max_hops=1
[ 1.792727] Bluetooth: RFCOMM TTY layer initialized
[ 1.792740] Bluetooth: RFCOMM socket layer initialized
[ 1.792760] Bluetooth: RFCOMM ver 1.11
[ 1.792772] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[ 1.792779] Bluetooth: BNEP filters: protocol multicast
[ 1.792791] Bluetooth: BNEP socket layer initialized
[ 1.792800] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
[ 1.792814] Bluetooth: HIDP socket layer initialized
[ 1.792975] 9pnet: Installing 9P2000 support
[ 1.793001] Key type dns_resolver registered
[ 1.793513] registered taskstats version 1
[ 1.793913] Btrfs loaded, crc32c=crc32c-generic
[ 1.800696] ff000000.serial: ttyPS1 at MMIO 0xff000000 (irq = 40, base_baud = 6249999) is a xuartps
[ 1.801591] ff010000.serial: ttyPS0 at MMIO 0xff010000 (irq = 41, base_baud = 6249999) is a xuartps
[ 1.868523] mmc0: error -110 whilst initialising SD card
[ 3.337586] console [ttyPS0] enabled
[ 3.342551] xilinx-psgtr fd400000.zynqmp_phy: Lane:1 type:8 protocol:4 pll_locked:yes
[ 3.350727] PLL: shutdown
[ 3.353395] PLL: shutdown
[ 3.356361] PLL: enable
[ 3.358861] PLL: shutdown
[ 3.362304] PLL: enable
[ 3.364779] xilinx-dp-snd-codec fd4a0000.zynqmp-display:zynqmp_dp_snd_codec0: Xilinx DisplayPort Sound Codec probed
[ 3.375371] xilinx-dp-snd-pcm zynqmp_dp_snd_pcm0: Xilinx DisplayPort Sound PCM probed
[ 3.383332] xilinx-dp-snd-pcm zynqmp_dp_snd_pcm1: Xilinx DisplayPort Sound PCM probed
[ 3.391671] xilinx-dp-snd-card fd4a0000.zynqmp-display:zynqmp_dp_snd_card: xilinx-dp-snd-codec-dai <-> xilinx-dp-snd-codec-dai mapping ok
[ 3.404056] xilinx-dp-snd-card fd4a0000.zynqmp-display:zynqmp_dp_snd_card: xilinx-dp-snd-codec-dai <-> xilinx-dp-snd-codec-dai mapping ok
[ 3.416734] xilinx-dp-snd-card fd4a0000.zynqmp-display:zynqmp_dp_snd_card: Xilinx DisplayPort Sound Card probed
[ 3.426839] OF: graph: no port node found in /amba/zynqmp-display@fd4a0000
[ 3.433741] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 3.440272] [drm] No driver support for vblank timestamp query.
[ 3.446247] xlnx-drm xlnx-drm.0: bound fd4a0000.zynqmp-display (ops 0xffffff8008af7228)
[ 3.475433] PLL: enable
[ 3.596080] Console: switching to colour frame buffer device 128x48
[ 3.621401] zynqmp-display fd4a0000.zynqmp-display: fb0: frame buffer device
[ 3.628740] [drm] Initialized xlnx 1.0.0 20130509 for fd4a0000.zynqmp-display on minor 0
[ 3.636846] zynqmp-display fd4a0000.zynqmp-display: ZynqMP DisplayPort Subsystem driver probed
[ 3.645796] dwc3-of-simple ff9d0000.usb0: dwc3_simple_set_phydata: Can't find usb3-phy
[ 3.654849] xhci-hcd xhci-hcd.0.auto: xHCI Host Controller
[ 3.660325] xhci-hcd xhci-hcd.0.auto: new USB bus registered, assigned bus number 1
[ 3.668220] xhci-hcd xhci-hcd.0.auto: hcc params 0x0238f625 hci version 0x100 quirks 0x22010010
[ 3.676958] xhci-hcd xhci-hcd.0.auto: irq 46, io mem 0xfe200000
[ 3.682987] usb usb1: New USB device found, idVendor=1d6b, idProduct=0002
[ 3.689759] usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[ 3.696961] usb usb1: Product: xHCI Host Controller
[ 3.701802] usb usb1: Manufacturer: Linux 4.14.0-xilinx-v2018.2 xhci-hcd
[ 3.708500] usb usb1: SerialNumber: xhci-hcd.0.auto
[ 3.713652] hub 1-0:1.0: USB hub found
[ 3.717378] hub 1-0:1.0: 1 port detected
[ 3.721459] xhci-hcd xhci-hcd.0.auto: xHCI Host Controller
[ 3.726927] xhci-hcd xhci-hcd.0.auto: new USB bus registered, assigned bus number 2
[ 3.734623] usb usb2: We don't know the algorithms for LPM for this host, disabling LPM.
[ 3.742780] usb usb2: New USB device found, idVendor=1d6b, idProduct=0003
[ 3.749550] usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[ 3.756756] usb usb2: Product: xHCI Host Controller
[ 3.761596] usb usb2: Manufacturer: Linux 4.14.0-xilinx-v2018.2 xhci-hcd
[ 3.770526] usb usb2: SerialNumber: xhci-hcd.0.auto
[ 3.777796] hub 2-0:1.0: USB hub found
[ 3.783731] hub 2-0:1.0: 1 port detected
[ 3.790031] dwc3 fe200000.dwc3: couldn't register cable notifier
[ 3.798262] dwc3 fe200000.dwc3: failed to initialize dual-role
[ 3.806361] platform fe200000.dwc3: runtime PM trying to suspend device but active child
[ 3.816760] dwc3: probe of fe200000.dwc3 failed with error -22
[ 3.825310] dwc3-of-simple ff9e0000.usb1: dwc3_simple_set_phydata: Can't find usb3-phy
[ 3.836885] xhci-hcd xhci-hcd.1.auto: xHCI Host Controller
[ 3.844667] xhci-hcd xhci-hcd.1.auto: new USB bus registered, assigned bus number 3
[ 3.854984] xhci-hcd xhci-hcd.1.auto: hcc params 0x0238f625 hci version 0x100 quirks 0x22010010
[ 3.866001] xhci-hcd xhci-hcd.1.auto: irq 49, io mem 0xfe300000
[ 3.874267] usb usb3: New USB device found, idVendor=1d6b, idProduct=0002
[ 3.883295] usb usb3: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[ 3.892734] usb usb3: Product: xHCI Host Controller
[ 3.899800] usb usb3: Manufacturer: Linux 4.14.0-xilinx-v2018.2 xhci-hcd
[ 3.908685] usb usb3: SerialNumber: xhci-hcd.1.auto
[ 3.915983] hub 3-0:1.0: USB hub found
[ 3.921863] hub 3-0:1.0: 1 port detected
[ 3.928036] xhci-hcd xhci-hcd.1.auto: xHCI Host Controller
[ 3.935609] xhci-hcd xhci-hcd.1.auto: new USB bus registered, assigned bus number 4
[ 3.945405] usb usb4: We don't know the algorithms for LPM for this host, disabling LPM.
[ 3.955680] usb usb4: New USB device found, idVendor=1d6b, idProduct=0003
[ 3.964547] usb usb4: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[ 3.973866] usb usb4: Product: xHCI Host Controller
[ 3.980824] usb usb4: Manufacturer: Linux 4.14.0-xilinx-v2018.2 xhci-hcd
[ 3.989600] usb usb4: SerialNumber: xhci-hcd.1.auto
[ 3.996758] hub 4-0:1.0: USB hub found
[ 4.002536] hub 4-0:1.0: 1 port detected
[ 4.008615] dwc3 fe300000.dwc3: couldn't register cable notifier
[ 4.016592] dwc3 fe300000.dwc3: failed to initialize dual-role
[ 4.024416] platform fe300000.dwc3: runtime PM trying to suspend device but active child
[ 4.034494] dwc3: probe of fe300000.dwc3 failed with error -22
[ 4.043378] rtc_zynqmp ffa60000.rtc: setting system clock to 1970-01-01 00:00:08 UTC (8)
[ 4.053601] clk: Not disabling unused clocks
[ 4.059922] ALSA device list:
[ 4.064865] #0: DisplayPort monitor
[ 4.071142] Freeing unused kernel memory: 512K
INIT: version 2.88 booting
Starting udev
[ 4.193831] udevd[1702]: starting version 3.2.2
[ 4.205490] udevd[1703]: starting eudev-3.2.2
Mon Feb 18 12:46:35 UTC 2019
Starting internet superserver: inetd.
Configuring packages on first boot....
(This may take several minutes. Please do not power off the machine.)
Running postinst /etc/rpm-postinsts/100-sysvinit-inittab...
update-rc.d: /etc/init.d/run-postinsts exists during rc.d purge (continuing)
INIT: Entering runlevel: 5
Configuring network interfaces... Cannot find device "eth0"
Starting Dropbear SSH server: Generating key, this may take a while...
Public key portion is:
ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQCUGFzO00efnXsncyQHUufMsAetLDsuUTWcifAJBzhnXPHcMZayHVUMxn93QVDtzEO+13NHy0mC5LY3uGb0e3PaZCF66aBR1yr7nHHeyxHbUMpm5TGBgbNhSP5kh4HYhCvKae3XZztuJlCxiYt76PkEjJ18bm2QDzePrN09nt5UooAhToGJ4I3aHdTPg5xH5gpQc6869OfaeUkJMnkoAvnYt5EOTaL6WV0fBH/pCUy2+Cs5ytqrBsJDmT4FTPnRrPY+ksm/JZ++5+O+x6u++g1PqqfM4yIZVglNY+hDdl0ArE7b7H6MjUaOwsnIwG4VlUX909APccrrfFWQ9xMzqDs/ root@cam_test_ov5642_182
Fingerprint: md5 6d:2f:72:10:6e:db:cc:32:d4:e4:fa:a4:b0:20:dd:f0
dropbear.
Starting syslogd/klogd: done
Starting tcf-agent: OK
PetaLinux 2018.2 cam_test_ov5642_182 /dev/ttyPS0
cam_test_ov5642_182 login:
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182$ LANG=en_US.UTF-8 petalinux-package --image -c kernel --format uImage
SDK environment now set up; additionally you may now run devtool to perform development tasks.
Run devtool --help for further details.
### Shell environment set up for builds. ###
You can now run 'bitbake <target>'
Common targets are:
core-image-minimal
core-image-sato
meta-toolchain
meta-ide-support
You can also run generated qemu images with a command like 'runqemu qemux86'
NOTE: Starting bitbake server...
INFO: Adding user layer: /home/masaaki/PetaLProj/cam_test_ov5642_182/project-spec/meta-user
NOTE: Starting bitbake server...
INFO: generating uImage
INFO: bitbake -R /home/masaaki/PetaLProj/cam_test_ov5642_182/build/conf/kerneltype.conf virtual/kernel
Parsing recipes: 100% |##########################################| Time: 0:01:25
Parsing of 2552 .bb files complete (0 cached, 2552 parsed). 3441 targets, 139 skipped, 0 masked, 0 errors.
NOTE: Resolving any missing task queue dependencies
Initialising tasks: 100% |#######################################| Time: 0:00:06
Checking sstate mirror object availability: 100% |###############| Time: 0:00:09
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
NOTE: fsbl: compiling from external source tree /home/masaaki/pkg/petalinux/tools/hsm/data/embeddedsw
NOTE: pmu-firmware: compiling from external source tree /home/masaaki/pkg/petalinux/tools/hsm/data/embeddedsw
WARNING: petalinux-user-image-1.0-r0 do_rootfs: [log_check] petalinux-user-image: found 1 warning message in the logfile:
[log_check] warning: %post(sysvinit-inittab-2.88dsf-r10.plnx_zynqmp) scriptlet failed, exit status 1
NOTE: Tasks Summary: Attempted 2383 tasks of which 2310 didn't need to be rerun and all succeeded.
Summary: There was 1 WARNING message shown.
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182
masaaki@722533dbc823:~/PetaLProj/cam_test_ov5642_182$ petalinux-boot --qemu --prebuilt 3
WARNING: Failed to load PMUFW, doesn't exist in prebuilts.See help to load custom PMUFW
petalinux-boot: line 1406: xxd: command not found
petalinux-boot: line 1417: xxd: command not found
petalinux-boot: line 1406: xxd: command not found
petalinux-boot: line 1417: xxd: command not found
petalinux-boot: line 1406: xxd: command not found
petalinux-boot: line 1417: xxd: command not found
INFO: No DTB has been specified, use the default one "/home/masaaki/PetaLProj/cam_test_ov5642_182/pre-built/linux/images/system.dtb".
INFO: No DTB has been specified, use the default one "/home/masaaki/PetaLProj/cam_test_ov5642_182/pre-built/linux/images/system.dtb".
INFO: Set QEMU tftp to /home/masaaki/PetaLProj/cam_test_ov5642_182/images/linux
petalinux-qemu-boot: line 88: xxd: command not found
INFO: TCP PORT is free
INFO: Starting aarch64 QEMU
INFO: qemu-system-aarch64 -M arm-generic-fdt -serial mon:stdio -serial /dev/null -display none -device loader,file=/tmp/tmp.ljJNJfCqew,cpu-num=0 -gdb tcp::9000 -dtb /home/masaaki/PetaLProj/cam_test_ov5642_182/pre-built/linux/images/system.dtb -hw-dtb /home/masaaki/PetaLProj/cam_test_ov5642_182/images/linux/zynqmp-qemu-multiarch-arm.dtb -machine-path /tmp/tmp.bdgrcojgBS -global xlnx,zynqmp-boot.cpu-num=0 -global xlnx,zynqmp-boot.use-pmufw=true -m 4G
qemu-system-aarch64: Failed to connect socket /tmp/tmp.bdgrcojgBS/qemu-rport-_pmu@0: No such file or directory
qemu-system-aarch64: info: QEMU waiting for connection on: disconnected:unix:/tmp/tmp.bdgrcojgBS/qemu-rport-_pmu@0,server
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182$ petalinux-config --get-hw-description=/home/masaaki/PetaLProj/cam_test_ov5642_182/cam_test_182.sdk/
INFO: Getting hardware description...
INFO: Rename cam_test_wrapper.hdf to system.hdf
[INFO] generating Kconfig for project
[INFO] menuconfig project
/home/masaaki/PetaLProj/cam_test_ov5642_182/build/misc/config/Kconfig.syshw:30:warning: defaults for choice values not supported
/home/masaaki/PetaLProj/cam_test_ov5642_182/build/misc/config/Kconfig:597:warning: config symbol defined without type
*** End of the configuration.
*** Execute 'make' to start the build or try 'make help'.
[INFO] sourcing bitbake
[INFO] generating plnxtool conf
[INFO] generating meta-plnx-generated layer
[INFO] generating machine configuration
[INFO] generating bbappends for project . This may take time !
[INFO] generating u-boot configuration files
[INFO] generating kernel configuration files
[INFO] generating kconfig for Rootfs
[INFO] oldconfig rootfs
[INFO] generating petalinux-user-image.bb
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182$
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182$ LANG=C petalinux-build
[INFO] building project
[INFO] sourcing bitbake
INFO: bitbake petalinux-user-image
Please use a locale setting which supports UTF-8 (such as LANG=en_US.UTF-8).
Python can't change the filesystem locale after loading so we need a UTF-8 when Python starts or things won't work.
ERROR: Failed to build project
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182$ LANG=en_US.UTF-8 petalinux-build
[INFO] building project
[INFO] sourcing bitbake
INFO: bitbake petalinux-user-image
Loading cache: 100% |############################################| Time: 0:00:00
Loaded 3439 entries from dependency cache.
Parsing recipes: 100% |##########################################| Time: 0:00:04
Parsing of 2552 .bb files complete (2509 cached, 43 parsed). 3441 targets, 139 skipped, 0 masked, 0 errors.
NOTE: Resolving any missing task queue dependencies
Initialising tasks: 100% |#######################################| Time: 0:00:08
Checking sstate mirror object availability: 100% |###############| Time: 0:00:14
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
NOTE: fsbl: compiling from external source tree /home/masaaki/pkg/petalinux/tools/hsm/data/embeddedsw
NOTE: pmu-firmware: compiling from external source tree /home/masaaki/pkg/petalinux/tools/hsm/data/embeddedsw
WARNING: petalinux-user-image-1.0-r0 do_rootfs: [log_check] petalinux-user-image: found 1 warning message in the logfile:
[log_check] warning: %post(sysvinit-inittab-2.88dsf-r10.plnx_zynqmp) scriptlet failed, exit status 1
NOTE: Tasks Summary: Attempted 3115 tasks of which 2259 didn't need to be rerun and all succeeded.
Summary: There was 1 WARNING message shown.
INFO: Copying Images from deploy to images
INFO: Creating images/linux directory
NOTE: Failed to copy built images to tftp dir: /tftpboot
[INFO] successfully built project
masaaki@d601aa5658aa:~/PetaLProj/cam_test_ov5642_182$
masaaki@6f49669c6591:~$ ./petalinux-v2018.2-final-installer.run /home/masaaki/pkg/petalinux/
INFO: Checking installer checksum...
INFO: Extracting PetaLinux installer...
LICENSE AGREEMENTS
PetaLinux SDK contains software from a number of sources. Please review
the following licenses and indicate your acceptance of each to continue.
You do not have to accept the licenses, however if you do not then you may
not use PetaLinux SDK.
Use PgUp/PgDn to navigate the license viewer, and press 'q' to close
Press Enter to display the license agreements
Do you accept Xilinx End User License Agreement? [y/N] > y
Do you accept Webtalk Terms and Conditions? [y/N] > y
Do you accept Third Party End User License Agreement? [y/N] > y
INFO: Checking installation environment requirements...
INFO: Checking free disk space
INFO: Checking installed tools
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 302: [: ==: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 304: [: ==: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 306: [: ==: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 251: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 251: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 251: [: =: unary operator expected
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 219: [: =: unary operator expected
INFO: Checking installed development libraries
/tmp/tmp.fW4nO5GJRg/./tools/common/petalinux/utils/petalinux-env-check: line 388: [: ==: unary operator expected
INFO: Checking network and other services
WARNING: No tftp server found - please refer to "PetaLinux SDK Installation Guide" for its impact and solution
INFO: Installing PetaLinux...
*********************************************
WARNING: PetaLinux installation directory: /home/masaaki/pkg/petalinux//. is not empty!
*********************************************
Please input "y" to continue to install PetaLinux in that directory?[n]y
INFO: Checking PetaLinux installer integrity...
INFO: Installing PetaLinux SDK to "/home/masaaki/pkg/petalinux//."
................................................................................................................................................................................................................................................................................INFO: Installing aarch64 Yocto SDK to "/home/masaaki/pkg/petalinux//./components/yocto/source/aarch64"...
PetaLinux Extensible SDK installer version 2018.2
=================================================
You are about to install the SDK to "/home/masaaki/pkg/petalinux/components/yocto/source/aarch64". Proceed[Y/n]? Y
Extracting SDK................................done
Setting it up...
Extracting buildtools...
done
SDK has been successfully set up and is ready to be used.
Each time you wish to use the SDK in a new shell session, you need to source the environment setup script e.g.
$ . /home/masaaki/pkg/petalinux/components/yocto/source/aarch64/environment-setup-aarch64-xilinx-linux
INFO: Installing arm Yocto SDK to "/home/masaaki/pkg/petalinux//./components/yocto/source/arm"...
PetaLinux Extensible SDK installer version 2018.2
=================================================
You are about to install the SDK to "/home/masaaki/pkg/petalinux/components/yocto/source/arm". Proceed[Y/n]? Y
Extracting SDK..............................done
Setting it up...
Extracting buildtools...
done
SDK has been successfully set up and is ready to be used.
Each time you wish to use the SDK in a new shell session, you need to source the environment setup script e.g.
$ . /home/masaaki/pkg/petalinux/components/yocto/source/arm/environment-setup-cortexa9hf-neon-xilinx-linux-gnueabi
INFO: Installing microblaze_full Yocto SDK to "/home/masaaki/pkg/petalinux//./components/yocto/source/microblaze_full"...
PetaLinux Extensible SDK installer version 2018.2
=================================================
You are about to install the SDK to "/home/masaaki/pkg/petalinux/components/yocto/source/microblaze_full". Proceed[Y/n]? Y
Extracting SDK.............................done
Setting it up...
Extracting buildtools...
done
SDK has been successfully set up and is ready to be used.
Each time you wish to use the SDK in a new shell session, you need to source the environment setup script e.g.
$ . /home/masaaki/pkg/petalinux/components/yocto/source/microblaze_full/environment-setup-microblazeel-v10.0-bs-cmp-re-mh-div-xilinx-linux
INFO: Installing microblaze_lite Yocto SDK to "/home/masaaki/pkg/petalinux//./components/yocto/source/microblaze_lite"...
PetaLinux Extensible SDK installer version 2018.2
=================================================
You are about to install the SDK to "/home/masaaki/pkg/petalinux/components/yocto/source/microblaze_lite". Proceed[Y/n]? Y
Extracting SDK.............................done
Setting it up...
Extracting buildtools...
done
SDK has been successfully set up and is ready to be used.
Each time you wish to use the SDK in a new shell session, you need to source the environment setup script e.g.
$ . /home/masaaki/pkg/petalinux/components/yocto/source/microblaze_lite/environment-setup-microblazeel-v10.0-bs-cmp-re-ml-xilinx-linux
INFO: PetaLinux SDK has been installed to /home/masaaki/pkg/petalinux//.
masaaki@6f49669c6591:~$
FROM ubuntu:16.04
RUN apt-get update && apt-get install -y x11-apps
RUN apt-get install -y sudo
RUN apt-get install -y net-tools gedit nautilus
# Replace 1000 with your user / group id
ARG uid=1000
ARG gid=1000
RUN groupadd -g ${uid} masaaki && \
useradd -u ${gid} -g masaaki -G sudo -r masaaki && \
mkdir /home/masaaki && \
chown ${uid}:${gid} -R /home/masaaki
RUN echo 'Defaults visiblepw' >> /etc/sudoers
RUN echo 'masaaki ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers
USER masaaki
WORKDIR /home/masaaki
docker build -t vivado182ub16 .
docker run -ti --rm --name vivado182ub16c -e DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix -v /home/masaaki/Docker/vivado182ub16/masaaki:/home/masaaki vivado182ub16 /bin/bash
FROM ubuntu:16.04
RUN apt-get update && apt-get install -y x11-apps
RUN apt-get install -y sudo
# Replace 1000 with your user / group id
ARG uid=1000
ARG gid=1000
RUN groupadd -g ${uid} masaaki && \
useradd -u ${gid} -g masaaki -G sudo -r masaaki && \
mkdir /home/masaaki && \
chown ${uid}:${gid} -R /home/masaaki
RUN echo 'Defaults visiblepw' >> /etc/sudoers
RUN echo 'masaaki ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers
USER masaaki
WORKDIR /home/masaaki
docker build -t ubuntu16gui .
docker run -ti --rm --name ubuntu16gui -e DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix ubuntu16gui /bin/bash
docker container run ubuntu:latest /bin/echo ‘Hello world’
docker: Got permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock: Post http://%2Fvar%2Frun%2Fdocker.sock/v1.39/containers/create: dial unix /var/run/docker.sock: connect: permission denied.
See 'docker run --help'.
masaaki@masaaki-H110M4-M01:~$ docker container run ubuntu:latest /bin/echo ‘Hello world’
Unable to find image 'ubuntu:latest' locally
latest: Pulling from library/ubuntu
6cf436f81810: Pull complete
987088a85b96: Pull complete
b4624b3efe06: Pull complete
d42beb8ded59: Pull complete
Digest: sha256:7a47ccc3bbe8a451b500d2b53104868b46d60ee8f5b35a24b41a86077c650210
Status: Downloaded newer image for ubuntu:latest
‘Hello world’
#!/bin/bash
sudo ./memwrite fd4ab070 54
sudo ./memwrite fd4aa00c 01
// cam_dp_ov5642.cpp (for Ultra96)
// 2018/12/14 by marsee
//
// This software converts the left and right of the camera image to BMP file.
// -b : bmp file name
// -n : Start File Number
// -h : help
//
// 2018/12/20 : completed.
// I am using the SVGA driver register setting of https://github.com/virajkanwade/rk3188_android_kernel/blob/master/drivers/media/video/ov5642.c
// 2018/12/22 : fixed
// 2018/12/30 : ov5642_inf_axis[0] fixed
// 2019/02/06 : for DisplayPort
#include <opencv2/opencv.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <assert.h>
#include <sys/mman.h>
#include <fcntl.h>
#include <string.h>
#define PIXEL_NUM_OF_BYTES 4
#define NUMBER_OF_WRITE_FRAMES 3
#define SVGA_HORIZONTAL_PIXELS 800
#define SVGA_VERTICAL_LINES 600
#define SVGA_ALL_DISP_ADDRESS (SVGA_HORIZONTAL_PIXELS * SVGA_VERTICAL_LINES * PIXEL_NUM_OF_BYTES)
#define SVGA_3_PICTURES (SVGA_ALL_DISP_ADDRESS * NUMBER_OF_WRITE_FRAMES)
#define XGA_HORIZONTAL_PIXELS 1024
#define XGA_VERTICAL_LINES 768
#define XGA_ALL_DISP_ADDRESS (XGA_HORIZONTAL_PIXELS * XGA_VERTICAL_LINES * PIXEL_NUM_OF_BYTES)
#define XGA_3_PICTURES (XGA_ALL_DISP_ADDRESS * NUMBER_OF_WRITE_FRAMES)
#define HD_HORIZONTAL_PIXELS 1920
#define HD_VERTICAL_LINES 1080
#define HD_ALL_DISP_ADDRESS (HD_HORIZONTAL_PIXELS * HD_VERTICAL_LINES * PIXEL_NUM_OF_BYTES)
#define HD_3_PICTURES (HD_ALL_DISP_ADDRESS * NUMBER_OF_WRITE_FRAMES)
int WriteBMPfile(char *bmp_file, volatile unsigned int *frame_buffer, int active_frame, int resolution);
void cam_i2c_init(volatile unsigned *ov5642_axi_iic) {
ov5642_axi_iic[64] = 0x2; // reset tx fifo ,address is 0x100, i2c_control_reg
ov5642_axi_iic[64] = 0x1; // enable i2c
}
void cam_i2x_write_sync(void) {
// unsigned c;
// c = *cam_i2c_rx_fifo;
// while ((c & 0x84) != 0x80)
// c = *cam_i2c_rx_fifo; // No Bus Busy and TX_FIFO_Empty = 1
usleep(1000);
}
void cam_i2c_write(volatile unsigned *ov5642_axi_iic, unsigned int device_addr, unsigned int write_addr, unsigned int write_data){
ov5642_axi_iic[66] = 0x100 | (device_addr & 0xfe); // Slave IIC Write Address, address is 0x108, i2c_tx_fifo
ov5642_axi_iic[66] = (write_addr >> 8) & 0xff; // address upper byte
ov5642_axi_iic[66] = write_addr & 0xff; // address lower byte
ov5642_axi_iic[66] = 0x200 | (write_data & 0xff); // data
cam_i2x_write_sync();
}
int cam_reg_set(volatile unsigned *axi_iic, unsigned int device_addr);
int main(int argc, char *argv[]){
int opt;
int c, help_flag=0;
char bmp_fn[256] = "bmp_file";
char attr[1024];
unsigned long phys_addr;
int file_no = -1;
int fd1, fd2, fd3, fd4, fd5, fd6, fd10, fd11;
volatile unsigned int *ov5642_inf_axis, *axi_iic, *disp_dmar_axis, *vflip_dma_write;
volatile unsigned int *axi_gpio_0, *axi_gpio_1;
volatile unsigned int *frame_buffer;
int active_frame;
int resolution;
int all_disp_addr;
resolution = 1; // XGA
while ((opt=getopt(argc, argv, "b:n:h:r:")) != -1){
switch (opt){
case 'b':
strcpy(bmp_fn, optarg);
break;
case 'n':
file_no = atoi(optarg);
printf("file_no = %d\n", file_no+1);
break;
case 'r':
resolution = atoi(optarg);
break;
case 'h':
help_flag = 1;
break;
}
}
if(resolution == 0){
printf("SVGA\n");
} else if(resolution == 1){
printf("XGA\n");
} else {
printf("HD\n");
}
if (help_flag == 1){ // help
printf("Usage : cam_capture [-b <bmp file name>] [-n <Start File Number>] [-h]\n");
printf(" -r [0|1|2](0:SVGA, 1:XGA, 2:HD)\n");
exit(0);
}
// all_disp_addr
switch(resolution){
case 0 :
all_disp_addr = SVGA_ALL_DISP_ADDRESS;
break;
case 1 :
all_disp_addr = XGA_ALL_DISP_ADDRESS;
break;
default : // 2
all_disp_addr = HD_ALL_DISP_ADDRESS;
break;
}
// ov5642_inf_axis-uio IP
fd1 = open("/dev/uio1", O_RDWR|O_SYNC); // Read/Write, The chache is disable
if (fd1 < 1){
fprintf(stderr, "/dev/uio1 (ov5642_inf_axis) open error\n");
exit(-1);
}
ov5642_inf_axis = (volatile unsigned *)mmap(NULL, 0x1000, PROT_READ|PROT_WRITE, MAP_SHARED, fd1, 0);
if (!ov5642_inf_axis){
fprintf(stderr, "ov5642_inf_axis mmap error\n");
exit(-1);
}
// axi_iic-uio IP
fd2 = open("/dev/uio2", O_RDWR|O_SYNC); // Read/Write, The chache is disable
if (fd2 < 1){
fprintf(stderr, "/dev/uio2 (axi_iic) open error\n");
exit(-1);
}
axi_iic = (volatile unsigned int *)mmap(NULL, 0x1000, PROT_READ|PROT_WRITE, MAP_SHARED, fd2, 0);
if (!axi_iic){
fprintf(stderr, "axi_iic mmap error\n");
exit(-1);
}
// disp_dmar_axis-uio IP
fd3 = open("/dev/uio3", O_RDWR|O_SYNC); // Read/Write, The chache is disable
if (fd3 < 1){
fprintf(stderr, "/dev/uio3 (disp_dmar_axis) open error\n");
exit(-1);
}
disp_dmar_axis = (volatile unsigned int *)mmap(NULL, 0x10000, PROT_READ|PROT_WRITE, MAP_SHARED, fd3, 0);
if (!disp_dmar_axis){
fprintf(stderr, "disp_dmar_axis mmap error\n");
exit(-1);
}
// vflip_dma_write-uio IP
fd4 = open("/dev/uio4", O_RDWR|O_SYNC); // Read/Write, The chache is disable
if (fd4 < 1){
fprintf(stderr, "/dev/uio4 (vflip_dma_write) open error\n");
exit(-1);
}
vflip_dma_write = (volatile unsigned int *)mmap(NULL, 0x10000, PROT_READ|PROT_WRITE, MAP_SHARED, fd4, 0);
if (!vflip_dma_write){
fprintf(stderr, "vflip_dma_write mmap error\n");
exit(-1);
}
// axi_gpio_0-uio IP (init_done output)
fd5 = open("/dev/uio5", O_RDWR|O_SYNC); // Read/Write, The chache is disable
if (fd5 < 1){
fprintf(stderr, "/dev/uio5 (axi_gpio_0) open error\n");
exit(-1);
}
axi_gpio_0 = (volatile unsigned int *)mmap(NULL, 0x1000, PROT_READ|PROT_WRITE, MAP_SHARED, fd5, 0);
if (!axi_gpio_0){
fprintf(stderr, "axi_gpio_0 mmap error\n");
exit(-1);
}
// axi_gpio_1-uio IP (active_frame input)
fd6 = open("/dev/uio6", O_RDWR|O_SYNC); // Read/Write, The chache is disable
if (fd6 < 1){
fprintf(stderr, "/dev/uio6 (axi_gpio_1) open error\n");
exit(-1);
}
axi_gpio_1 = (volatile unsigned int *)mmap(NULL, 0x1000, PROT_READ|PROT_WRITE, MAP_SHARED, fd6, 0);
if (!axi_gpio_1){
fprintf(stderr, "axi_gpio_1 mmap error\n");
exit(-1);
}
// udmabuf4
fd10 = open("/dev/udmabuf4", O_RDWR | O_SYNC); // frame_buffer, The chache is disabled.
if (fd10 == -1){
fprintf(stderr, "/dev/udmabuf4 open error\n");
exit(-1);
}
frame_buffer = (volatile unsigned int *)mmap(NULL, all_disp_addr*NUMBER_OF_WRITE_FRAMES, PROT_READ|PROT_WRITE, MAP_SHARED, fd10, 0);
if (!frame_buffer){
fprintf(stderr, "frame_buffer4 mmap error\n");
exit(-1);
}
// phys_addr of udmabuf4
fd11 = open("/sys/class/udmabuf/udmabuf4/phys_addr", O_RDONLY);
if (fd11 == -1){
fprintf(stderr, "/sys/class/udmabuf/udmabuf4/phys_addr open error\n");
exit(-1);
}
read(fd11, attr, 1024);
sscanf(attr, "%lx", &phys_addr);
close(fd11);
printf("phys_addr = %x\n", (int)phys_addr);
// vflip_dma_write start
vflip_dma_write[6] = phys_addr; // fb0
vflip_dma_write[8] = phys_addr+all_disp_addr; // fb1
vflip_dma_write[10] = phys_addr+2*all_disp_addr; // fb2
vflip_dma_write[12] = resolution;
vflip_dma_write[0] = 0x1; // start
vflip_dma_write[0] = 0x80; // EnableAutoRestart
// CMOS Camera initialize, ov5642
cam_i2c_init(axi_iic);
cam_reg_set(axi_iic, 0x78); // OV5642 register set
ov5642_inf_axis[0] = phys_addr; // ov5642 AXI4-Stream Start
ov5642_inf_axis[1] = 0;
// disp_dmar_axis start
disp_dmar_axis[4] = phys_addr; // fb0
disp_dmar_axis[6] = phys_addr+all_disp_addr; // fb1
disp_dmar_axis[8] = phys_addr+2*all_disp_addr; // fb2
disp_dmar_axis[10] = resolution;
axi_gpio_0[0] = 1; // disp_dmar_axis start(init_done = 1)
char bmp_file[256];
// All 0 set
int all_disp_frame_index = all_disp_addr/PIXEL_NUM_OF_BYTES*NUMBER_OF_WRITE_FRAMES;
for (int i=0; i<all_disp_frame_index; i++){
frame_buffer[i] = 0x0;
}
// w - writed the left and right eye's bmp files. q - exit.
c = getc(stdin);
while(c != 'q'){
switch ((char)c) {
case 'w' : // w - writed a bmp files.
// writed the frame buffer
file_no++;
sprintf(bmp_file, "%s%d.bmp", bmp_fn, file_no);
active_frame = (int)(axi_gpio_1[0] & 0x3); // Data signal of active_frame_V
WriteBMPfile(bmp_file, frame_buffer, active_frame, resolution);
printf("file No. = %d\n", file_no);
break;
case 'e' : // e - writed a same bmp files.
// writed the frame buffer
if (file_no == -1)
file_no = 0;
sprintf(bmp_file, "%s%d.bmp", bmp_fn, file_no);
active_frame = (int)(axi_gpio_1[0] & 0x3); // Data signal of active_frame_V
WriteBMPfile(bmp_file, frame_buffer, active_frame, resolution);
printf("file No. = %d\n", file_no);
break;
}
c = getc(stdin);
}
munmap((void *)ov5642_inf_axis, 0x1000);
munmap((void *)axi_iic, 0x1000);
munmap((void *)disp_dmar_axis, 0x10000);
munmap((void *)vflip_dma_write, 0x10000);
munmap((void *)axi_gpio_0, 0x1000);
munmap((void *)axi_gpio_1, 0x1000);
munmap((void *)frame_buffer, all_disp_addr*3);
close(fd1);
close(fd2);
close(fd3);
close(fd4);
close(fd5);
close(fd6);
close(fd10);
return(0);
}
int WriteBMPfile(char *bmp_file, volatile unsigned int *frame_buffer, int active_frame, int resolution){
int read_frame;
int img_width, img_height;
if (active_frame == 0)
read_frame = 2;
else if (active_frame == 1)
read_frame = 0;
else // active_frame == 2
read_frame = 1;
switch(resolution){
case 0 :
img_width = SVGA_HORIZONTAL_PIXELS;
img_height = SVGA_VERTICAL_LINES;
break;
case 1 :
img_width = XGA_HORIZONTAL_PIXELS;
img_height = XGA_VERTICAL_LINES;
break;
default : // case 2 :
img_width = HD_HORIZONTAL_PIXELS;
img_height = HD_VERTICAL_LINES;
break;
}
int offset_addr = read_frame * img_width * img_height;
cv::Mat img(img_height, img_width, CV_8UC3);
cv::Mat_<cv::Vec3b> dst_vec3b = cv::Mat_<cv::Vec3b>(img);
for(int y=0; y<img.rows; y++){
for(int x=0; x<img.cols; x++){
cv::Vec3b pixel;
int rgb = frame_buffer[offset_addr+y*img.cols+x];
pixel[0] = (rgb & 0xff); // blue
pixel[1] = (rgb & 0xff00) >> 8; // green
pixel[2] = (rgb & 0xff0000) >> 16; // red
dst_vec3b(y,x) = pixel;
}
}
cv::imwrite(bmp_file, img);
return(0);
}
int cam_reg_set(volatile unsigned *axi_iic, unsigned int device_addr){
cam_i2c_write(axi_iic, device_addr, 0x3103, 0x93);
cam_i2c_write(axi_iic, device_addr, 0x3008, 0x82);
cam_i2c_write(axi_iic, device_addr, 0x3017, 0x7f);
cam_i2c_write(axi_iic, device_addr, 0x3018, 0xfc);
cam_i2c_write(axi_iic, device_addr, 0x3810, 0xc2);
cam_i2c_write(axi_iic, device_addr, 0x3615, 0xf0);
cam_i2c_write(axi_iic, device_addr, 0x3000, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3001, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3002, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3003, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3000, 0xf8);
cam_i2c_write(axi_iic, device_addr, 0x3001, 0x48);
cam_i2c_write(axi_iic, device_addr, 0x3002, 0x5c);
cam_i2c_write(axi_iic, device_addr, 0x3003, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x3004, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x3005, 0xb7);
cam_i2c_write(axi_iic, device_addr, 0x3006, 0x43);
cam_i2c_write(axi_iic, device_addr, 0x3007, 0x37);
cam_i2c_write(axi_iic, device_addr, 0x3011, 0x08); // 0x08 - 15fps, 0x10 - 30fps
cam_i2c_write(axi_iic, device_addr, 0x3010, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x460c, 0x22);
cam_i2c_write(axi_iic, device_addr, 0x3815, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x370d, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x370c, 0xa0);
cam_i2c_write(axi_iic, device_addr, 0x3602, 0xfc);
cam_i2c_write(axi_iic, device_addr, 0x3612, 0xff);
cam_i2c_write(axi_iic, device_addr, 0x3634, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x3613, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3605, 0x7c);
cam_i2c_write(axi_iic, device_addr, 0x3621, 0x09);
cam_i2c_write(axi_iic, device_addr, 0x3622, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3604, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x3603, 0xa7);
cam_i2c_write(axi_iic, device_addr, 0x3603, 0x27);
cam_i2c_write(axi_iic, device_addr, 0x4000, 0x21);
cam_i2c_write(axi_iic, device_addr, 0x401d, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x3600, 0x54);
cam_i2c_write(axi_iic, device_addr, 0x3605, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x3606, 0x3f);
cam_i2c_write(axi_iic, device_addr, 0x3c01, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5000, 0x4f);
cam_i2c_write(axi_iic, device_addr, 0x5020, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x5181, 0x79);
cam_i2c_write(axi_iic, device_addr, 0x5182, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5185, 0x22);
cam_i2c_write(axi_iic, device_addr, 0x5197, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x5001, 0xff);
cam_i2c_write(axi_iic, device_addr, 0x5500, 0x0a);
cam_i2c_write(axi_iic, device_addr, 0x5504, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5505, 0x7f);
cam_i2c_write(axi_iic, device_addr, 0x5080, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x300e, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x4610, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x471d, 0x05);
cam_i2c_write(axi_iic, device_addr, 0x4708, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x3710, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x3632, 0x41);
cam_i2c_write(axi_iic, device_addr, 0x3702, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x3620, 0x37);
cam_i2c_write(axi_iic, device_addr, 0x3631, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x3808, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x3809, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x380a, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x380b, 0xe0);
cam_i2c_write(axi_iic, device_addr, 0x380e, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x380f, 0xd0);
cam_i2c_write(axi_iic, device_addr, 0x501f, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5000, 0x4f);
cam_i2c_write(axi_iic, device_addr, 0x4300, 0x61); // RGB565
cam_i2c_write(axi_iic, device_addr, 0x3503, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x3501, 0x73);
cam_i2c_write(axi_iic, device_addr, 0x3502, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x350b, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3503, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x3824, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x3501, 0x1e);
cam_i2c_write(axi_iic, device_addr, 0x3502, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x350b, 0x7f);
cam_i2c_write(axi_iic, device_addr, 0x380c, 0x0c);
cam_i2c_write(axi_iic, device_addr, 0x380d, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x380e, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x380f, 0xe8);
cam_i2c_write(axi_iic, device_addr, 0x3a0d, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x3a0e, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x3818, 0xc1);
cam_i2c_write(axi_iic, device_addr, 0x3705, 0xdb);
cam_i2c_write(axi_iic, device_addr, 0x370a, 0x81);
cam_i2c_write(axi_iic, device_addr, 0x3801, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x3621, 0xc7);
cam_i2c_write(axi_iic, device_addr, 0x3801, 0x50);
cam_i2c_write(axi_iic, device_addr, 0x3803, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x3827, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x3810, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x3804, 0x05);
cam_i2c_write(axi_iic, device_addr, 0x3805, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5682, 0x05);
cam_i2c_write(axi_iic, device_addr, 0x5683, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3806, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x3807, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x5686, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x5687, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x3a00, 0x78);
cam_i2c_write(axi_iic, device_addr, 0x3a1a, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x3a13, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x3a18, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3a19, 0x7c);
cam_i2c_write(axi_iic, device_addr, 0x3a08, 0x12);
cam_i2c_write(axi_iic, device_addr, 0x3a09, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x3a0a, 0x0f);
cam_i2c_write(axi_iic, device_addr, 0x3a0b, 0xa0);
cam_i2c_write(axi_iic, device_addr, 0x3004, 0xff);
cam_i2c_write(axi_iic, device_addr, 0x350c, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x350d, 0xd0);
cam_i2c_write(axi_iic, device_addr, 0x3500, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3501, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3502, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x350a, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x350b, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3503, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x528a, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x528b, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x528c, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x528d, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x528e, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x528f, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5290, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5292, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5293, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5294, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5295, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5296, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5297, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5298, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5299, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x529a, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529b, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x529c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529d, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x529e, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529f, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x3a0f, 0x3c);
cam_i2c_write(axi_iic, device_addr, 0x3a10, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x3a1b, 0x3c);
cam_i2c_write(axi_iic, device_addr, 0x3a1e, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x3a11, 0x70);
cam_i2c_write(axi_iic, device_addr, 0x3a1f, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x3030, 0x0b);
cam_i2c_write(axi_iic, device_addr, 0x3a02, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3a03, 0x7d);
cam_i2c_write(axi_iic, device_addr, 0x3a04, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3a14, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3a15, 0x7d);
cam_i2c_write(axi_iic, device_addr, 0x3a16, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3a00, 0x78);
cam_i2c_write(axi_iic, device_addr, 0x3a08, 0x09);
cam_i2c_write(axi_iic, device_addr, 0x3a09, 0x60);
cam_i2c_write(axi_iic, device_addr, 0x3a0a, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x3a0b, 0xd0);
cam_i2c_write(axi_iic, device_addr, 0x3a0d, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x3a0e, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x5193, 0x70);
cam_i2c_write(axi_iic, device_addr, 0x3620, 0x57);
cam_i2c_write(axi_iic, device_addr, 0x3703, 0x98);
cam_i2c_write(axi_iic, device_addr, 0x3704, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x589b, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x589a, 0xc5);
cam_i2c_write(axi_iic, device_addr, 0x528a, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x528b, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x528c, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x528d, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x528e, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x528f, 0x28);
cam_i2c_write(axi_iic, device_addr, 0x5290, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5292, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5293, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5294, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5295, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5296, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5297, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5298, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5299, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x529a, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529b, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x529c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529d, 0x28);
cam_i2c_write(axi_iic, device_addr, 0x529e, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529f, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5282, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5300, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5301, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5302, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5303, 0x7c);
cam_i2c_write(axi_iic, device_addr, 0x530c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x530d, 0x0c);
cam_i2c_write(axi_iic, device_addr, 0x530e, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x530f, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5310, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5311, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5308, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5309, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x5304, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5305, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5306, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5307, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5314, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5315, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5319, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5316, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5317, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5318, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5380, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x5381, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5382, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5383, 0x4e);
cam_i2c_write(axi_iic, device_addr, 0x5384, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5385, 0x0f);
cam_i2c_write(axi_iic, device_addr, 0x5386, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5387, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5388, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x5389, 0x15);
cam_i2c_write(axi_iic, device_addr, 0x538a, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538b, 0x31);
cam_i2c_write(axi_iic, device_addr, 0x538c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538d, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538e, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538f, 0x0f);
cam_i2c_write(axi_iic, device_addr, 0x5390, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5391, 0xab);
cam_i2c_write(axi_iic, device_addr, 0x5392, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5393, 0xa2);
cam_i2c_write(axi_iic, device_addr, 0x5394, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5480, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5481, 0x21);
cam_i2c_write(axi_iic, device_addr, 0x5482, 0x36);
cam_i2c_write(axi_iic, device_addr, 0x5483, 0x57);
cam_i2c_write(axi_iic, device_addr, 0x5484, 0x65);
cam_i2c_write(axi_iic, device_addr, 0x5485, 0x71);
cam_i2c_write(axi_iic, device_addr, 0x5486, 0x7d);
cam_i2c_write(axi_iic, device_addr, 0x5487, 0x87);
cam_i2c_write(axi_iic, device_addr, 0x5488, 0x91);
cam_i2c_write(axi_iic, device_addr, 0x5489, 0x9a);
cam_i2c_write(axi_iic, device_addr, 0x548a, 0xaa);
cam_i2c_write(axi_iic, device_addr, 0x548b, 0xb8);
cam_i2c_write(axi_iic, device_addr, 0x548c, 0xcd);
cam_i2c_write(axi_iic, device_addr, 0x548d, 0xdd);
cam_i2c_write(axi_iic, device_addr, 0x548e, 0xea);
cam_i2c_write(axi_iic, device_addr, 0x548f, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5490, 0x05);
cam_i2c_write(axi_iic, device_addr, 0x5491, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5492, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x5493, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5494, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x5495, 0x60);
cam_i2c_write(axi_iic, device_addr, 0x5496, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5497, 0xb8);
cam_i2c_write(axi_iic, device_addr, 0x5498, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5499, 0x86);
cam_i2c_write(axi_iic, device_addr, 0x549a, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x549b, 0x5b);
cam_i2c_write(axi_iic, device_addr, 0x549c, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x549d, 0x3b);
cam_i2c_write(axi_iic, device_addr, 0x549e, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x549f, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x54a0, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x54a1, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x54a2, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54a3, 0xed);
cam_i2c_write(axi_iic, device_addr, 0x54a4, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54a5, 0xc5);
cam_i2c_write(axi_iic, device_addr, 0x54a6, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54a7, 0xa5);
cam_i2c_write(axi_iic, device_addr, 0x54a8, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54a9, 0x6c);
cam_i2c_write(axi_iic, device_addr, 0x54aa, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54ab, 0x41);
cam_i2c_write(axi_iic, device_addr, 0x54ac, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54ad, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x54ae, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x54af, 0x16);
cam_i2c_write(axi_iic, device_addr, 0x3406, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5192, 0x04); // 0x04
cam_i2c_write(axi_iic, device_addr, 0x5191, 0xf8); // 0xf8
cam_i2c_write(axi_iic, device_addr, 0x5193, 0x70);
cam_i2c_write(axi_iic, device_addr, 0x5194, 0xf0);
cam_i2c_write(axi_iic, device_addr, 0x5195, 0xf0);
cam_i2c_write(axi_iic, device_addr, 0x518d, 0x3d);
cam_i2c_write(axi_iic, device_addr, 0x518f, 0x54);
cam_i2c_write(axi_iic, device_addr, 0x518e, 0x3d);
cam_i2c_write(axi_iic, device_addr, 0x5190, 0x54);
cam_i2c_write(axi_iic, device_addr, 0x518b, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x518c, 0xbd);
cam_i2c_write(axi_iic, device_addr, 0x5187, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x5188, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x5189, 0x6e);
cam_i2c_write(axi_iic, device_addr, 0x518a, 0x68);
cam_i2c_write(axi_iic, device_addr, 0x5186, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x5181, 0x50);
cam_i2c_write(axi_iic, device_addr, 0x5184, 0x25);
cam_i2c_write(axi_iic, device_addr, 0x5182, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x5183, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5184, 0x25);
cam_i2c_write(axi_iic, device_addr, 0x5185, 0x24);
cam_i2c_write(axi_iic, device_addr, 0x5025, 0x82);
cam_i2c_write(axi_iic, device_addr, 0x5583, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x5584, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x5580, 0x02); // 0x02
cam_i2c_write(axi_iic, device_addr, 0x3633, 0x07);
cam_i2c_write(axi_iic, device_addr, 0x3702, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x3703, 0xb2);
cam_i2c_write(axi_iic, device_addr, 0x3704, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x370b, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x370d, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x3620, 0x52);
cam_i2c_write(axi_iic, device_addr, 0x3c00, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x5001, 0xFF);
cam_i2c_write(axi_iic, device_addr, 0x5282, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5300, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5301, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5302, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5303, 0x7c);
cam_i2c_write(axi_iic, device_addr, 0x530c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x530d, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x530e, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x530f, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5310, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5311, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5308, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5309, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x5304, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5305, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5306, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5307, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5314, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5315, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5319, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5316, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5317, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5318, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5500, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5502, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5503, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x5504, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5505, 0x7f);
cam_i2c_write(axi_iic, device_addr, 0x5025, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5300, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5301, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5302, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5303, 0x7c);
cam_i2c_write(axi_iic, device_addr, 0x530c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x530d, 0x0c);
cam_i2c_write(axi_iic, device_addr, 0x530e, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x530f, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5310, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5311, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5308, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5309, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x5304, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5305, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5306, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5307, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5314, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5315, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x5319, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5316, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5317, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5318, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5380, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x5381, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5382, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5383, 0x1f);
cam_i2c_write(axi_iic, device_addr, 0x5384, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5385, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x5386, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5387, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5388, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5389, 0xE1);
cam_i2c_write(axi_iic, device_addr, 0x538A, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538B, 0x2B);
cam_i2c_write(axi_iic, device_addr, 0x538C, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538D, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538E, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x538F, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5390, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5391, 0xB3);
cam_i2c_write(axi_iic, device_addr, 0x5392, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5393, 0xA6);
cam_i2c_write(axi_iic, device_addr, 0x5394, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5480, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5481, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x5482, 0x2a);
cam_i2c_write(axi_iic, device_addr, 0x5483, 0x49);
cam_i2c_write(axi_iic, device_addr, 0x5484, 0x56);
cam_i2c_write(axi_iic, device_addr, 0x5485, 0x62);
cam_i2c_write(axi_iic, device_addr, 0x5486, 0x6c);
cam_i2c_write(axi_iic, device_addr, 0x5487, 0x76);
cam_i2c_write(axi_iic, device_addr, 0x5488, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x5489, 0x88);
cam_i2c_write(axi_iic, device_addr, 0x548a, 0x96);
cam_i2c_write(axi_iic, device_addr, 0x548b, 0xa2);
cam_i2c_write(axi_iic, device_addr, 0x548c, 0xb8);
cam_i2c_write(axi_iic, device_addr, 0x548d, 0xcc);
cam_i2c_write(axi_iic, device_addr, 0x548e, 0xe0);
cam_i2c_write(axi_iic, device_addr, 0x548f, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5490, 0x3);
cam_i2c_write(axi_iic, device_addr, 0x5491, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x5492, 0x3);
cam_i2c_write(axi_iic, device_addr, 0x5493, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5494, 0x2);
cam_i2c_write(axi_iic, device_addr, 0x5495, 0xa0);
cam_i2c_write(axi_iic, device_addr, 0x5496, 0x2);
cam_i2c_write(axi_iic, device_addr, 0x5497, 0x48);
cam_i2c_write(axi_iic, device_addr, 0x5498, 0x2);
cam_i2c_write(axi_iic, device_addr, 0x5499, 0x26);
cam_i2c_write(axi_iic, device_addr, 0x549a, 0x2);
cam_i2c_write(axi_iic, device_addr, 0x549b, 0xb);
cam_i2c_write(axi_iic, device_addr, 0x549c, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x549d, 0xee);
cam_i2c_write(axi_iic, device_addr, 0x549e, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x549f, 0xd8);
cam_i2c_write(axi_iic, device_addr, 0x54a0, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54a1, 0xc7);
cam_i2c_write(axi_iic, device_addr, 0x54a2, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54a3, 0xb3);
cam_i2c_write(axi_iic, device_addr, 0x54a4, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54a5, 0x90);
cam_i2c_write(axi_iic, device_addr, 0x54a6, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54a7, 0x62);
cam_i2c_write(axi_iic, device_addr, 0x54a8, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54a9, 0x27);
cam_i2c_write(axi_iic, device_addr, 0x54aa, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54ab, 0x09);
cam_i2c_write(axi_iic, device_addr, 0x54ac, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x54ad, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x54ae, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x54af, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x54b0, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54b1, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x54b2, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54b3, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x54b4, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x54b5, 0xf0);
cam_i2c_write(axi_iic, device_addr, 0x54b6, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x54b7, 0xdf);
cam_i2c_write(axi_iic, device_addr, 0x5583, 0x5d);
cam_i2c_write(axi_iic, device_addr, 0x5584, 0x5d);
cam_i2c_write(axi_iic, device_addr, 0x5580, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x5587, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5588, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x558a, 0x09);
cam_i2c_write(axi_iic, device_addr, 0x5589, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5000, 0xcf);
cam_i2c_write(axi_iic, device_addr, 0x5800, 0x48);
cam_i2c_write(axi_iic, device_addr, 0x5801, 0x31);
cam_i2c_write(axi_iic, device_addr, 0x5802, 0x21);
cam_i2c_write(axi_iic, device_addr, 0x5803, 0x1b);
cam_i2c_write(axi_iic, device_addr, 0x5804, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x5805, 0x1e);
cam_i2c_write(axi_iic, device_addr, 0x5806, 0x29);
cam_i2c_write(axi_iic, device_addr, 0x5807, 0x38);
cam_i2c_write(axi_iic, device_addr, 0x5808, 0x26);
cam_i2c_write(axi_iic, device_addr, 0x5809, 0x17);
cam_i2c_write(axi_iic, device_addr, 0x580a, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x580b, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x580c, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x580d, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x580e, 0x13);
cam_i2c_write(axi_iic, device_addr, 0x580f, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x5810, 0x15);
cam_i2c_write(axi_iic, device_addr, 0x5811, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5812, 0x8);
cam_i2c_write(axi_iic, device_addr, 0x5813, 0x5);
cam_i2c_write(axi_iic, device_addr, 0x5814, 0x4);
cam_i2c_write(axi_iic, device_addr, 0x5815, 0x5);
cam_i2c_write(axi_iic, device_addr, 0x5816, 0x9);
cam_i2c_write(axi_iic, device_addr, 0x5817, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5818, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x5819, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x581a, 0x4);
cam_i2c_write(axi_iic, device_addr, 0x581b, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x581c, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x581d, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x581e, 0x6);
cam_i2c_write(axi_iic, device_addr, 0x581f, 0x9);
cam_i2c_write(axi_iic, device_addr, 0x5820, 0x12);
cam_i2c_write(axi_iic, device_addr, 0x5821, 0xb);
cam_i2c_write(axi_iic, device_addr, 0x5822, 0x4);
cam_i2c_write(axi_iic, device_addr, 0x5823, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5824, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5825, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x5826, 0x6);
cam_i2c_write(axi_iic, device_addr, 0x5827, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x5828, 0x17);
cam_i2c_write(axi_iic, device_addr, 0x5829, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x582a, 0x9);
cam_i2c_write(axi_iic, device_addr, 0x582b, 0x6);
cam_i2c_write(axi_iic, device_addr, 0x582c, 0x5);
cam_i2c_write(axi_iic, device_addr, 0x582d, 0x6);
cam_i2c_write(axi_iic, device_addr, 0x582e, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x582f, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x5830, 0x28);
cam_i2c_write(axi_iic, device_addr, 0x5831, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x5832, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x5833, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x5834, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x5835, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x5836, 0x15);
cam_i2c_write(axi_iic, device_addr, 0x5837, 0x1d);
cam_i2c_write(axi_iic, device_addr, 0x5838, 0x6e);
cam_i2c_write(axi_iic, device_addr, 0x5839, 0x39);
cam_i2c_write(axi_iic, device_addr, 0x583a, 0x27);
cam_i2c_write(axi_iic, device_addr, 0x583b, 0x1f);
cam_i2c_write(axi_iic, device_addr, 0x583c, 0x1e);
cam_i2c_write(axi_iic, device_addr, 0x583d, 0x23);
cam_i2c_write(axi_iic, device_addr, 0x583e, 0x2f);
cam_i2c_write(axi_iic, device_addr, 0x583f, 0x41);
cam_i2c_write(axi_iic, device_addr, 0x5840, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x5841, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5842, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5843, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5844, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5845, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5846, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5847, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5848, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5849, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x584a, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x584b, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x584c, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x584d, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x584e, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x584f, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5850, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x5851, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x5852, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x5853, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x5854, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5855, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5856, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5857, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x5858, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x5859, 0xe);
cam_i2c_write(axi_iic, device_addr, 0x585a, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x585b, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x585c, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x585d, 0xa);
cam_i2c_write(axi_iic, device_addr, 0x585e, 0x9);
cam_i2c_write(axi_iic, device_addr, 0x585f, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5860, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x5861, 0xb);
cam_i2c_write(axi_iic, device_addr, 0x5862, 0xd);
cam_i2c_write(axi_iic, device_addr, 0x5863, 0x7);
cam_i2c_write(axi_iic, device_addr, 0x5864, 0x17);
cam_i2c_write(axi_iic, device_addr, 0x5865, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5866, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x5867, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x5868, 0x16);
cam_i2c_write(axi_iic, device_addr, 0x5869, 0x12);
cam_i2c_write(axi_iic, device_addr, 0x586a, 0x1b);
cam_i2c_write(axi_iic, device_addr, 0x586b, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x586c, 0x16);
cam_i2c_write(axi_iic, device_addr, 0x586d, 0x16);
cam_i2c_write(axi_iic, device_addr, 0x586e, 0x18);
cam_i2c_write(axi_iic, device_addr, 0x586f, 0x1f);
cam_i2c_write(axi_iic, device_addr, 0x5870, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x5871, 0x16);
cam_i2c_write(axi_iic, device_addr, 0x5872, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x5873, 0xf);
cam_i2c_write(axi_iic, device_addr, 0x5874, 0x13);
cam_i2c_write(axi_iic, device_addr, 0x5875, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x5876, 0x1e);
cam_i2c_write(axi_iic, device_addr, 0x5877, 0x17);
cam_i2c_write(axi_iic, device_addr, 0x5878, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x5879, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x587a, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x587b, 0x1e);
cam_i2c_write(axi_iic, device_addr, 0x587c, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x587d, 0x1c);
cam_i2c_write(axi_iic, device_addr, 0x587e, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x587f, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x5880, 0x1b);
cam_i2c_write(axi_iic, device_addr, 0x5881, 0x1f);
cam_i2c_write(axi_iic, device_addr, 0x5882, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5883, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x5884, 0x1d);
cam_i2c_write(axi_iic, device_addr, 0x5885, 0x1e);
cam_i2c_write(axi_iic, device_addr, 0x5886, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x5887, 0x1a);
cam_i2c_write(axi_iic, device_addr, 0x528a, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x528b, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x528c, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x528d, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x528e, 0x40);
cam_i2c_write(axi_iic, device_addr, 0x528f, 0x50);
cam_i2c_write(axi_iic, device_addr, 0x5290, 0x60);
cam_i2c_write(axi_iic, device_addr, 0x5292, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5293, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x5294, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5295, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x5296, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5297, 0x08);
cam_i2c_write(axi_iic, device_addr, 0x5298, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5299, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x529a, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529b, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x529c, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529d, 0x28);
cam_i2c_write(axi_iic, device_addr, 0x529e, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x529f, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x5282, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5680, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5681, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5682, 0x05);
cam_i2c_write(axi_iic, device_addr, 0x5683, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5684, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5685, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x5686, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x5687, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x5180, 0xff);
cam_i2c_write(axi_iic, device_addr, 0x5181, 0x52);
cam_i2c_write(axi_iic, device_addr, 0x5182, 0x11);
cam_i2c_write(axi_iic, device_addr, 0x5183, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5184, 0x25);
cam_i2c_write(axi_iic, device_addr, 0x5185, 0x24);
cam_i2c_write(axi_iic, device_addr, 0x5186, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5187, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5188, 0x14);
cam_i2c_write(axi_iic, device_addr, 0x5189, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x518a, 0x60);
cam_i2c_write(axi_iic, device_addr, 0x518b, 0xa2);
cam_i2c_write(axi_iic, device_addr, 0x518c, 0x9c);
cam_i2c_write(axi_iic, device_addr, 0x518d, 0x36);
cam_i2c_write(axi_iic, device_addr, 0x518e, 0x34);
cam_i2c_write(axi_iic, device_addr, 0x518f, 0x54);
cam_i2c_write(axi_iic, device_addr, 0x5190, 0x4c);
cam_i2c_write(axi_iic, device_addr, 0x5191, 0xf8);
cam_i2c_write(axi_iic, device_addr, 0x5192, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x5193, 0x70);
cam_i2c_write(axi_iic, device_addr, 0x5194, 0xf0);
cam_i2c_write(axi_iic, device_addr, 0x5195, 0xf0);
cam_i2c_write(axi_iic, device_addr, 0x5196, 0x03);
cam_i2c_write(axi_iic, device_addr, 0x5197, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x5198, 0x05);
cam_i2c_write(axi_iic, device_addr, 0x5199, 0x2f);
cam_i2c_write(axi_iic, device_addr, 0x519a, 0x04);
cam_i2c_write(axi_iic, device_addr, 0x519b, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x519c, 0x06);
cam_i2c_write(axi_iic, device_addr, 0x519d, 0xa0);
cam_i2c_write(axi_iic, device_addr, 0x519e, 0xa0);
cam_i2c_write(axi_iic, device_addr, 0x3a0f, 0x3c);
cam_i2c_write(axi_iic, device_addr, 0x3a10, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x3a1b, 0x3c);
cam_i2c_write(axi_iic, device_addr, 0x3a1e, 0x30);
cam_i2c_write(axi_iic, device_addr, 0x3a11, 0x70);
cam_i2c_write(axi_iic, device_addr, 0x3a1f, 0x10);
cam_i2c_write(axi_iic, device_addr, 0x3800, 0x1);
cam_i2c_write(axi_iic, device_addr, 0x3801, 0x50);
cam_i2c_write(axi_iic, device_addr, 0x3802, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x3803, 0x8);
cam_i2c_write(axi_iic, device_addr, 0x3804, 0x5);
cam_i2c_write(axi_iic, device_addr, 0x3805, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x3806, 0x3);
cam_i2c_write(axi_iic, device_addr, 0x3807, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x3808, 0x3);
cam_i2c_write(axi_iic, device_addr, 0x3809, 0x20);
cam_i2c_write(axi_iic, device_addr, 0x380a, 0x2);
cam_i2c_write(axi_iic, device_addr, 0x380b, 0x58);
cam_i2c_write(axi_iic, device_addr, 0x380c, 0xc);
cam_i2c_write(axi_iic, device_addr, 0x380d, 0x80);
cam_i2c_write(axi_iic, device_addr, 0x380e, 0x3);
cam_i2c_write(axi_iic, device_addr, 0x380f, 0xe8);
cam_i2c_write(axi_iic, device_addr, 0x5001, 0x7f);
cam_i2c_write(axi_iic, device_addr, 0x5680, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5681, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5682, 0x5);
cam_i2c_write(axi_iic, device_addr, 0x5683, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5684, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5685, 0x0);
cam_i2c_write(axi_iic, device_addr, 0x5686, 0x3);
cam_i2c_write(axi_iic, device_addr, 0x5687, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x5687, 0xc0);
cam_i2c_write(axi_iic, device_addr, 0x3815, 0x02);
cam_i2c_write(axi_iic, device_addr, 0x3503, 0x00);
cam_i2c_write(axi_iic, device_addr, 0x3818, 0x81); // No Mirror
cam_i2c_write(axi_iic, device_addr, 0x3621, 0xa7);
cam_i2c_write(axi_iic, device_addr, 0x4740, 0x21);
cam_i2c_write(axi_iic, device_addr, 0x501e, 0x2a);
cam_i2c_write(axi_iic, device_addr, 0x5002, 0x78);
cam_i2c_write(axi_iic, device_addr, 0x501f, 0x01);
cam_i2c_write(axi_iic, device_addr, 0x4300, 0x61);
return(0);
}
/dts-v1/;
/ {
fragment@0 {
target-path = "/fpga-full";
__overlay__ {
firmware-name = "cam_dp_wrapper.bin";
};
};
};
/dts-v1/;/plugin/;
/ {
fragment@0 {
target-path = "/amba";
__overlay__ {
fclk0 {
compatible = "ikwzm,fclkcfg-0.10.a";
clocks = <&clk 0x47>;
insert-rate = "200000000";
insert-enable = <1>;
remove-rate = "1000000";
remove-enable = <0>;
};
fclk1 {
compatible = "ikwzm,fclkcfg-0.10.a";
clocks = <&clk 0x48>;
insert-rate = "24000000";
insert-enable = <1>;
remove-rate = "1000000";
remove-enable = <0>;
};
};
};
};
/dts-v1/;/plugin/;
/ {
fragment@0 {
target-path = "/amba_pl@0";
#address-cells = <2>;
#size-cells = <2>;
__overlay__ {
#address-cells = <2>;
#size-cells = <2>;
mt9d111_inf_axis-uio {
compatible = "generic-uio";
reg = <0x0 0xA0010000 0x0 0x1000>;
};
axi_iic-uio {
compatible = "generic-uio";
reg = <0x0 0xA0011000 0x0 0x1000>;
};
disp_dmar_axis-uio {
compatible = "generic-uio";
reg = <0x0 0xA0000000 0x0 0x10000>;
};
vflip_dma_write-uio {
compatible = "generic-uio";
reg = <0x0 0xA0020000 0x0 0x10000>;
};
axi_gpio_0-uio {
compatible = "generic-uio";
reg = <0x0 0xA0012000 0x0 0x1000>;
};
axi_gpio_1-uio {
compatible = "generic-uio";
reg = <0x0 0xA0013000 0x0 0x1000>;
};
cam_dp-udmabuf4 {
compatible = "ikwzm,udmabuf-0.10.a";
device-name = "udmabuf4";
size = <0x02000000>;
};
};
};
};
#!/bin/bash
dtc -I dts -O dtb -o fpga-load.dtb fpga-load.dts
dtc -I dts -O dtb -o fclk01-zynqmp.dtb fclk01-zynqmp.dts
dtc -I dts -O dtb -o cam_dp.dtb cam_dp.dts
#!/bin/bash
sudo mkdir /config/device-tree/overlays/fpga
sudo cp fpga-load.dtb /config/device-tree/overlays/fpga/dtbo
sudo mkdir /config/device-tree/overlays/fclk01-zynqmp
sudo cp fclk01-zynqmp.dtb /config/device-tree/overlays/fclk01-zynqmp/dtbo
sudo mkdir /config/device-tree/overlays/cam_dp
sudo cp cam_dp.dtb /config/device-tree/overlays/cam_dp/dtbo
sleep 0.5
sudo chmod 666 /dev/uio*
sudo chmod 666 /dev/udmabuf4
#!/bin/bash
sudo rmdir /config/device-tree/overlays/cam_dp/
sudo rmdir /config/device-tree/overlays/fclk01-zynqmp
sudo rmdir /config/device-tree/overlays/fpga/
をassign disp_pixel = {pfifo_dout[23:16], 4'd0, pfifo_dout[15:8], 4'd0, pfifo_dout[7:0], 4'd0};
に修正する。assign disp_pixel = {pfifo_dout[7:0], 4'd0, pfifo_dout[23:16], 4'd0, pfifo_dout[15:8], 4'd0}; //BRG
all:
{
[destination_device = pl] cam_dp_wrapper.bit
}
#!/bin/bash
bootgen -image cam_dp_wrapper.bif -arch zynqmp -w -o cam_dp_wrapper.bin
set_property PACKAGE_PIN G7 [get_ports pclk]
set_property PACKAGE_PIN G6 [get_ports {cam_data[7]}]
set_property PACKAGE_PIN D5 [get_ports {cam_data[6]}]
set_property PACKAGE_PIN E6 [get_ports {cam_data[5]}]
set_property PACKAGE_PIN C7 [get_ports {cam_data[4]}]
set_property PACKAGE_PIN E5 [get_ports {cam_data[3]}]
set_property PACKAGE_PIN B6 [get_ports {cam_data[2]}]
set_property PACKAGE_PIN D6 [get_ports {cam_data[1]}]
set_property PACKAGE_PIN C5 [get_ports {cam_data[0]}]
set_property PACKAGE_PIN G5 [get_ports href]
set_property PACKAGE_PIN F7 [get_ports standby]
set_property PACKAGE_PIN F8 [get_ports vsync]
set_property PACKAGE_PIN A7 [get_ports xck]
set_property PACKAGE_PIN D7 [get_ports iic_rtl_scl_io]
set_property PACKAGE_PIN F6 [get_ports iic_rtl_sda_io]
create_clock -period 13.700 -name pclk -waveform {0.000 6.850} [get_ports pclk]
set_input_delay -clock [get_clocks pclk] 10.800 [get_ports {{cam_data[0]} {cam_data[1]} {cam_data[2]} {cam_data[3]} {cam_data[4]} {cam_data[5]} {cam_data[6]} {cam_data[7]} href pclk vsync}]
set_property CLOCK_DEDICATED_ROUTE FALSE [get_nets pclk_IBUF_inst/O]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[7]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[6]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[5]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[4]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[3]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[2]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[1]}]
set_property IOSTANDARD LVCMOS18 [get_ports {cam_data[0]}]
set_property IOSTANDARD LVCMOS18 [get_ports pclk]
set_property IOSTANDARD LVCMOS18 [get_ports href]
set_property IOSTANDARD LVCMOS18 [get_ports standby]
set_property IOSTANDARD LVCMOS18 [get_ports vsync]
set_property IOSTANDARD LVCMOS18 [get_ports xck]
set_property IOSTANDARD LVCMOS18 [get_ports iic_rtl_scl_io]
set_property IOSTANDARD LVCMOS18 [get_ports iic_rtl_sda_io]
set_false_path -from [get_clocks clk_pl_0] -to [get_clocks pclk]
set_false_path -from [get_clocks pclk] -to [get_clocks clk_pl_0]
create_clock -period 6.734 -name disp_clk -waveform {0.000 3.367} [get_pins cam_dp_i/zynq_ultra_ps_e_0/dp_video_ref_clk]
set_false_path -from [get_clocks clk_pl_0] -to [get_clocks disp_clk]
set_false_path -from [get_clocks disp_clk] -to [get_clocks clk_pl_0]
// disp_dmar_axis.h
// 2019/02/01 by marsee
//
#ifndef __DISP_DMAR_AXIS_H__
#define __DISP_DMAR_AXIS_H__
#include "ap_axi_sdata.h"
#include "hls_video.h"
#define SVGA_WIDTH 800
#define SVGA_HEIGHT 600
#define XGA_WIDTH 1024
#define XGA_HEIGHT 768
#define HD_WIDTH 1920
#define HD_HEIGHT 1080
#define RESO_SVGA 0
#define RESO_XGA 1
#define RESO_HD 2
typedef hls::stream<ap_axiu<32,1,1,1> > AXI_STREAM;
typedef ap_axiu<32,1,1,1> AP_AXIU32;
typedef hls::Scalar<3, unsigned char> RGB_PIXEL;
typedef hls::Mat<HD_HEIGHT, HD_WIDTH, HLS_8UC3> RGB_IMAGE;
typedef hls::Mat<HD_HEIGHT, HD_WIDTH, HLS_8UC1> GRAY_IMAGE;
#endif
// disp_dmar_axis.cpp
// 2019/02/01 by marsee
//
#include <ap_int.h>
#include <hls_stream.h>
#include <ap_axi_sdata.h>
#include "disp_dmar_axis.h"
int disp_dmar_fb0(volatile ap_int<32> *fb0, AXI_STREAM &outs, int max_width, int max_height);
int disp_dmar_fb1(volatile ap_int<32> *fb1, AXI_STREAM &outs, int max_width, int max_height);
int disp_dmar_fb2(volatile ap_int<32> *fb2, AXI_STREAM &outs, int max_width, int max_height);
int disp_dmar_axis(volatile ap_int<32> *fb0, volatile ap_int<32> *fb1, volatile ap_int<32> *fb2,
AXI_STREAM &outs, ap_uint<32> &resolution, ap_uint<2> &active_frame){
#pragma HLS INTERFACE ap_ctrl_hs port=return
#pragma HLS INTERFACE ap_none port=active_frame
#pragma HLS INTERFACE s_axilite port=resolution
#pragma HLS INTERFACE axis register both port=outs
#pragma HLS INTERFACE m_axi depth=480000 port=fb2 offset=slave
#pragma HLS INTERFACE m_axi depth=480000 port=fb1 offset=slave
#pragma HLS INTERFACE m_axi depth=480000 port=fb0 offset=slave
AP_AXIU32 pix;
int max_width, max_height;
switch((int)resolution){
case 0: // SVGA
max_width = SVGA_WIDTH;
max_height = SVGA_HEIGHT;
break;
case 1: // XGA
max_width = XGA_WIDTH;
max_height = XGA_HEIGHT;
break;
default: // HD
max_width = HD_WIDTH;
max_height = HD_HEIGHT;
break;
}
if (active_frame == (ap_uint<2>)0)
disp_dmar_fb2(fb2, outs, max_width, max_height);
else if (active_frame == (ap_uint<2>)1)
disp_dmar_fb0(fb0, outs, max_width, max_height);
else
disp_dmar_fb1(fb1, outs, max_width, max_height);
return(0);
}
int disp_dmar_fb0(volatile ap_int<32> *fb0, AXI_STREAM &outs, int max_width, int max_height){
AP_AXIU32 pix;
LOOP_Y0: for (int y=0; y<max_height; y++){
#pragma HLS LOOP_TRIPCOUNT min=600 max=1080 avg=768
LOOP_X0: for (int x=0; x<max_width; x++){
#pragma HLS LOOP_TRIPCOUNT min=800 max=1920 avg=1024
#pragma HLS PIPELINE II=1
pix.data = fb0[(y*max_width)+x];
if (x==0 && y==0)
pix.user = 1;
else
pix.user = 0;
if (x == max_width-1)
pix.last = 1;
else
pix.last = 0;
outs << pix;
}
}
return(0);
}
int disp_dmar_fb1(volatile ap_int<32> *fb1, AXI_STREAM &outs, int max_width, int max_height){
AP_AXIU32 pix;
LOOP_Y1: for (int y=0; y<max_height; y++){
#pragma HLS LOOP_TRIPCOUNT min=600 max=1080 avg=768
LOOP_X1: for (int x=0; x<max_width; x++){
#pragma HLS LOOP_TRIPCOUNT min=800 max=1920 avg=1024
#pragma HLS PIPELINE II=1
pix.data = fb1[(y*max_width)+x];
if (x==0 && y==0)
pix.user = 1;
else
pix.user = 0;
if (x == max_width-1)
pix.last = 1;
else
pix.last = 0;
outs << pix;
}
}
return(0);
}
int disp_dmar_fb2(volatile ap_int<32> *fb2, AXI_STREAM &outs, int max_width, int max_height){
AP_AXIU32 pix;
LOOP_Y2: for (int y=0; y<max_height; y++){
#pragma HLS LOOP_TRIPCOUNT min=600 max=1080 avg=768
LOOP_X2: for (int x=0; x<max_width; x++){
#pragma HLS LOOP_TRIPCOUNT min=800 max=1920 avg=1024
#pragma HLS PIPELINE II=1
pix.data = fb2[(y*max_width)+x];
if (x==0 && y==0)
pix.user = 1;
else
pix.user = 0;
if (x == max_width-1)
pix.last = 1;
else
pix.last = 0;
outs << pix;
}
}
return(0);
}
// disp_dmar_axis_tb.cpp
// 2019/02/01 by marsee
//
#include <ap_int.h>
#include <hls_stream.h>
#include <iostream>
#include <fstream>
#include "hls_opencv.h"
#include "disp_dmar_axis.h"
int disp_dmar_axis(volatile ap_int<32> *fb0, volatile ap_int<32> *fb1, volatile ap_int<32> *fb2,
AXI_STREAM &outs, ap_uint<32> &resolution, ap_uint<2> &active_frame);
#define NUM_FRAME_BUFFER 3
#define FB_NUMBER 0
int main(){
using namespace cv;
AXI_STREAM outs;
AP_AXIU32 pix;
ap_uint<32> *frame_buffer;
// OpenCV で 画像を読み込む
Mat src = imread("bmp_file0.bmp");
// 画像をセーブするメモリをアロケート
if ((frame_buffer =(ap_uint<32> *)malloc(NUM_FRAME_BUFFER * sizeof(ap_int<32>) * (src.cols * src.rows))) == NULL){
fprintf(stderr, "Can't allocate frame_buffer0 ~ 2\n");
exit(1);
}
// Mat フォーマットからフレームバッファに画像をロード
Mat_<Vec3b> src_vec3b = Mat_<Vec3b>(src);
for(int i=0; i<NUM_FRAME_BUFFER; i++){
for(int y=0; y<src.rows; y++){
for(int x=0; x<src.cols; x++){
Vec3b pixel;
pixel = src_vec3b(y,x);
int rgb = ((int)(pixel(2))<<16) + ((int)(pixel(1))<<8) + pixel(0);
frame_buffer[(src.cols*src.rows)*i + (src.cols*y) + x] = rgb;
}
}
}
ap_uint<2> fb_number = FB_NUMBER;
ap_uint<32> resolution = RESO_SVGA;
disp_dmar_axis((volatile ap_int<32> *)frame_buffer,
(volatile ap_int<32> *)&frame_buffer[src.cols * src.rows],
(volatile ap_int<32> *)&frame_buffer[2 * (src.cols * src.rows)],
outs, resolution, fb_number);
// AXI4 Stream から Mat フォーマットに変換しファイルに書き込み
Mat dst(src.rows, src.cols, CV_8UC3);
Mat_<Vec3b> dst_vec3b = Mat_<Vec3b>(dst);
for(int y=0; y<dst.rows; y++){
for(int x=0; x<dst.cols; x++){
outs >> pix;
int rgb = pix.data;
//int rgb = frame_buffer[(src.cols*y) + x];
Vec3b pixel;
pixel[0] = (rgb & 0xff); // blue
pixel[1] = (rgb & 0xff00) >> 8; // green
pixel[2] = (rgb & 0xff0000) >> 16; // red
dst_vec3b(y,x) = pixel;
}
}
imwrite("dmar_result.bmp", dst);
free(frame_buffer);
return(0);
}
| 日 | 月 | 火 | 水 | 木 | 金 | 土 |
|---|---|---|---|---|---|---|
| - | - | - | - | - | 1 | 2 |
| 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| 10 | 11 | 12 | 13 | 14 | 15 | 16 |
| 17 | 18 | 19 | 20 | 21 | 22 | 23 |
| 24 | 25 | 26 | 27 | 28 | - | - |