yolov3-tiny-onnx-TensorRT
yolov3-tiny-onnx-TensorRT copied to clipboard
zombie0117
Error with new node ‘Upsample’
Hi, it looks like you have done some work with new node upsample in code, but after convert I got error about upsample below. Have you met problem like this? Thanks look forward your reply!
Layer of type yolo not supported, skipping ONNX node generation.
Layer of type yolo not supported, skipping ONNX node generation.
graph YOLOv3-tiny-416 (
%000_net[FLOAT, 1x3x416x416]
) initializers (
%001_convolutional_bn_scale[FLOAT, 16]
%001_convolutional_bn_bias[FLOAT, 16]
%001_convolutional_bn_mean[FLOAT, 16]
%001_convolutional_bn_var[FLOAT, 16]
%001_convolutional_conv_weights[FLOAT, 16x3x3x3]
%003_convolutional_bn_scale[FLOAT, 32]
%003_convolutional_bn_bias[FLOAT, 32]
%003_convolutional_bn_mean[FLOAT, 32]
%003_convolutional_bn_var[FLOAT, 32]
%003_convolutional_conv_weights[FLOAT, 32x16x3x3]
%005_convolutional_bn_scale[FLOAT, 64]
%005_convolutional_bn_bias[FLOAT, 64]
%005_convolutional_bn_mean[FLOAT, 64]
%005_convolutional_bn_var[FLOAT, 64]
%005_convolutional_conv_weights[FLOAT, 64x32x3x3]
%007_convolutional_bn_scale[FLOAT, 128]
%007_convolutional_bn_bias[FLOAT, 128]
%007_convolutional_bn_mean[FLOAT, 128]
%007_convolutional_bn_var[FLOAT, 128]
%007_convolutional_conv_weights[FLOAT, 128x64x3x3]
%009_convolutional_bn_scale[FLOAT, 256]
%009_convolutional_bn_bias[FLOAT, 256]
%009_convolutional_bn_mean[FLOAT, 256]
%009_convolutional_bn_var[FLOAT, 256]
%009_convolutional_conv_weights[FLOAT, 256x128x3x3]
%011_convolutional_bn_scale[FLOAT, 512]
%011_convolutional_bn_bias[FLOAT, 512]
%011_convolutional_bn_mean[FLOAT, 512]
%011_convolutional_bn_var[FLOAT, 512]
%011_convolutional_conv_weights[FLOAT, 512x256x3x3]
%013_convolutional_bn_scale[FLOAT, 1024]
%013_convolutional_bn_bias[FLOAT, 1024]
%013_convolutional_bn_mean[FLOAT, 1024]
%013_convolutional_bn_var[FLOAT, 1024]
%013_convolutional_conv_weights[FLOAT, 1024x512x3x3]
%014_convolutional_bn_scale[FLOAT, 256]
%014_convolutional_bn_bias[FLOAT, 256]
%014_convolutional_bn_mean[FLOAT, 256]
%014_convolutional_bn_var[FLOAT, 256]
%014_convolutional_conv_weights[FLOAT, 256x1024x1x1]
%015_convolutional_bn_scale[FLOAT, 512]
%015_convolutional_bn_bias[FLOAT, 512]
%015_convolutional_bn_mean[FLOAT, 512]
%015_convolutional_bn_var[FLOAT, 512]
%015_convolutional_conv_weights[FLOAT, 512x256x3x3]
%016_convolutional_conv_bias[FLOAT, 21]
%016_convolutional_conv_weights[FLOAT, 21x512x1x1]
%019_convolutional_bn_scale[FLOAT, 128]
%019_convolutional_bn_bias[FLOAT, 128]
%019_convolutional_bn_mean[FLOAT, 128]
%019_convolutional_bn_var[FLOAT, 128]
%019_convolutional_conv_weights[FLOAT, 128x256x1x1]
%020_upsample_scale[FLOAT, 4]
%022_convolutional_bn_scale[FLOAT, 256]
%022_convolutional_bn_bias[FLOAT, 256]
%022_convolutional_bn_mean[FLOAT, 256]
%022_convolutional_bn_var[FLOAT, 256]
%022_convolutional_conv_weights[FLOAT, 256x384x3x3]
%023_convolutional_conv_bias[FLOAT, 21]
%023_convolutional_conv_weights[FLOAT, 21x256x1x1]
) {
%001_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%000_net, %001_convolutional_conv_weights)
%001_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%001_convolutional, %001_convolutional_bn_scale, %001_convolutional_bn_bias, %001_convolutional_bn_mean, %001_convolutional_bn_var)
%001_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%002_maxpool = MaxPoolauto_pad = u'SAME_UPPER', kernel_shape = [2, 2], strides = [2, 2]
%003_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%002_maxpool, %003_convolutional_conv_weights)
%003_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%003_convolutional, %003_convolutional_bn_scale, %003_convolutional_bn_bias, %003_convolutional_bn_mean, %003_convolutional_bn_var)
%003_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%004_maxpool = MaxPoolauto_pad = u'SAME_UPPER', kernel_shape = [2, 2], strides = [2, 2]
%005_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%004_maxpool, %005_convolutional_conv_weights)
%005_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%005_convolutional, %005_convolutional_bn_scale, %005_convolutional_bn_bias, %005_convolutional_bn_mean, %005_convolutional_bn_var)
%005_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%006_maxpool = MaxPoolauto_pad = u'SAME_UPPER', kernel_shape = [2, 2], strides = [2, 2]
%007_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%006_maxpool, %007_convolutional_conv_weights)
%007_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%007_convolutional, %007_convolutional_bn_scale, %007_convolutional_bn_bias, %007_convolutional_bn_mean, %007_convolutional_bn_var)
%007_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%008_maxpool = MaxPoolauto_pad = u'SAME_UPPER', kernel_shape = [2, 2], strides = [2, 2]
%009_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%008_maxpool, %009_convolutional_conv_weights)
%009_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%009_convolutional, %009_convolutional_bn_scale, %009_convolutional_bn_bias, %009_convolutional_bn_mean, %009_convolutional_bn_var)
%009_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%010_maxpool = MaxPoolauto_pad = u'SAME_UPPER', kernel_shape = [2, 2], strides = [2, 2]
%011_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%010_maxpool, %011_convolutional_conv_weights)
%011_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%011_convolutional, %011_convolutional_bn_scale, %011_convolutional_bn_bias, %011_convolutional_bn_mean, %011_convolutional_bn_var)
%011_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%012_maxpool = MaxPoolauto_pad = u'SAME_UPPER', kernel_shape = [2, 2], strides = [1, 1]
%013_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%012_maxpool, %013_convolutional_conv_weights)
%013_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%013_convolutional, %013_convolutional_bn_scale, %013_convolutional_bn_bias, %013_convolutional_bn_mean, %013_convolutional_bn_var)
%013_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%014_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [1, 1], strides = [1, 1]](%013_convolutional_lrelu, %014_convolutional_conv_weights)
%014_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%014_convolutional, %014_convolutional_bn_scale, %014_convolutional_bn_bias, %014_convolutional_bn_mean, %014_convolutional_bn_var)
%014_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%015_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%014_convolutional_lrelu, %015_convolutional_conv_weights)
%015_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%015_convolutional, %015_convolutional_bn_scale, %015_convolutional_bn_bias, %015_convolutional_bn_mean, %015_convolutional_bn_var)
%015_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%016_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [1, 1], strides = [1, 1]](%015_convolutional_lrelu, %016_convolutional_conv_weights, %016_convolutional_conv_bias)
%019_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [1, 1], strides = [1, 1]](%014_convolutional_lrelu, %019_convolutional_conv_weights)
%019_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%019_convolutional, %019_convolutional_bn_scale, %019_convolutional_bn_bias, %019_convolutional_bn_mean, %019_convolutional_bn_var)
%019_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%020_upsample = Upsample[mode = u'nearest'](%019_convolutional_lrelu, %020_upsample_scale)
%021_route = Concat[axis = 1](%020_upsample, %009_convolutional_lrelu)
%022_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [3, 3], strides = [1, 1]](%021_route, %022_convolutional_conv_weights)
%022_convolutional_bn = BatchNormalization[epsilon = 9.99999974737875e-06, momentum = 0.990000009536743](%022_convolutional, %022_convolutional_bn_scale, %022_convolutional_bn_bias, %022_convolutional_bn_mean, %022_convolutional_bn_var)
%022_convolutional_lrelu = LeakyRelualpha = 0.100000001490116
%023_convolutional = Conv[auto_pad = u'SAME_LOWER', dilations = [1, 1], kernel_shape = [1, 1], strides = [1, 1]](%022_convolutional_lrelu, %023_convolutional_conv_weights, %023_convolutional_conv_bias)
return %016_convolutional, %023_convolutional
}
Traceback (most recent call last):
File "/home/jxiao/github/yolov3-tiny-onnx-TensorRT-master/yolov3_to_onnx.py", line 830, in
==> Context: Bad node spec: input: "019_convolutional_lrelu" input: "020_upsample_scale" output: "020_upsample" name: "020_upsample" op_type: "Upsample" attribute { name: "mode" s: "nearest" type: STRING }
Process finished with exit code 1
