Forums - (SNPE v.1.6.0) UDL Tutorial does not work

3 posts / 0 new

or Register

Last post

(SNPE v.1.6.0) UDL Tutorial does not work

mudria

Join Date: 25 Jun 17

Posts: 4

Posted: Thu, 2017-11-02 19:56

Top

I'm working with SNPE v.1.6.0.

I have followed steps in UDL Tutorial. (snpe-1.6.0/doc/html/udl_tutorial.html).

When I run snpe-caffe-to-dlc-udl, it gave me following error!

Encountered Error'module' object has no attribute 'UdlBlobOutput'

These are my outcomes. Please fix the problem.

sujungbae@sujungbae-VirtualBox:~/Qualcomm/snpe-1.6.0/examples/Python/UdlExample$ python snpe-caffe-to-dlc-udl \
> --caffe_txt $CAFFE_HOME/examples/mnist/mycustomlenet.prototxt \
> --caffe_bin $CAFFE_HOME/mycustomlenet_iter_10000.caffemodel \
> --dlc mycustomlenet.dlc
WARNING: Logging before InitGoogleLogging() is written to STDERR
W1103 11:42:46.729535 29946 _caffe.cpp:139] DEPRECATION WARNING - deprecated use of Python interface
W1103 11:42:46.729598 29946 _caffe.cpp:140] Use this instead (with the named "weights" parameter):
W1103 11:42:46.729606 29946 _caffe.cpp:142] Net('/home/sujungbae/caffe/examples/mnist/mycustomlenet.prototxt', 1, weights='/home/sujungbae/caffe/mycustomlenet_iter_10000.caffemodel')
I1103 11:42:46.730643 29946 net.cpp:51] Initializing net from parameters:
name: "LeNet"
state {
phase: TEST
level: 0
}
layer {
name: "data"
type: "Input"
top: "data"
input_param {
    shape {
      dim: 64
      dim: 1
      dim: 28
      dim: 28
    }
}
}
layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
    lr_mult: 1
}
param {
    lr_mult: 2
}
convolution_param {
    num_output: 20
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "conv1"
top: "pool1"
pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
    lr_mult: 1
}
param {
    lr_mult: 2
}
convolution_param {
    num_output: 50
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
}
}
layer {
name: "pool2"
type: "Pooling"
bottom: "conv2"
top: "pool2"
pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
}
}
layer {
name: "ip1"
type: "InnerProduct"
bottom: "pool2"
top: "ip1"
param {
    lr_mult: 1
}
param {
    lr_mult: 2
}
inner_product_param {
    num_output: 500
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "ip1"
top: "ip1"
}
layer {
name: "scale"
type: "MyCustomScale"
bottom: "ip1"
top: "scale"
scale_param {
    bias_term: false
}
}
layer {
name: "ip2"
type: "InnerProduct"
bottom: "scale"
top: "ip2"
param {
    lr_mult: 1
}
param {
    lr_mult: 2
}
inner_product_param {
    num_output: 10
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
}
}
layer {
name: "prob"
type: "Softmax"
bottom: "ip2"
top: "prob"
}
I1103 11:42:46.730976 29946 layer_factory.hpp:77] Creating layer data
I1103 11:42:46.730993 29946 net.cpp:84] Creating Layer data
I1103 11:42:46.731003 29946 net.cpp:380] data -> data
I1103 11:42:46.731030 29946 net.cpp:122] Setting up data
I1103 11:42:46.731045 29946 net.cpp:129] Top shape: 64 1 28 28 (50176)
I1103 11:42:46.731052 29946 net.cpp:137] Memory required for data: 200704
I1103 11:42:46.731081 29946 layer_factory.hpp:77] Creating layer conv1
I1103 11:42:46.731096 29946 net.cpp:84] Creating Layer conv1
I1103 11:42:46.731106 29946 net.cpp:406] conv1 <- data
I1103 11:42:46.731117 29946 net.cpp:380] conv1 -> conv1
I1103 11:42:46.731377 29946 net.cpp:122] Setting up conv1
I1103 11:42:46.731392 29946 net.cpp:129] Top shape: 64 20 24 24 (737280)
I1103 11:42:46.731400 29946 net.cpp:137] Memory required for data: 3149824
I1103 11:42:46.731415 29946 layer_factory.hpp:77] Creating layer pool1
I1103 11:42:46.731427 29946 net.cpp:84] Creating Layer pool1
I1103 11:42:46.731436 29946 net.cpp:406] pool1 <- conv1
I1103 11:42:46.731446 29946 net.cpp:380] pool1 -> pool1
I1103 11:42:46.731462 29946 net.cpp:122] Setting up pool1
I1103 11:42:46.731472 29946 net.cpp:129] Top shape: 64 20 12 12 (184320)
I1103 11:42:46.731479 29946 net.cpp:137] Memory required for data: 3887104
I1103 11:42:46.731488 29946 layer_factory.hpp:77] Creating layer conv2
I1103 11:42:46.731499 29946 net.cpp:84] Creating Layer conv2
I1103 11:42:46.731508 29946 net.cpp:406] conv2 <- pool1
I1103 11:42:46.731519 29946 net.cpp:380] conv2 -> conv2
I1103 11:42:46.731659 29946 net.cpp:122] Setting up conv2
I1103 11:42:46.731673 29946 net.cpp:129] Top shape: 64 50 8 8 (204800)
I1103 11:42:46.731681 29946 net.cpp:137] Memory required for data: 4706304
I1103 11:42:46.731693 29946 layer_factory.hpp:77] Creating layer pool2
I1103 11:42:46.731703 29946 net.cpp:84] Creating Layer pool2
I1103 11:42:46.731712 29946 net.cpp:406] pool2 <- conv2
I1103 11:42:46.731722 29946 net.cpp:380] pool2 -> pool2
I1103 11:42:46.731735 29946 net.cpp:122] Setting up pool2
I1103 11:42:46.731745 29946 net.cpp:129] Top shape: 64 50 4 4 (51200)
I1103 11:42:46.731753 29946 net.cpp:137] Memory required for data: 4911104
I1103 11:42:46.731761 29946 layer_factory.hpp:77] Creating layer ip1
I1103 11:42:46.731771 29946 net.cpp:84] Creating Layer ip1
I1103 11:42:46.731781 29946 net.cpp:406] ip1 <- pool2
I1103 11:42:46.731789 29946 net.cpp:380] ip1 -> ip1
I1103 11:42:46.734031 29946 net.cpp:122] Setting up ip1
I1103 11:42:46.734048 29946 net.cpp:129] Top shape: 64 500 (32000)
I1103 11:42:46.734056 29946 net.cpp:137] Memory required for data: 5039104
I1103 11:42:46.734068 29946 layer_factory.hpp:77] Creating layer relu1
I1103 11:42:46.734079 29946 net.cpp:84] Creating Layer relu1
I1103 11:42:46.734088 29946 net.cpp:406] relu1 <- ip1
I1103 11:42:46.734097 29946 net.cpp:367] relu1 -> ip1 (in-place)
I1103 11:42:46.734108 29946 net.cpp:122] Setting up relu1
I1103 11:42:46.734118 29946 net.cpp:129] Top shape: 64 500 (32000)
I1103 11:42:46.734125 29946 net.cpp:137] Memory required for data: 5167104
I1103 11:42:46.734133 29946 layer_factory.hpp:77] Creating layer scale
I1103 11:42:46.734143 29946 net.cpp:84] Creating Layer scale
I1103 11:42:46.734151 29946 net.cpp:406] scale <- ip1
I1103 11:42:46.734161 29946 net.cpp:380] scale -> scale
I1103 11:42:46.734177 29946 net.cpp:122] Setting up scale
I1103 11:42:46.734187 29946 net.cpp:129] Top shape: 64 500 (32000)
I1103 11:42:46.734195 29946 net.cpp:137] Memory required for data: 5295104
I1103 11:42:46.734205 29946 layer_factory.hpp:77] Creating layer ip2
I1103 11:42:46.734215 29946 net.cpp:84] Creating Layer ip2
I1103 11:42:46.734223 29946 net.cpp:406] ip2 <- scale
I1103 11:42:46.734233 29946 net.cpp:380] ip2 -> ip2
I1103 11:42:46.734272 29946 net.cpp:122] Setting up ip2
I1103 11:42:46.734283 29946 net.cpp:129] Top shape: 64 10 (640)
I1103 11:42:46.734290 29946 net.cpp:137] Memory required for data: 5297664
I1103 11:42:46.734302 29946 layer_factory.hpp:77] Creating layer prob
I1103 11:42:46.734313 29946 net.cpp:84] Creating Layer prob
I1103 11:42:46.734321 29946 net.cpp:406] prob <- ip2
I1103 11:42:46.734330 29946 net.cpp:380] prob -> prob
I1103 11:42:46.734344 29946 net.cpp:122] Setting up prob
I1103 11:42:46.734361 29946 net.cpp:129] Top shape: 64 10 (640)
I1103 11:42:46.734369 29946 net.cpp:137] Memory required for data: 5300224
I1103 11:42:46.734378 29946 net.cpp:200] prob does not need backward computation.
I1103 11:42:46.734386 29946 net.cpp:200] ip2 does not need backward computation.
I1103 11:42:46.734395 29946 net.cpp:200] scale does not need backward computation.
I1103 11:42:46.734403 29946 net.cpp:200] relu1 does not need backward computation.
I1103 11:42:46.734411 29946 net.cpp:200] ip1 does not need backward computation.
I1103 11:42:46.734421 29946 net.cpp:200] pool2 does not need backward computation.
I1103 11:42:46.734428 29946 net.cpp:200] conv2 does not need backward computation.
I1103 11:42:46.734436 29946 net.cpp:200] pool1 does not need backward computation.
I1103 11:42:46.734446 29946 net.cpp:200] conv1 does not need backward computation.
I1103 11:42:46.734453 29946 net.cpp:200] data does not need backward computation.
I1103 11:42:46.734462 29946 net.cpp:242] This network produces output prob
I1103 11:42:46.734472 29946 net.cpp:255] Network initialization done.
I1103 11:42:46.735786 29946 net.cpp:744] Ignoring source layer mnist
I1103 11:42:46.736049 29946 net.cpp:744] Ignoring source layer loss
Encountered Error'module' object has no attribute 'UdlBlobOutput'

Forum vote up/down

Re: (SNPE v.1.6.0) UDL Tutorial does not work #1

phongnhhn92

Join Date: 27 Oct 17

Posts: 8

Posted: Sat, 2017-11-11 00:50

Top

I am having the same issue as well. Please check !

Re: (SNPE v.1.6.0) UDL Tutorial does not work #2

howardd

Join Date: 3 Nov 17

Posts: 5

Posted: Mon, 2017-11-13 10:52

Top

Hello,

It seems that the example script is attempting to access the UdlBlobOutput utiliy from the wrong module. Could you please modify line 86 of snpe-caffe-to-dlc-udl to read as follows:

return snpe_udl_utils.UdlBlobOutput(blob=blob, out_dims=input_dims)

and let us know if that resolves the error?

or Register

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.

Sort By

Filter Results