124075 - rpi5 aidge – dAIEdge-VLab Status

dAIEdge-VLab Weekly Tests

rpi5

0000019c71638acf37948e8933e2246c7ee24eb06307d06986d991b6e014d767

aidge

124075 - rpi5 aidge

✅

Summary: Pipeline failed, but error.log is filled.

Model Details

model name : bvlcalexnet-12-int8.onnx
model url : Download here

Logs Details

user.log

[[33mWARNING[0m] - Aidge currently only supports layerwise scaling and not channelwise for
[[33mWARNING[0m]   QLinearConv[Weight DequantizeLinear] node. This node will be filled by a
[[33mWARNING[0m]   GenericOperator.
[[94mNOTICE[0m] - Loaded node [[1m[3mConv_nc_rename_0_quant[0m] of type [[1m[3mQLinearConv[0m] as a GenericOperator.
[[94mNOTICE[0m]   	* auto_pad : b'NOTSET'
[[94mNOTICE[0m]   	* group : 1
[[94mNOTICE[0m]   	* kernel_shape : [11, 11]
[[94mNOTICE[0m]   	* pads : [0, 0, 0, 0]
[[94mNOTICE[0m]   	* strides : [4, 4]
[[33mWARNING[0m] - Trying to load node named [[1m[3mConv_nc_rename_4_quant[0m] of type [[1m[3mQLinearConv[0m].
[[33mWARNING[0m]   Loading node using a [[1m[3mGenericOperator[0m].
[[33mWARNING[0m]   Please report this issue at https://gitlab.eclipse.org/eclipse/aidge/aidge_onnx by
[[33mWARNING[0m]   providing your ONNX model and the following error:
[[33mWARNING[0m]   "ONNX_NODE_CONVERTER_ returned: list index out of range"
[[94mNOTICE[0m] - Loaded node [[1m[3mConv_nc_rename_4_quant[0m] of type [[1m[3mQLinearConv[0m] as a GenericOperator.
[[94mNOTICE[0m]   	* auto_pad : b'NOTSET'
[[94mNOTICE[0m]   	* group : 2
[[94mNOTICE[0m]   	* kernel_shape : [5, 5]
[[94mNOTICE[0m]   	* pads : [2, 2, 2, 2]
[[94mNOTICE[0m]   	* strides : [1, 1]
[[33mWARNING[0m] - Aidge currently only supports layerwise scaling and not channelwise for
[[33mWARNING[0m]   QLinearConv[Weight DequantizeLinear] node. This node will be filled by a
[[33mWARNING[0m]   GenericOperator.
[[94mNOTICE[0m] - Loaded node [[1m[3mConv_nc_rename_8_quant[0m] of type [[1m[3mQLinearConv[0m] as a GenericOperator.
[[94mNOTICE[0m]   	* auto_pad : b'NOTSET'
[[94mNOTICE[0m]   	* group : 1
[[94mNOTICE[0m]   	* kernel_shape : [3, 3]
[[94mNOTICE[0m]   	* pads : [1, 1, 1, 1]
[[94mNOTICE[0m]   	* strides : [1, 1]
[[33mWARNING[0m] - Trying to load node named [[1m[3mConv_nc_rename_10_quant[0m] of type [[1m[3mQLinearConv[0m].
[[33mWARNING[0m]   Loading node using a [[1m[3mGenericOperator[0m].
[[33mWARNING[0m]   Please report this issue at https://gitlab.eclipse.org/eclipse/aidge/aidge_onnx by
[[33mWARNING[0m]   providing your ONNX model and the following error:
[[33mWARNING[0m]   "ONNX_NODE_CONVERTER_ returned: list index out of range"
[[94mNOTICE[0m] - Loaded node [[1m[3mConv_nc_rename_10_quant[0m] of type [[1m[3mQLinearConv[0m] as a GenericOperator.
[[94mNOTICE[0m]   	* auto_pad : b'NOTSET'
[[94mNOTICE[0m]   	* group : 2
[[94mNOTICE[0m]   	* kernel_shape : [3, 3]
[[94mNOTICE[0m]   	* pads : [1, 1, 1, 1]
[[94mNOTICE[0m]   	* strides : [1, 1]
[[33mWARNING[0m] - Trying to load node named [[1m[3mConv_nc_rename_12_quant[0m] of type [[1m[3mQLinearConv[0m].
[[33mWARNING[0m]   Loading node using a [[1m[3mGenericOperator[0m].
[[33mWARNING[0m]   Please report this issue at https://gitlab.eclipse.org/eclipse/aidge/aidge_onnx by
[[33mWARNING[0m]   providing your ONNX model and the following error:
[[33mWARNING[0m]   "ONNX_NODE_CONVERTER_ returned: list index out of range"
[[94mNOTICE[0m] - Loaded node [[1m[3mConv_nc_rename_12_quant[0m] of type [[1m[3mQLinearConv[0m] as a GenericOperator.
[[94mNOTICE[0m]   	* auto_pad : b'NOTSET'
[[94mNOTICE[0m]   	* group : 2
[[94mNOTICE[0m]   	* kernel_shape : [3, 3]
[[94mNOTICE[0m]   	* pads : [1, 1, 1, 1]
[[94mNOTICE[0m]   	* strides : [1, 1]
[[94mNOTICE[0m] - Node name "MatMul_nc_rename_16_quant" is a duplicate, renaming it to
[[94mNOTICE[0m]   MatMul_nc_rename_16_quant_1.
[[94mNOTICE[0m] - Loaded node [[1m[3mAdd_nc_rename_17_quant[0m] of type [[1m[3mQLinearAdd[0m] as a GenericOperator.
[[94mNOTICE[0m] - Node name "MatMul_nc_rename_19_quant" is a duplicate, renaming it to
[[94mNOTICE[0m]   MatMul_nc_rename_19_quant_1.
[[94mNOTICE[0m] - Loaded node [[1m[3mAdd_nc_rename_20_quant[0m] of type [[1m[3mQLinearAdd[0m] as a GenericOperator.
[[94mNOTICE[0m] - Node name "MatMul_nc_rename_22_quant" is a duplicate, renaming it to
[[94mNOTICE[0m]   MatMul_nc_rename_22MODEL : bvlcalexnet12int8.onnx

===============
ONNX Graph
===============

data_0 [1, 3, 224, 224]

===============
Aidge Graph
===============

  Node(name='fc8_1_MatMul_scale', optype='Producer', children: [[1]])
  Node(name='fc6_1_MatMul_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_1_scale', optype='Producer', children: [[1]])
  Node(name='fc7_1_MatMul_zero_point', optype='Producer', children: [[1]])
  Node(name='fc8_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='fc6_b_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_1_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_b_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_1_MatMul_scale', optype='Producer', children: [[1]])
  Node(name='conv4_1_zero_point', optype='Producer', children: [[1, 1]])
  Node(name='conv4_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='fc7_b_0_scale', optype='Producer', children: [[1]])
  Node(name='pool2_1_zero_point', optype='Producer', children: [[1]])
  Node(name='conv3_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv4_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='conv3_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='fc6_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv3_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv2_1_scale', optype='Producer', children: [[1]])
  Node(name='conv4_w_0_scale', optype='Producer', children: [[1]])
  Node(name='conv5_1_zero_point', optype='Producer', children: [[1]])
  Node(name='conv5_w_0_scale', optype='Producer', children: [[1]])
  Node(name='conv5_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv4_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv2_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv4_1_scale', optype='Producer', children: [[1, 1]])
  Node(name='fc6_1_MatMul_scale', optype='Producer', children: [[1]])
  Node(name='fc6_1_scale', optype='Producer', children: [[1]])
  Node(name='fc6_1_zero_point', optype='Producer', children: [[1]])
  Node(name='conv1_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='LRN_nc_rename_2', optype='LRN', parents: [1], children: [[1]])
  Node(name='LRN_nc_rename_6', optype='LRN', parents: [1], children: [[1]])
  Node(name='conv1_w_0_scale', optype='Producer', children: [[1]])
  Node(name='fc8_1_DequantizeLinear', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='conv1_1_scale', optype='Producer', children: [[1]])
  Node(name='conv3_w_0_scale', optype='Producer', children: [[1]])
  Node(name='conv3_1_scale', optype='Producer', children: [[1, 1]])
  Node(name='MatMul_nc_rename_22_quant', optype='QLinearMatMul', parents: [1, 1], children: [[1]])
  Node(name='MatMul_nc_rename_19_quant', optype='QLinearMatMul', parents: [1, 1], children: [[1]])
  Node(name='fc6_b_0_scale', optype='Producer', children: [[1]])
  Node(name='Add_nc_rename_17_quant', optype='QLinearAdd', parents: [1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='MatMul_nc_rename_16_quant', optype='QLinearMatMul', parents: [1, 1], children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear', optype='Quantizer', parents: [1], children: [[1]])
  Node(name='Reshape_nc_rename_15', optype='Reshape', parents: [1, 1], children: [[1]])
  Node(name='OC2_DUMMY_1', optype='Producer', children: [[1]])
  Node(name='pool5_1_DequantizeLinear', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='conv2_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_12_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='fc6_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_8_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='conv2_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_10_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='pool2_1_QuantizeLinear', optype='Quantizer', parents: [1], children: [[1]])
  Node(name='conv2_1_zero_point', optype='Producer', children: [[1]])
  Node(name='MaxPool_nc_rename_14_quant', optype='PaddedMaxPooling2D', parents: [1, 0, 0, 0], children: [[1], []])
  Node(name='conv2_w_0_scale', optype='Producer', children: [[1]])
  Node(name='conv2_2_DequantizeLinear', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='MaxPool_nc_rename_7', optype='MaxPooling2D', parents: [1], children: [[1], []])
  Node(name='pool1_1_QuantizeLinear', optype='Quantizer', parents: [1], children: [[1]])
  Node(name='pool2_1_scale', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_4_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='data_0_zero_point', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_0_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='conv1_2_DequantizeLinear', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='conv1_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='MaxPool_nc_rename_3', optype='MaxPooling2D', parents: [1], children: [[1], []])
  Node(name='conv5_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='data_0_QuantizeLinear', optype='Quantizer', parents: [0], children: [[1]])
  Node(name='conv5_1_scale', optype='Producer', children: [[1]])
  Node(name='fc8_b_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc8_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv3_1_zero_point', optype='Producer', children: [[1, 1]])
  Node(name='fc8_b_0_scale', optype='Producer', children: [[1]])
  Node(name='fc8_1_scale', optype='Producer', children: [[1]])
  Node(name='fc8_1_zero_point', optype='Producer', children: [[1]])
  Node(name='fc8_1_MatMul_zero_point', optype='Producer', children: [[1]])
  Node(name='conv5_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='data_0_scale', optype='Producer', children: [[1]])
  Node(name='Softmax_nc_rename_24', optype='Softmax', parents: [1], children: [[]])
  Node(name='pool1_1_zero_point', optype='Producer', children: [[1]])
  Node(name='pool1_1_scale', optype='Producer', children: [[1]])
  Node(name='conv1_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv1_1_zero_point', optype='Producer', children: [[1]])
  Node(name='Add_nc_rename_23_quant', optype='QLinearAdd', parents: [1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='Add_nc_rename_20_quant', optype='QLinearAdd', parents: [1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])

===============
Supported nodes
===============

Native operators: 79 (9 types)
- Dequantizer: 4
- LRN: 2
- MaxPooling2D: 2
- PaddedMaxPooling2D: 1
- Producer: 61
- QLinearMatMul: 3
- Quantizer: 4
- Reshape: 1
- Softmax: 1
Generic operators: 8 (2 types)
- QLinearAdd: 3
- QLinearConv: 5
Native types coverage: 81.8% (9/11)
Native operators coverage: 90.8% (79/87)
  (defaultdict(<class 'int'>, {'Producer': 61, 'LRN': 2, 'Softmax': 1, 'Dequantizer': 4, 'QLinearMatMul': 3, 'Quantizer': 4, 'Reshape': 1, 'PaddedMaxPooling2D': 1, 'MaxPooling2D': 2}), defaultdict(<class 'int'>, {'QLinearAdd': 3, 'QLinearConv': 5}))

===============\Graph manipulation
===============

Remove flatten
Fuse batchnorm
Expand metaop
Fuse to metaop

===============
New Aidge Graph
===============

  Node(name='fc7_b_0_scale', optype='Producer', children: [[1]])
  Node(name='fc6_1_scale', optype='Producer', children: [[1]])
  Node(name='fc6_b_0_scale', optype='Producer', children: [[1]])
  Node(name='fc6_1_zero_point', optype='Producer', children: [[1]])
  Node(name='ScalingFactor_2', optype='Producer', children: [[1]])
  Node(name='', optype='PadMaxPool', parents: [1, 0, 0, 0], children: [[1], []])
  Node(name='fc6_1_MatMul_scale', optype='Producer', children: [[1]])
  Node(name='conv2_2_DequantizeLinear_MulDeQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='ScalingFactor', optype='Producer', children: [[1]])
  Node(name='fc6_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv5_1_zero_point', optype='Producer', children: [[1]])
  Node(name='fc6_1_MatMul_zero_point', optype='Producer', children: [[1]])
  Node(name='conv5_w_0_scale', optype='Producer', children: [[1]])
  Node(name='conv4_1_scale', optype='Producer', children: [[1, 1]])
  Node(name='conv5_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv4_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv4_1_zero_point', optype='Producer', children: [[1, 1]])
  Node(name='conv4_w_0_scale', optype='Producer', children: [[1]])
  Node(name='conv4_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='conv3_1_scale', optype='Producer', children: [[1, 1]])
  Node(name='conv4_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv1_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='LRN_nc_rename_2', optype='LRN', parents: [1], children: [[1]])
  Node(name='data_0_scale', optype='Producer', children: [[1]])
  Node(name='LRN_nc_rename_6', optype='LRN', parents: [1], children: [[1]])
  Node(name='conv1_w_0_scale', optype='Producer', children: [[1]])
  Node(name='Softmax_nc_rename_24', optype='Softmax', parents: [1], children: [[]])
  Node(name='', optype='QMatMul', parents: [1, 1], children: [[1]])
  Node(name='fc8_1_DequantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='conv1_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv3_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='fc8_1_DequantizeLinear_MulDeQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='fc8_1_DequantizeLinear_SubDeQuant', optype='Sub', parents: [1, 1], children: [[1]])
  Node(name='conv1_1_zero_point', optype='Producer', children: [[1]])
  Node(name='MatMul_nc_rename_22_quant_w_dq', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='MatMul_nc_rename_19_quant_w_dq', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='', optype='QMatMul', parents: [1, 1], children: [[1]])
  Node(name='Add_nc_rename_17_quant', optype='QLinearAdd', parents: [1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='conv3_1_zero_point', optype='Producer', children: [[1, 1]])
  Node(name='conv3_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='MatMul_nc_rename_16_quant_x_dq', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='ScalingFactor_4', optype='Producer', children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_ClipQuantMax', optype='Producer', children: [[1]])
  Node(name='ZeroPoint_4', optype='Producer', children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_ClipQuant', optype='Clip', parents: [1, 1, 1], children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_AddQuant', optype='Add', parents: [1, 1], children: [[1]])
  Node(name='Add_nc_rename_23_quant', optype='QLinearAdd', parents: [1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_MulQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='pool5_1_DequantizeLinear_MulDeQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='MatMul_nc_rename_22_quant_x_dq', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='ScalingFactor_3', optype='Producer', children: [[1]])
  Node(name='ZeroPoint_3', optype='Producer', children: [[1]])
  Node(name='MatMul_nc_rename_19_quant_x_dq', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='Conv_nc_rename_10_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='pool2_1_QuantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='MatMul_nc_rename_16_quant_w_dq', optype='Dequantizer', parents: [1], children: [[1]])
  Node(name='conv5_1_scale', optype='Producer', children: [[1]])
  Node(name='pool2_1_QuantizeLinear_ClipQuantMax', optype='Producer', children: [[1]])
  Node(name='pool2_1_QuantizeLinear_ClipQuant', optype='Clip', parents: [1, 1, 1], children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_CastOut', optype='Cast', parents: [1], children: [[1]])
  Node(name='pool2_1_QuantizeLinear_AddQuant', optype='Add', parents: [1, 1], children: [[1]])
  Node(name='pool2_1_QuantizeLinear_MulQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_RoundQuant', optype='Round', parents: [1], children: [[1]])
  Node(name='fc6_b_0_zero_point', optype='Producer', children: [[1]])
  Node(name='conv1_1_scale', optype='Producer', children: [[1]])
  Node(name='pool1_1_zero_point', optype='Producer', children: [[1]])
  Node(name='conv3_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='fc7_b_0_zero_point', optype='Producer', children: [[1]])
  Node(name='', optype='QMatMul', parents: [1, 1], children: [[1]])
  Node(name='fc7_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='OC2_DUMMY_0_QuantizeLinear_ClipQuantMin', optype='Producer', children: [[1]])
  Node(name='pool2_1_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='ZeroPoint', optype='Producer', children: [[1]])
  Node(name='fc7_1_zero_point', optype='Producer', children: [[1]])
  Node(name='ZeroPoint_2', optype='Producer', children: [[1]])
  Node(name='Reshape_nc_rename_15', optype='Reshape', parents: [1, 1], children: [[1]])
  Node(name='conv2_2_DequantizeLinear_SubDeQuant', optype='Sub', parents: [1, 1], children: [[1]])
  Node(name='MaxPool_nc_rename_7', optype='MaxPooling2D', parents: [1], children: [[1], []])
  Node(name='pool5_1_DequantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='pool1_1_QuantizeLinear_CastOut', optype='Cast', parents: [1], children: [[1]])
  Node(name='pool1_1_QuantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='Conv_nc_rename_12_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='conv2_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv5_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv5_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='pool1_1_QuantizeLinear_RoundQuant', optype='Round', parents: [1], children: [[1]])
  Node(name='pool1_1_QuantizeLinear_ClipQuantMax', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_8_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='conv2_1_scale', optype='Producer', children: [[1]])
  Node(name='pool1_1_QuantizeLinear_ClipQuantMin', optype='Producer', children: [[1]])
  Node(name='pool1_1_QuantizeLinear_ClipQuant', optype='Clip', parents: [1, 1, 1], children: [[1]])
  Node(name='conv2_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='ZeroPoint', optype='Producer', children: [[1]])
  Node(name='conv2_1_zero_point', optype='Producer', children: [[1]])
  Node(name='pool1_1_QuantizeLinear_AddQuant', optype='Add', parents: [1, 1], children: [[1]])
  Node(name='pool2_1_QuantizeLinear_CastOut', optype='Cast', parents: [1], children: [[1]])
  Node(name='pool2_1_scale', optype='Producer', children: [[1]])
  Node(name='ScalingFactor', optype='Producer', children: [[1]])
  Node(name='pool1_1_QuantizeLinear_MulQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='pool2_1_QuantizeLinear_RoundQuant', optype='Round', parents: [1], children: [[1]])
  Node(name='conv2_w_0_zero_point', optype='Producer', children: [[1]])
  Node(name='OC2_DUMMY_1', optype='Producer', children: [[1]])
  Node(name='Conv_nc_rename_4_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  Node(name='Conv_nc_rename_0_quant', optype='QLinearConv', parents: [1, 1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])
  _quant_1.
[[94mNOTICE[0m] - Loaded node [[1m[3mAdd_nc_rename_23_quant[0m] of type [[1m[3mQLinearAdd[0m] as a GenericOperator.
[[94mNOTICE[0m] - Node name "ScalingFactor" is a duplicate, renaming it to ScalingFactor_1.
[[94mNOTICE[0m] - Node name "ZeroPoint" is a duplicate, renaming it to ZeroPoint_1.
[[94mNOTICE[0m] - Node name "ZeroPoint" is a duplicate, renaming it to ZeroPoint_2.
[[94mNOTICE[0m] - Node name "ScalingFactor" is a duplicate, renaming it to ScalingFactor_2.
[[94mNOTICE[0m] - Node name "ZeroPoint" is a duplicate, renaming it to ZeroPoint_3.
[[94mNOTICE[0m] - Node name "ScalingFactor" is a duplicate, renaming it to ScalingFactor_3.
[[94mNOTICE[0m] - Node name "ZeroPoint" is a duplicate, renaming it to ZeroPoint_4.
[[94mNOTICE[0m] - Node name "ScalingFactor" is a duplicate, renaming it to ScalingFactor_4.
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[33mWARNING[0m] - GenericOperator::setBackend(): cannot set backend for a generic operator, as no
[[33mWARNING[0m]   implementation has been provided!
[[31mERROR[0m] - Unable to forward data type for node Conv_nc_rename_0_quant (of type QLinearConv)
[[31mERROR[0m] - Unable to forward data type for node Conv_nc_rename_0_quant (of type QLinearConv)
[[31mERROR[0m] - Unable to forward data type for node Conv_nc_rename_0_quant (of type QLinearConv)
[[33mWARNING[0m] - Unable to forward data type (circular dependency and/or wrong dimensions and/or data
[[33mWARNING[0m]   dependent dimension?). Unable to compute output data type for nodes [" (QMatMul)", "
[[33mWARNING[0m]   (PadMaxPool)", "conv1_2_DequantizeLinear_SubDeQuant (Sub)",
[[33mWARNING[0m]   "conv1_2_DequantizeLinear_MulDeQuant (Mul)", "Conv_nc_rename_0_quant (QLinearConv)",
[[33mWARNING[0m]   "Conv_nc_rename_4_quant (QLinearConv)", "pool1_1_QuantizeLinear_MulQuant (Mul)",
[[33mWARNING[0m]   "pool1_1_QuantizeLinear_AddQuant (Add)", "pool1_1_QuantizeLinear_ClipQuant (Clip)",
[[33mWARNING[0m]   "conv2_2_DequantizeLinear_SubDeQuant (Sub)", "conv2_2_DequantizeLinear_MulDeQuant
[[33mWARNING[0m]   (Mul)", "pool5_1_DequantizeLinear_SubDeQuant (Sub)", "pool2_1_QuantizeLinear_MulQuant
[[33mWARNING[0m]   (Mul)", "pool2_1_QuantizeLinear_AddQuant (Add)", "pool2_1_QuantizeLinear_ClipQuant
[[33mWARNING[0m]   (Clip)", "Conv_nc_rename_10_quant (QLinearConv)", "Conv_nc_rename_8_quant
[[33mWARNING[0m]   (QLinearConv)", "Conv_nc_rename_12_quant (QLinearConv)",
[[33mWARNING[0m]   "pool5_1_DequantizeLinear_MulDeQuant (Mul)", "OC2_DUMMY_0_QuantizeLinear_MulQuant
[[33mWARNING[0m]   (Mul)", "Reshape_nc_rename_15 (Reshape)", "OC2_DUMMY_0_QuantizeLinear_AddQuant (Add)",
[[33mWARNING[0m]   "OC2_DUMMY_0_QuantizeLinear_ClipQuant (Clip)", "Add_nc_rename_17_quant (QLinearAdd)",
[[33mWARNING[0m]   "Add_nc_rename_20_quant (QLinearAdd)", "Add_nc_rename_23_quant (QLinearAdd)",
[[33mWARNING[0m]   "fc8_1_DequantizeLinear_SubDeQuant (Sub)", "fc8_1_DequantizeLinear_MulDeQuant (Mul)", "
[[33mWARNING[0m]   (QMatMul)", " (QMatMul)"].
[[33mWARNING[0m] - GenNode(name='pool2_1_QuantizeLinear_ClipQuantMin', optype='Producer', children: [[1]])
  Node(name='conv1_2_DequantizeLinear_MulDeQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='conv1_2_DequantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='ZeroPoint', optype='Producer', children: [[1]])
  Node(name='conv2_w_0_scale', optype='Producer', children: [[1]])
  Node(name='MaxPool_nc_rename_3', optype='MaxPooling2D', parents: [1], children: [[1], []])
  Node(name='ScalingFactor_1', optype='Producer', children: [[1]])
  Node(name='conv1_2_DequantizeLinear_SubDeQuant', optype='Sub', parents: [1, 1], children: [[1]])
  Node(name='ScalingFactor', optype='Producer', children: [[1]])
  Node(name='ZeroPoint_1', optype='Producer', children: [[1]])
  Node(name='data_0_QuantizeLinear_CastOut', optype='Cast', parents: [1], children: [[1]])
  Node(name='pool5_1_DequantizeLinear_SubDeQuant', optype='Sub', parents: [1, 1], children: [[1]])
  Node(name='data_0_QuantizeLinear_CastIn', optype='Cast', parents: [0], children: [[1]])
  Node(name='data_0_QuantizeLinear_RoundQuant', optype='Round', parents: [1], children: [[1]])
  Node(name='conv2_2_DequantizeLinear_CastIn', optype='Cast', parents: [1], children: [[1]])
  Node(name='data_0_QuantizeLinear_ClipQuantMax', optype='Producer', children: [[1]])
  Node(name='pool1_1_scale', optype='Producer', children: [[1]])
  Node(name='fc6_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='data_0_QuantizeLinear_ClipQuantMin', optype='Producer', children: [[1]])
  Node(name='data_0_QuantizeLinear_ClipQuant', optype='Clip', parents: [1, 1, 1], children: [[1]])
  Node(name='conv1_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='fc7_1_MatMul_zero_point', optype='Producer', children: [[1]])
  Node(name='ZeroPoint', optype='Producer', children: [[1]])
  Node(name='data_0_QuantizeLinear_AddQuant', optype='Add', parents: [1, 1], children: [[1]])
  Node(name='ScalingFactor', optype='Producer', children: [[1]])
  Node(name='fc7_1_MatMul_scale', optype='Producer', children: [[1]])
  Node(name='data_0_QuantizeLinear_MulQuant', optype='Mul', parents: [1, 1], children: [[1]])
  Node(name='fc8_b_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc7_1_scale', optype='Producer', children: [[1]])
  Node(name='fc8_b_0_quantized', optype='Producer', children: [[1]])
  Node(name='conv3_w_0_scale', optype='Producer', children: [[1]])
  Node(name='fc8_1_scale', optype='Producer', children: [[1]])
  Node(name='fc8_b_0_scale', optype='Producer', children: [[1]])
  Node(name='fc8_1_zero_point', optype='Producer', children: [[1]])
  Node(name='fc8_1_MatMul_scale', optype='Producer', children: [[1]])
  Node(name='fc8_1_MatMul_zero_point', optype='Producer', children: [[1]])
  Node(name='data_0_zero_point', optype='Producer', children: [[1]])
  Node(name='fc8_w_0_quantized', optype='Producer', children: [[1]])
  Node(name='Add_nc_rename_20_quant', optype='QLinearAdd', parents: [1, 1, 1, 1, 1, 1, 1, 1], children: [[1]])

===============
Supported nodes 2
===============

Native operators: 137 (14 types)
- Add: 4
- Cast: 12
- Clip: 4
- Dequantizer: 6
- LRN: 2
- MaxPooling2D: 2
- Mul: 8
- PadMaxPool: 1
- Producer: 85
- QMatMul: 3
- Reshape: 1
- Round: 4
- Softmax: 1
- Sub: 4
Generic operators: 8 (2 types)
- QLinearAdd: 3
- QLinearConv: 5
Native types coverage: 87.5% (14/16)
Native operators coverage: 94.5% (137/145)
  (defaultdict(<class 'int'>, {'Producer': 85, 'PadMaxPool': 1, 'Softmax': 1, 'LRN': 2, 'QMatMul': 3, 'Cast': 12, 'Mul': 8, 'Sub': 4, 'Dequantizer': 6, 'Round': 4, 'Clip': 4, 'Add': 4, 'Reshape': 1, 'MaxPooling2D': 2}), defaultdict(<class 'int'>, {'QLinearAdd': 3, 'QLinearConv': 5}))

===============
Supported nodes
===============

Native operators: 137 (14 types)
- Add: 4
- Cast: 12
- Clip: 4
- Dequantizer: 6
- LRN: 2
- MaxPooling2D: 2
- Mul: 8
- PadMaxPool: 1
- Producer: 85
- QMatMul: 3
- Reshape: 1
- Round: 4
- Softmax: 1
- Sub: 4
Generic operators: 8 (2 types)
- QLinearAdd: 3
- QLinearConv: 5
Native types coverage: 87.5% (14/16)
Native operators coverage: 94.5% (137/145)
  (defaultdict(<class 'int'>, {'Producer': 85, 'PadMaxPool': 1, 'LRN': 2, 'Softmax': 1, 'QMatMul': 3, 'Cast': 12, 'Mul': 8, 'Sub': 4, 'Dequantizer': 6, 'Round': 4, 'Clip': 4, 'Reshape': 1, 'Add': 4, 'MaxPooling2D': 2}), defaultdict(<class 'int'>, {'QLinearAdd': 3, 'QLinearConv': 5}))

===============
Compile
===============

OK

===============
Create Scheduler
===============

OK

===============
Name nodes
===============

data_0_zero_point (Producer)
OC2_DUMMY_1 (Producer)
ScalingFactor_1 (Producer)
conv1_b_0_quantized (Producer)
conv2_w_0_zero_point (Producer)
conv2_1_zero_point (Producer)
conv2_b_0_quantized (Producer)
conv2_1_scale (Producer)
pool2_1_zero_point (Producer)
conv1_1_scale (Producer)
pool1_1_zero_point (Producer)
pool1_1_scale (Producer)
conv2_w_0_scale (Producer)
pool2_1_scale (Producer)
conv3_w_0_quantized (Producer)
conv3_w_0_zero_point (Producer)
conv3_w_0_scale (Producer)
conv3_1_zero_point (Producer)
conv3_b_0_quantized (Producer)
conv3_1_scale (Producer)
conv4_w_0_quantized (Producer)
conv4_w_0_zero_point (Producer)
conv4_w_0_scale (Producer)
conv4_1_zero_point (Producer)
conv4_b_0_quantized (Producer)
conv4_1_scale (Producer)
conv5_w_0_quantized (Producer)
conv5_w_0_scale (Producer)
conv5_w_0_zero_point (Producer)
conv5_b_0_quantized (Producer)
conv5_1_zero_point (Producer)
conv5_1_scale (Producer)
fc6_w_0_quantized (Producer)
fc6_1_MatMul_zero_point (Producer)
ScalingFactor (Producer)
fc6_1_MatMul_scale (Producer)
fc6_1_zero_point (Producer)
fc6_1_scale (Producer)
fc6_b_0_scale (Producer)
fc6_b_0_zero_point (Producer)
fc6_b_0_quantized (Producer)
fc7_w_0_quantized (Producer)
fc7_1_MatMul_zero_point (Producer)
fc7_1_MatMul_scale (Producer)
fc7_1_zero_point (Producer)
fc7_1_scale (Producer)
fc7_b_0_scale (Producer)
fc7_b_0_zero_point (Producer)
fc7_b_0_quantized (Producer)
fc8_w_0_quantized (Producer)
fc8_1_MatMul_zero_point (Producer)
fc8_1_MatMul_scale (Producer)
fc8_1_zero_point (Producer)
fc8_1_scale (Producer)
fc8_b_0_scale (Producer)
fc8_b_0_quantized (Producer)
fc8_b_0_zero_point (Producer)
ScalingFactor (Producer)
ZeroPoint (Producer)
data_0_QuantizeLinear_ClipQuantMin (Producer)
data_0_QuantizeLinear_ClipQuantMax (Producer)
ZeroPoint_1 (Producer)
ScalingFactor (Producer)
ZeroPoint (Producer)
pool1_1_QuantizeLinear_ClipQuantMin (Producer)
pool1_1_QuantizeLinear_ClipQuantMax (Producer)
ZeroPoint_2 (Producer)
ScalingFactor_2 (Producer)
ScalingFactor (Producer)
ZeroPoint (Producer)
pool2_1_QuantizeLinear_ClipQuantMin (Producer)
pool2_1_QuantizeLinear_ClipQuantMax (Producer)
conv2_w_0_quantized (Producer)
ZeroPoint_3 (Producer)
ScalingFactor_3 (Producer)
ZeroPoint (Producer)
OC2_DUMMY_0_QuantizeLinear_ClipQuantMin (Producer)
OC2_DUMMY_0_QuantizeLinear_ClipQuantMax (Producer)
ZeroPoint_4 (Producer)
ScalingFactor_4 (Producer)
conv1_1_zero_point (Producer)
conv1_w_0_scale (Producer)
conv1_w_0_quantized (Producer)
data_0_scale (Producer)
conv1_w_0_zero_point (Producer)
_Cast_0 (Cast)
_Dequantizer_0 (Dequantizer)
_Dequantizer_1 (Dequantizer)
_Dequantizer_2 (Dequantizer)
_Mul_0 (Mul)
_Round_0 (Round)
_Add_0 (Add)
_Clip_0 (Clip)
_Cast_1 (Cast)
_QLinearConv_0 (QLinearConv)
_Cast_2 (Cast)
_Sub_0 (Sub)
_Mul_1 (Mul)
_LRN_0 (LRN)
_MaxPooling2D_0 (MaxPooling2D)
_Cast_3 (Cast)
_Mul_2 (Mul)
_Round_1 (Round)
_Add_1 (Add)
_Clip_1 (Clip)
_Cast_4 (Cast)
_QLinearConv_1 (QLinearConv)
_Cast_5 (Cast)
_Sub_1 (Sub)
_Mul_3 (Mul)
_LRN_1 (LRN)
_MaxPooling2D_1 (MaxPooling2D)
_Cast_6 (Cast)
_Mul_4 (Mul)
_Round_2 (Round)
_Add_2 (Add)
_Clip_2 (Clip)
_Cast_7 (Cast)
_QLinearConv_2 (QLinearConv)
_QLinearConv_3 (QLinearConv)
_QLinearConv_4 (QLinearConv)
_PadMaxPool_0 (PadMaxPool)
_Cast_8 (Cast)
_Sub_2 (Sub)
_Mul_5 (Mul)
_Reshape_0 (Reshape)
_Cast_9 (Cast)
_Mul_6 (Mul)
_Round_3 (Round)
_Add_3 (Add)
_Clip_3 (Clip)
_Cast_10 (Cast)
_Dequantizer_3 (Dequantizer)
_QMatMul_0 (QMatMul)
_QLinearAdd_0 (QLinearAdd)
_Dequantizer_4 (Dequantizer)
_QMatMul_1 (QMatMul)
_QLinearAdd_1 (QLinearAdd)
_Dequantizer_5 (Dequantizer)
_QMatMul_2 (QMatMul)
_QLinearAdd_2 (QLinearAdd)
_Cast_11 (Cast)
_Sub_3 (Sub)
_Mul_7 (Mul)
_Softmax_0 (Softmax)

===============
Set backend
===============

ericOperator: cannot compute output dims, no ComputeDimsFunc function provided.
[[33mWARNING[0m] - GenericOperator: cannot compute output dims, no ComputeDimsFunc function provided.
[[33mWARNING[0m] - GenericOperator: cannot compute output dims, no ComputeDimsFunc function provided.
[[33mWARNING[0m] - Unable to forward dimensions (circular dependency and/or wrong dimensions and/or data
[[33mWARNING[0m]   dependent dimension?). Unable to compute output dims for nodes [" (QMatMul)", "
[[33mWARNING[0m]   (PadMaxPool)", "conv1_2_DequantizeLinear_SubDeQuant (Sub)",
[[33mWARNING[0m]   "conv1_2_DequantizeLinear_MulDeQuant (Mul)", "Conv_nc_rename_0_quant (QLinearConv)",
[[33mWARNING[0m]   "Conv_nc_rename_4_quant (QLinearConv)", "pool1_1_QuantizeLinear_MulQuant (Mul)",
[[33mWARNING[0m]   "pool1_1_QuantizeLinear_AddQuant (Add)", "pool1_1_QuantizeLinear_ClipQuant (Clip)",
[[33mWARNING[0m]   "conv2_2_DequantizeLinear_SubDeQuant (Sub)", "conv2_2_DequantizeLinear_MulDeQuant
[[33mWARNING[0m]   (Mul)", "pool5_1_DequantizeLinear_SubDeQuant (Sub)", "pool2_1_QuantizeLinear_MulQuant
[[33mWARNING[0m]   (Mul)", "pool2_1_QuantizeLinear_AddQuant (Add)", "pool2_1_QuantizeLinear_ClipQuant
[[33mWARNING[0m]   (Clip)", "Conv_nc_rename_10_quant (QLinearConv)", "Conv_nc_rename_8_quant
[[33mWARNING[0m]   (QLinearConv)", "Conv_nc_rename_12_quant (QLinearConv)",
[[33mWARNING[0m]   "pool5_1_DequantizeLinear_MulDeQuant (Mul)", "OC2_DUMMY_0_QuantizeLinear_MulQuant
[[33mWARNING[0m]   (Mul)", "Reshape_nc_rename_15 (Reshape)", "OC2_DUMMY_0_QuantizeLinear_AddQuant (Add)",
[[33mWARNING[0m]   "OC2_DUMMY_0_QuantizeLinear_ClipQuant (Clip)", "Add_nc_rename_17_quant (QLinearAdd)",
[[33mWARNING[0m]   "Add_nc_rename_20_quant (QLinearAdd)", "Add_nc_rename_23_quant (QLinearAdd)",
[[33mWARNING[0m]   "fc8_1_DequantizeLinear_SubDeQuant (Sub)", "fc8_1_DequantizeLinear_MulDeQuant (Mul)", "
[[33mWARNING[0m]   (QMatMul)", " (QMatMul)"].
[[31mERROR[0m] - Assertion failed: exists(key) in /opt/aidge/aidge/aidge_core/include/aidge/utils/Registrar.hpp:87
[[95mFATAL[0m] - missing or invalid registrar key: "export_cpp" for registrable object N5Aidge8Round_OpE
[[95mFATAL[0m]   Did you include/import the corresponding module?
[[95mFATAL[0m]   If so, it is possible that the object is not yet supported.
[[95mFATAL[0m]   
[[95mFATAL[0m]   Available registrar keys are:
[[95mFATAL[0m]       cpu
Traceback (most recent call last):
  File "/app/AI_Project/01_generate_cpp.py", line 115, in <module>
    model_aidge.set_backend(aidge_export_cpp.ExportLibCpp._name)
RuntimeError: missing or invalid registrar key: "export_cpp" for registrable object N5Aidge8Round_OpE
Did you include/import the corresponding module?
If so, it is possible that the object is not yet supported.

Available registrar keys are:
    cpu
Error: export_model/data directory does not exist.

error.log

sed: can't read export_model/Makefile: No such file or directory
sed: can't read export_model/Makefile: No such file or directory
sed: can't read export_model/Makefile: No such file or directory
sed: can't read export_model/Makefile: No such file or directory
./AI_Project/03_build.sh: line 17: cd: export_model: No such file or directory
make: AI_Build AI_Deploy AI_Manager AI_Project AI_Support ConvNet.onnx MLP_MNIST.onnx MobileNet-v2.onnx README.md __pycache__ bvlcalexnet12int8.onnx config.json docker examples exit_functions.sh mnist.onnx mnist_test_input_type2.bin model_1D_classifier.onnx model_type2.onnx print_raw_output.py type1_test.sh type2_test.sh type3_test.sh No targets specified and no makefile found. Stop.

Report Details

report.json

null

124074 - rpi5 aidge 124076 - rpi5 aidge