In [1]:

    

using DataFrames
using TensorFlow
using Distributions
using Images


    

    




I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcublas.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:116] Couldn't open CUDA library libcudnn.so.5. LD_LIBRARY_PATH: /usr/local/cuda/lib64
I tensorflow/stream_executor/cuda/cuda_dnn.cc:3459] Unable to load cuDNN DSO
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcufft.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcurand.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcublas.so locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcudnn.so locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcufft.so locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcurand.so locally

In [2]:

    

data_set_folder = "/raided/datasets/sketches_png/png/"
list_of_imgs = open(data_set_folder * "filelist.txt")
img_paths = readlines(list_of_imgs)


    

    
Out[2]:





20000-element Array{String,1}:
 "airplane/1.png\n" 
 "airplane/10.png\n"
 "airplane/11.png\n"
 "airplane/12.png\n"
 "airplane/13.png\n"
 "airplane/14.png\n"
 "airplane/15.png\n"
 "airplane/16.png\n"
 "airplane/17.png\n"
 "airplane/18.png\n"
 "airplane/19.png\n"
 "airplane/2.png\n" 
 "airplane/20.png\n"
 ⋮                  
 "zebra/19989.png\n"
 "zebra/19990.png\n"
 "zebra/19991.png\n"
 "zebra/19992.png\n"
 "zebra/19993.png\n"
 "zebra/19994.png\n"
 "zebra/19995.png\n"
 "zebra/19996.png\n"
 "zebra/19997.png\n"
 "zebra/19998.png\n"
 "zebra/19999.png\n"
 "zebra/20000.png\n"

In [3]:

    

test_paths = Array(String, 2500)
path_index = 71
count = 1
for i = 1:2500
    test_paths[i] = img_paths[path_index]
    path_index += 1
    count += 1
    if count % 11 == 0
        path_index += 70
        count = 1
    end
end


    

In [4]:

    

files_in_folder = readdir(data_set_folder)
labels = filter(x->isdir(data_set_folder * x) == true, files_in_folder)


    

    
Out[4]:





250-element Array{String,1}:
 "airplane"     
 "alarm clock"  
 "angel"        
 "ant"          
 "apple"        
 "arm"          
 "armchair"     
 "ashtray"      
 "axe"          
 "backpack"     
 "banana"       
 "barn"         
 "baseball bat" 
 ⋮              
 "umbrella"     
 "van"          
 "vase"         
 "violin"       
 "walkie talkie"
 "wheel"        
 "wheelbarrow"  
 "windmill"     
 "wine-bottle"  
 "wineglass"    
 "wrist-watch"  
 "zebra"        

In [5]:

    

x = TensorFlow.placeholder(Float32)
y_ = TensorFlow.placeholder(Float32)
W = get_variable("weights", [77841, 250], Float32)
b = get_variable("bias", [250], Float32)


    

    
Out[5]:





TensorFlow.Variable(<Operation 'bias'>,<Operation 'bias/Assign'>)

In [6]:

    

y = nn.softmax(x*W + b)
cross_entropy = reduce_mean(-reduce_sum(y_ .* log(y), reduction_indices=[2]))
train_step = train.minimize(train.GradientDescentOptimizer(.00001), cross_entropy)


    

    




W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.






    
Out[6]:





<Tensor NoOp:1 shape=unknown dtype=?>

In [7]:

    

saver = train.Saver()
gpu_options = TensorFlow.tensorflow.GPUOptions(allow_growth=true, per_process_gpu_memory_fraction=0.4)
config = TensorFlow.tensorflow.ConfigProto(log_device_placement=false, gpu_options=gpu_options)
# println(config)
sess = Session(config=config)
train.restore(saver, sess, "4percAcc-allCats")


    

    




I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:910] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
I tensorflow/core/common_runtime/gpu/gpu_device.cc:885] Found device 0 with properties: 
name: GeForce GTX 1070
major: 6 minor: 1 memoryClockRate (GHz) 1.683
pciBusID 0000:01:00.0
Total memory: 7.92GiB
Free memory: 7.64GiB
I tensorflow/core/common_runtime/gpu/gpu_device.cc:906] DMA: 0 
I tensorflow/core/common_runtime/gpu/gpu_device.cc:916] 0:   Y 
I tensorflow/core/common_runtime/gpu/gpu_device.cc:975] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0)

In [8]:

    

type DataLoader
    cur_id::Int
    order::Vector{Int}
end

TrainLoader() = DataLoader(1, shuffle(1:17500))
TestLoader() = DataLoader(1, shuffle(1:2500))


    

    
Out[8]:





TestLoader (generic function with 1 method)

In [9]:

    

function get_test(loader::DataLoader)
    x = zeros(Float32, 2500, 77841)
    y = zeros(Float32, 2500, 250)
    for i = 1:2500
        path = test_paths[loader.order[loader.cur_id]][1:end-1]
        sketch_name = split(path, "/")[1]
        img = load(data_set_folder * path)
        img = restrict(restrict(img))
        flatten_img = reshape(img, 77841)
        
        x[i, :] = flatten_img
        label = find((x -> x == sketch_name), labels)[1]
        y[i, label] = 1.0
        loader.cur_id += 1
        if loader.cur_id > 2500
            println("========================")
            println(sketch_name)
            break
        end
    end
    x, y
end


    

    
Out[9]:





get_test (generic function with 1 method)

In [10]:

    

correct_prediction = indmax(y, 2) .== indmax(y_, 2)
accuracy=reduce_mean(cast(correct_prediction, Float64))
test_loader = TestLoader()
testx, testy = get_test(test_loader)

println(run(sess, indmax(y, 2), Dict(x=>testx, y_=>testy)))
println(run(sess, indmax(y_, 2), Dict(x=>testx, y_=>testy)))
println(run(sess, accuracy, Dict(x=>testx, y_=>testy)))


    

    




========================
head
[168,65,247,32,65,37,28,231,57,190,48,179,53,200,169,60,193,224,223,135,212,193,57,48,84,26,108,67,109,168,62,10,187,126,200,230,190,112,84,58,53,3,247,39,209,32,58,224,26,177,6,48,48,213,76,187,128,233,3,193,190,20,26,36,137,126,209,39,249,20,65,181,193,169,45,63,115,173,26,225,28,20,134,96,212,109,115,119,205,36,58,67,74,168,181,173,33,198,9,83,209,5,25,53,182,109,86,97,65,190,140,39,135,146,207,48,11,60,229,200,109,237,156,241,31,169,115,247,141,246,193,58,193,124,39,149,168,193,152,173,229,109,192,200,55,244,32,53,112,6,191,45,10,32,20,214,233,11,58,26,73,137,66,124,201,149,173,211,26,180,28,7,83,193,17,32,172,33,60,139,5,20,60,190,112,212,229,191,97,229,57,54,106,206,224,168,247,181,190,20,181,82,249,83,228,200,53,244,247,249,133,173,149,128,229,20,179,71,48,229,109,37,37,26,84,84,56,208,67,210,66,7,6,249,139,139,180,115,229,115,72,190,120,66,28,240,84,109,32,209,60,141,128,244,109,67,246,229,97,180,39,73,212,190,181,210,180,45,97,57,126,120,181,190,48,57,6,36,207,83,32,48,119,57,168,145,191,58,181,48,128,168,140,237,190,39,63,168,149,173,129,31,229,45,86,200,53,218,26,112,191,190,36,141,215,244,48,42,119,233,57,181,192,58,39,181,92,187,240,181,169,28,225,244,156,26,190,190,120,168,141,181,135,61,34,83,71,168,39,58,109,249,48,137,109,200,146,168,60,0,169,190,169,247,229,168,182,177,223,221,229,192,73,224,212,198,179,33,84,168,224,48,243,84,5,180,192,48,173,28,71,39,148,200,148,102,210,192,201,11,5,180,11,168,127,98,76,7,193,168,32,31,173,26,71,206,190,87,182,168,71,109,168,237,57,224,228,247,212,208,240,126,189,233,57,39,26,172,26,207,106,17,97,33,168,162,190,190,168,233,208,109,191,55,247,237,190,90,210,10,229,212,48,67,198,141,213,240,26,172,210,180,137,192,209,1,230,221,229,36,229,156,20,229,5,26,80,126,48,168,76,58,140,10,191,48,244,37,187,210,193,140,5,233,200,182,83,212,140,244,80,31,149,84,127,48,182,229,152,28,115,216,26,72,115,229,221,237,106,240,72,137,244,211,83,20,90,58,36,208,182,83,53,168,191,182,109,172,32,37,26,55,200,151,58,137,134,151,32,124,149,26,244,53,200,158,28,201,20,61,88,141,57,60,181,145,65,109,133,84,201,246,66,37,124,0,16,53,80,25,139,128,172,201,181,200,32,55,53,149,137,229,7,169,5,247,82,168,180,214,53,58,190,72,209,48,133,207,5,48,244,224,192,71,215,109,187,53,192,210,29,187,211,10,109,48,168,48,209,7,39,48,32,83,60,212,8,26,60,247,200,242,58,126,133,129,48,152,210,73,71,193,48,241,32,168,20,32,92,83,119,198,168,221,33,7,215,149,200,180,32,124,128,168,200,124,56,172,211,230,243,36,96,242,26,173,149,32,215,32,58,5,113,201,126,180,32,221,58,108,127,66,210,212,58,210,25,94,200,181,48,152,115,71,137,201,88,192,11,187,32,191,58,193,193,229,53,240,48,210,193,213,58,26,28,97,41,55,148,179,25,191,3,20,88,5,38,242,109,233,145,244,24,48,246,39,36,241,173,181,120,126,140,195,191,60,244,56,32,246,32,168,222,33,140,97,145,190,6,58,31,48,28,96,233,181,71,181,97,53,48,48,198,113,109,20,187,73,200,190,25,237,191,212,109,84,216,169,173,208,98,124,149,55,133,86,249,229,48,5,200,57,60,180,42,31,246,36,120,213,203,26,230,210,26,216,48,65,32,148,173,182,209,72,25,200,20,195,229,137,244,182,211,31,200,181,80,56,33,137,39,200,5,201,53,0,26,201,26,28,111,168,84,186,66,169,53,109,149,209,145,48,20,129,156,222,112,237,92,20,26,97,224,65,188,212,247,60,119,65,119,120,141,126,192,210,193,148,31,181,173,58,57,247,109,146,247,161,57,109,65,180,137,148,228,169,84,219,129,58,207,85,181,120,137,66,96,48,172,112,223,211,36,80,227,57,187,9,5,66,193,20,211,240,135,26,57,148,48,92,209,83,20,57,169,174,172,20,20,26,5,224,73,5,191,11,106,32,128,139,58,247,115,174,72,244,244,229,83,190,124,109,53,26,9,115,32,145,20,246,57,168,39,39,229,7,247,57,82,190,181,53,229,3,6,45,240,149,181,182,67,20,179,20,173,148,48,65,53,192,201,174,48,224,224,229,137,72,173,65,208,60,53,39,149,65,209,223,210,112,57,149,58,146,28,57,190,209,58,146,129,10,181,210,55,58,84,60,66,69,182,48,173,6,11,141,124,229,126,243,109,58,193,135,39,115,36,159,61,73,193,207,56,249,173,120,221,181,58,168,181,181,83,182,229,127,135,128,120,39,172,39,210,180,168,151,244,58,209,28,33,120,48,168,192,9,122,240,201,26,229,127,191,99,145,58,80,82,180,210,244,156,229,208,179,209,36,190,200,48,58,28,200,48,84,48,20,181,244,48,246,240,229,209,193,139,181,48,65,109,8,16,172,149,168,168,84,203,201,48,32,39,187,20,246,65,132,28,62,115,161,84,168,200,168,172,173,53,65,247,106,191,63,242,193,124,181,115,97,180,247,244,177,193,229,36,192,200,139,173,20,141,26,210,32,200,123,11,246,28,181,181,60,152,97,247,84,56,73,94,169,65,48,3,20,233,168,84,156,57,39,48,249,244,173,39,26,173,187,48,48,65,28,48,140,215,126,36,25,208,6,36,26,146,191,223,56,233,191,48,55,190,209,57,72,57,31,63,109,5,58,94,126,97,33,247,148,229,187,216,169,12,210,127,173,83,16,195,55,249,45,212,209,247,108,48,25,28,48,57,115,92,26,201,191,224,126,140,65,207,124,28,80,65,210,169,109,41,127,65,168,141,39,94,188,83,0,191,31,124,20,161,20,124,208,168,174,210,191,193,72,191,181,65,181,28,244,156,244,140,210,48,190,32,109,208,26,137,84,229,159,122,96,82,32,215,190,26,237,149,55,201,180,60,58,113,181,229,57,48,210,98,84,57,11,216,210,212,63,32,32,208,247,128,39,180,201,246,20,90,83,209,200,224,246,83,5,107,7,26,200,207,237,115,244,152,36,115,184,173,58,115,25,173,168,53,181,146,149,195,187,3,173,48,48,240,84,168,135,57,198,172,211,65,246,198,128,48,88,26,26,84,180,82,168,173,209,124,36,128,180,26,60,193,120,187,32,25,48,212,57,190,159,58,65,193,67,193,31,243,209,211,96,60,206,115,230,193,28,179,109,39,17,200,181,193,33,126,180,66,135,31,210,168,72,65,139,112,229,244,48,169,137,181,115,32,209,146,12,128,124,244,195,192,11,200,200,11,249,71,98,84,190,229,173,32,200,156,56,20,5,181,109,58,229,54,126,32,61,57,192,246,5,32,48,210,3,82,5,169,32,94,210,201,172,246,48,216,17,63,83,124,210,61,187,58,215,25,193,209,20,191,135,25,32,32,192,191,210,151,153,190,187,20,181,201,109,231,126,84,28,192,115,115,26,187,20,48,168,39,5,136,5,173,174,139,33,198,98,31,102,48,182,129,201,181,139,32,222,173,61,82,63,168,206,48,168,61,97,206,62,210,137,179,137,190,120,181,191,84,37,67,28,244,200,26,229,191,219,201,53,190,60,109,229,140,57,71,60,61,237,168,247,233,31,58,237,222,66,120,181,237,192,229,60,53,200,36,32,139,187,244,84,223,58,48,182,200,65,72,28,57,181,58,180,124,172,97,193,124,229,244,180,48,84,200,242,247,201,20,169,36,187,209,210,26,38,230,247,48,28,140,106,26,181,208,9,191,174,55,56,169,124,137,26,17,72,173,229,242,72,148,221,32,124,240,26,229,28,190,210,149,5,181,10,61,146,10,237,109,106,32,233,129,210,48,139,7,55,28,76,90,206,230,57,60,140,10,246,16,224,60,201,206,26,83,84,212,72,96,203,193,5,73,39,173,146,17,109,33,84,201,146,106,28,229,130,102,55,200,169,169,106,141,193,169,210,168,26,181,115,26,58,124,190,20,48,230,28,182,65,211,58,5,98,240,229,99,244,109,48,210,66,6,191,32,249,109,200,45,156,187,191,65,57,5,209,20,206,200,32,83,115,0,210,206,26,60,180,53,55,58,192,57,102,73,72,55,32,242,123,20,224,141,48,214,10,58,237,42,141,66,26,168,83,57,190,37,48,181,209,112,83,223,212,128,58,28,73,26,145,229,229,191,188,229,146,149,187,32,28,112,20,193,141,149,124,28,65,149,71,240,32,210,115,181,128,182,109,72,48,168,201,109,224,48,249,126,221,97,26,53,206,57,168,55,198,207,135,28,216,212,115,72,57,58,20,32,192,210,200,145,244,209,31,98,53,48,247,48,221,145,9,55,20,73,140,135,206,211,193,112,109,247,73,223,212,135,115,145,5,149,67,198,249,244,32,124,57,187,156,39,5,17,53,26,58,146,235,63,229,48,229,20,82,71,98,168,224,32,193,60,58,94,113,244,139,28,31,60,76,139,97,48,65,229,221,109,130,26,187,90,26,8,63,168,32,11,244,211,229,180,229,169,246,48,26,5,126,8,32,139,130,141,48,109,193,20,244,84,146,212,229,190,7,209,48,209,39,109,48,71,11,39,140,182,146,237,109,58,224,192,190,57,37,82,168,5,32,48,7,191,7,28,28,229,190,9,139,58,115,115,221,244,209,181,200,146,228,215,26,48,224,191,211,26,148,107,20,76,33,57,126,26,58,190,32,128,210,190,55,36,233,124,10,20,48,26,246,209,56,32,156,84,200,83,198,32,48,243,249,69,20,98,72,58,48,242,54,48,109,159,99,108,71,212,209,11,212,82,181,221,193,215,180,56,207,246,55,17,26,190,37,66,109,206,168,48,181,133,191,213,66,247,83,28,26,244,223,212,83,173,230,191,249,60,48,145,140,55,42,129,10,137,214,48,129,149,222,20,20,57,60,60,33,39,156,8,126,221,26,26,26,146,80,39,244,120,97,149,211,28,55,36,249,193,244,36,224,113,48,25,48,58,82,126,193,180,181,109,32,11,180,126,180,190,247,215,36,25,233,115,247,5,96,25,84,72,3,124,115,32,168,3,28,83,26,222,32,247,41,210,26,32,180,32,84,222,48,82,25,84,84,201,173,224,216,209,188,39,20,173,106,98,229,57,11,249,65,86,241,190,128,55,126,57,20,48,233,216,190,209,187,57,72,200,67,169,193,209,136,211,55,65,141,63,94,210,106,115,115,200,181,169,20,58,198,200,115,36,141,7,246,115,180,7,58,146,187,169,161,5,201,31,130,207,173,129,173,224,181,182,61,97,88,146,17,206,229,128,5,168,241,209,5,168,97,58,55,53,210,200,168,190,102,20,212,90,48,151,247,240,28,122,127,124,115,137,58,12,115,233,98,243,60,119,25,128,20,20,228,55,181,67,209,233,198,210,135,109]
[32,230,60,28,248,218,188,44,140,117,181,51,66,107,42,103,31,226,59,27,0,162,133,195,12,148,241,235,100,108,113,13,109,32,84,242,70,99,167,117,66,11,74,166,218,113,158,164,137,159,216,52,28,117,146,190,43,76,63,232,150,54,244,114,162,108,200,175,105,233,2,81,134,155,42,39,56,206,21,35,202,146,23,18,69,153,135,113,37,126,53,139,56,213,103,85,173,24,65,151,1,229,13,130,181,55,101,247,88,9,157,211,171,50,43,116,239,176,82,92,247,64,121,242,69,183,209,3,43,161,49,11,32,226,236,196,7,40,221,185,36,117,111,136,4,31,189,57,247,28,221,233,169,34,52,95,222,155,3,105,204,33,216,22,185,96,151,240,226,134,193,21,9,17,243,58,81,49,65,189,164,42,12,48,100,226,43,22,210,181,97,243,200,27,47,155,134,76,190,11,203,219,71,149,190,72,33,32,79,212,138,36,180,8,160,236,237,61,158,64,74,74,29,26,162,131,1,88,195,44,210,26,218,24,216,83,56,179,204,207,191,43,72,67,36,216,197,53,41,72,39,63,238,48,10,157,214,202,151,157,87,58,105,198,90,53,102,58,95,45,102,143,87,79,0,27,16,220,237,118,36,10,41,225,107,162,213,84,46,234,220,185,111,42,108,224,185,54,75,2,166,6,188,195,80,57,40,75,137,110,111,119,230,102,18,153,5,217,29,75,210,103,73,51,58,199,106,129,45,223,241,151,34,168,219,26,42,4,223,12,127,205,115,192,71,80,79,143,237,243,96,28,217,41,208,180,191,98,88,30,142,48,69,112,80,248,131,124,19,41,90,152,14,242,113,31,159,141,156,166,45,225,74,201,168,18,79,85,8,119,59,131,173,19,217,239,224,32,63,194,205,204,128,151,44,50,193,93,86,20,113,177,231,8,73,35,103,116,198,239,91,67,150,158,234,24,127,88,137,148,82,145,78,140,133,78,215,107,132,208,95,134,201,172,141,59,9,134,101,88,122,57,227,87,122,247,249,240,129,66,203,96,185,25,31,181,82,65,18,125,57,233,6,61,230,224,18,81,78,191,126,109,129,211,127,131,236,170,235,119,150,16,58,71,225,206,24,136,197,85,3,99,76,97,241,149,83,6,206,193,54,189,235,224,168,153,173,234,45,211,78,130,41,154,73,95,179,208,100,45,34,183,158,199,116,40,106,22,34,175,1,130,20,192,102,60,216,172,168,118,153,105,218,91,199,13,104,238,177,62,249,20,147,191,167,128,231,163,243,124,0,66,165,6,228,34,60,154,11,197,248,183,234,230,199,53,152,84,215,152,15,31,221,121,128,28,228,66,163,68,50,178,55,49,223,127,240,26,30,11,173,32,189,142,205,61,158,156,164,161,46,110,92,21,6,106,78,26,207,163,99,87,93,161,44,10,86,201,159,7,245,149,178,225,212,120,203,32,144,246,187,138,85,179,242,8,33,65,57,139,101,40,45,11,20,111,125,120,174,149,188,46,112,143,244,147,111,156,14,76,33,36,16,150,86,16,0,27,233,211,113,9,226,18,179,202,83,73,202,24,36,10,173,68,66,91,91,193,214,155,16,204,249,7,246,12,59,75,138,190,167,59,204,234,209,4,245,129,194,106,119,105,129,134,55,94,189,30,59,105,227,21,109,161,166,123,75,106,135,19,25,104,93,2,186,68,178,142,14,240,165,49,209,119,215,85,239,127,130,242,244,181,66,207,72,64,91,41,77,12,198,223,210,232,93,186,11,237,23,129,50,129,88,153,211,184,229,130,160,115,111,46,38,235,97,238,206,225,55,204,38,74,126,133,197,35,209,110,64,101,65,228,228,159,17,114,232,95,104,34,50,221,26,163,90,52,104,165,71,165,99,205,143,80,94,92,43,199,171,125,99,72,37,161,228,25,30,159,248,108,237,34,215,126,125,62,158,0,118,239,38,185,21,127,48,10,15,67,25,73,38,108,157,223,202,4,1,120,202,180,238,228,241,109,144,194,112,240,113,170,28,148,215,83,166,174,67,75,237,218,155,193,214,56,174,170,119,77,50,188,50,4,146,5,129,102,193,183,132,79,14,66,45,14,187,152,219,118,186,136,225,170,220,173,249,53,200,200,12,207,186,116,152,0,118,140,4,39,9,235,39,61,17,25,144,166,160,222,224,17,54,100,245,26,243,151,247,100,194,237,171,82,189,47,213,43,29,144,146,157,75,124,192,29,239,207,69,196,186,108,177,44,107,64,2,128,125,89,86,199,87,118,102,213,37,49,147,118,129,197,7,115,225,110,235,249,182,25,31,193,229,80,185,213,68,212,37,143,169,141,197,216,115,162,123,87,205,210,71,126,221,49,238,99,225,94,155,83,223,127,231,78,76,39,107,37,70,236,160,219,13,56,167,91,100,162,29,237,183,207,85,40,141,110,195,86,5,144,139,175,112,12,140,93,91,100,8,47,156,67,45,57,248,61,158,130,23,164,130,58,9,5,74,89,97,104,175,126,103,26,231,198,98,204,241,31,228,132,52,123,143,142,11,176,145,230,109,120,168,165,119,222,141,25,172,139,226,120,35,248,86,50,148,94,66,191,94,168,95,50,86,34,237,62,246,136,143,198,233,169,190,220,97,244,106,211,187,136,55,242,48,151,65,96,79,212,162,246,209,49,213,102,172,72,114,14,194,145,43,112,8,186,120,212,89,141,79,16,214,13,20,85,0,60,209,48,181,0,122,41,17,165,147,46,172,162,112,63,40,42,94,86,9,165,201,74,14,109,17,67,137,203,104,108,169,156,0,30,196,76,2,127,238,161,154,180,201,216,175,47,35,155,131,42,125,89,185,125,64,211,15,191,70,198,157,244,98,27,78,179,237,245,76,73,8,80,41,24,233,51,48,185,53,195,69,203,147,219,42,222,133,136,190,221,126,244,140,80,194,91,69,184,46,117,226,59,81,84,188,212,12,47,60,128,51,198,51,95,16,93,178,142,211,129,247,141,224,245,154,231,137,84,233,187,183,86,2,164,156,101,54,135,229,153,80,230,188,230,192,19,194,200,144,174,177,173,60,147,174,142,167,220,99,71,224,232,16,174,230,76,47,179,1,192,221,160,217,234,107,41,93,114,162,164,87,206,7,204,121,70,201,135,119,170,7,193,65,206,25,121,191,35,2,232,126,11,35,132,182,105,41,62,196,77,89,191,155,90,134,123,84,227,89,192,77,163,116,24,208,137,222,74,21,187,211,6,128,29,101,215,37,180,152,160,203,197,217,10,55,200,75,231,177,12,23,227,81,246,5,62,152,75,244,85,92,107,195,188,49,236,123,133,158,175,145,6,248,114,135,29,136,168,210,210,169,241,181,169,8,241,55,23,4,212,16,217,100,186,193,38,112,231,63,218,6,139,113,102,193,91,82,117,233,96,171,29,228,80,236,178,110,186,51,154,160,139,86,247,47,40,22,117,176,248,209,239,18,192,52,246,132,82,210,25,201,155,135,135,227,44,238,187,102,98,22,114,176,145,27,115,159,184,220,248,243,90,137,63,120,103,172,167,30,54,177,171,206,238,96,114,32,215,57,155,124,97,144,132,61,176,190,122,27,148,139,194,198,79,180,108,124,209,70,242,110,208,117,132,215,163,197,209,115,244,97,159,53,96,47,248,242,245,105,218,147,142,208,85,166,72,62,94,116,92,183,131,236,152,10,64,97,199,243,183,168,139,218,98,13,213,48,17,149,183,216,39,10,58,168,1,127,63,84,42,175,38,79,201,31,195,229,165,48,99,79,20,44,164,243,68,116,39,217,106,235,124,147,15,140,172,73,134,88,98,58,214,84,35,72,19,236,167,204,213,191,119,13,117,249,80,205,44,3,217,224,0,109,175,2,182,10,20,246,88,83,207,140,226,83,152,235,221,124,46,175,150,149,37,199,55,182,49,56,234,5,116,32,89,36,140,124,9,232,82,222,232,125,95,231,198,246,196,27,229,232,114,81,215,182,3,249,128,196,25,70,230,202,221,141,225,98,102,149,14,147,69,38,177,153,153,154,180,36,57,84,202,205,181,224,78,142,229,131,196,122,132,57,137,184,229,51,232,215,154,67,195,159,210,182,236,105,81,63,63,138,39,220,37,24,69,36,16,172,44,71,234,160,51,111,34,238,146,92,58,145,213,30,120,220,171,76,106,89,43,196,207,227,64,212,77,121,192,158,33,206,61,213,130,66,171,60,222,53,4,46,123,223,100,22,47,46,24,115,55,81,144,232,40,121,167,170,114,219,230,118,139,61,40,122,94,90,72,138,203,50,56,204,44,145,15,18,103,98,182,236,90,148,150,227,163,138,27,19,143,179,112,161,123,225,166,186,121,19,110,12,150,114,109,197,169,39,244,106,233,123,87,103,200,234,122,151,229,68,43,1,77,3,40,207,138,172,157,222,8,81,7,1,149,20,210,104,124,208,9,249,70,178,88,15,6,36,13,163,249,90,26,60,126,73,173,92,60,74,182,249,246,176,56,180,150,112,37,33,131,223,234,178,165,17,1,214,205,154,133,184,15,59,145,28,111,104,17,23,194,3,108,23,54,222,68,146,90,164,82,169,7,130,28,240,101,241,63,29,242,54,67,174,105,141,71,88,120,187,30,206,214,21,219,38,184,189,214,245,228,78,69,161,113,68,228,115,126,151,133,94,212,136,81,157,107,171,173,96,172,23,184,178,3,177,70,120,184,52,84,227,184,239,182,101,121,67,107,205,195,112,70,38,55,239,124,4,156,92,158,171,33,76,223,149,138,53,128,133,163,111,240,17,192,45,181,85,62,52,122,240,59,219,207,133,13,170,118,159,15,138,122,182,197,121,189,65,22,87,33,154,19,51,203,83,161,226,30,164,3,220,7,5,96,19,39,121,47,151,137,3,18,201,150,148,195,160,167,110,62,135,180,183,30,218,1,54,181,33,122,2,49,35,20,173,174,15,208,34,128,146,200,165,93,14,161,83,180,21,113,227,65,77,127,145,174,132,227,103,176,31,57,78,156,242,77,148,240,48,4,29,35,75,54,239,56,190,21,115,61,209,73,135,60,171,190,238,187,224,46,247,22,77,142,108,116,128,144,82,71,123,177,184,186,99,103,116,67,140,146,99,189,111,170,104,22,8,26,217,143,71,119,97,178,20,98,146,137,187,9,10,156,176,214,5,223,70,31,153,130,14,160,72,125,117,136,194,163,157,189,231,132,233,191,226,95,178,73,45,235,205,150,37,241,200,87,166,53,166,69,235,208,167,95,202,148,203,201,134,68,22,185,176,143,222,5,148,74,106,59,217,139,162,64,19,38,91,208,219,6,243,220,168,188,179,18,61,177,170,134,90,24,82,64,13,170,28,68,109,240,11,23,123,89,21,154,42,65,56,179,218,62,198,190,92,98,174,109,169,247,169,89,245,140,211,144,221,229,28,51,104,23,135,199,188,246,175,156,214,97,110,203,58,7,247,33,146,145,115,83,206,196,93,131,125,179,241,212,138,96,101,100,141,32,192,159,27,147,118,243,216,153,199,107,5,200,244,176,231,142,216,164,196,2,136,62,188,52,92,52,15,131,187,245,77,133,219,157,245,202,149,93,152,94,52,101]
0.004400000000000001

In [11]:

    

function sketch_input(path, sketch_name)
    x = zeros(Float32, 1, 77841)
    y = zeros(Float32, 1, 250)
    img = load(path)
    img = restrict(restrict(img))
    flatten_img = reshape(img, 77841)
    x[1, :] = flatten_img
    
    label = find((x -> x == sketch_name), labels)[1]
    y[1, label] = 1.0
    
    x, y
end


    

    
Out[11]:





sketch_input (generic function with 1 method)

In [12]:

    

sketch_path = "/raided/datasets/sketches_png/png/bee/1371.png"
sketch_label = "bee"

test_img, test_label = sketch_input(sketch_path, sketch_label)
what_it_thinks = labels[run(sess, indmax(y,2), Dict(x=>test_img, y_=>test_label))[1] + 1]
println("Is it an " * what_it_thinks * "? Answer: ", (sketch_label == what_it_thinks), ".")


    

    




Is it an duck? Answer: false.

In [ ]: