Thursday, September 08, 2011

notMNIST dataset

I've taken some publicly available fonts and extracted glyphs from them to make a dataset similar to MNIST. There are 10 classes, with letters A-J taken from different fonts. Here are some examples of letter "A" Judging by the examples, one would expect this to be a harder task than MNIST. This seems to be the case -- logistic regression on top of stacked auto-encoder with fine-tuning gets about 89% accuracy whereas same approach gives got 98% on MNIST. Dataset consists of small hand-cleaned part, about 19k instances, and large uncleaned dataset, 500k instances. Two parts have approximately 0.5% and 6.5% label error rate. I got this by looking through glyphs and counting how often my guess of the letter didn't match it's unicode value in the font file. Matlab version of the dataset (.mat file) can be accessed as follows:
load('notMNIST_small.mat')
for i=1:5
    figure('Name',num2str(labels(i))),imshow(images(:,:,i)/255)
end
Zipped version is just a set of png images grouped by class. You can turn zipped version of dataset into Matlab version as follows
tar -xzf notMNIST_large.tar.gz
python matlab_convert.py notMNIST_large notMNIST_large.mat
Approaching 0.5% error rate on notMNIST_small would be very impressive. If you run your algorithm on this dataset, please let me know your results.

539 comments:

  1. Anonymous11:37 PM

    What does your baseline get on the negated version of the dataset? In other words, make the "ink" pixels have intensity 1 and the non-ink pixels have intensity zero. I would be curious to know if your baseline does better on one version or the other.

    ReplyDelete
  2. I don't expect it to make any difference -- pixel level features are learned by stacked autoencoder, and there's nothing biasing to learner to prefer 0's or 1's to start with

    ReplyDelete
  3. Anonymous7:09 PM

    It makes a difference on MNIST, which is why I asked.

    ReplyDelete
  4. Anonymous8:21 AM

    WE can use this data even if we do research on this?for instance if we obtain relatively good results
    or propose something novel are we allowed to publish anything on it?
    regards
    ML_random_guy

    ReplyDelete
  5. ML_random_guy -- that depends on whether your country has laws against publishing

    ReplyDelete
  6. Anonymous12:47 AM

    Hello, it's nice to have such a new challenging dataset. Do you recommend a specific evaluation protocol (number of training/test images) ? Otherwise people will work on different subsets and results will not be directly comparable.

    ReplyDelete
  7. Train on the whole "dirty" dataset, evaluate on the whole "clean" dataset.

    This is a better indicator of real-life performance of a system than traditional 60/30 split because there is often a ton of low-quality ground truth and small amount of high quality ground truth. For this task, I can get millions, possibly billions of distinct digital glyph images with 5-10% labels wrong, but I'm stuck with small amount of near perfectly labeled glyphs

    ReplyDelete
  8. Anonymous3:25 AM

    Thanks for the protocol info. Would it be possible to get a tar archive with PNG images of the small dataset like the huge one ? I'm not using Matlab.

    ReplyDelete
  9. Oops, small tar should've been in the directory to start with, fixed

    ReplyDelete
  10. I used this dataset to test some of my code and got about 3.8% error rate. Are there more results known for this dataset? A few lines of text are here.

    ReplyDelete
  11. Hey, that's pretty impressive! This is the highest accuracy I know. I'm working on a larger dataset to release publicly, but slowed down by some legal clearance hurdles

    ReplyDelete
  12. How do you do finetuning? Hinton's contrastive wake-sleep?

    ReplyDelete
  13. What is unicode370k.tar.gz?

    ReplyDelete
  14. It's a bunch of characters taken from the tail end of unicode values

    http://yaroslavvb.blogspot.com/2011/11/shapecatcher.html

    ReplyDelete
  15. How did you split your dataset into train,valid,test to get 89%?

    ReplyDelete
  16. Hi, myself and Zhen Zhou from the LISA lab at Université de Montréal trained a couple of 4 layer MLPs with 1024-300-50 hidden neurons respectively. We divided the noisy set into 5/6 train 1/6 valid and kept the clean set for testing. We 97.1% accuracy on the test set at 412 epoch with early stopping, linear decay of the learning rate, a hard constraint on the norm of the weights and tanh activation units. We get approximately 93 on valid and 98 on train. The train set is easy to overfit (you can get 100% accuracy on train if you continue training). One could probably do better if they pursue hyper-optimization further. We used Torch 7.

    ReplyDelete
  17. I got with a simple neural network (784,1024,10), whereas the activation functions where RELU and then just a normal softmax. Without activation decay, pre stop, dropout & co and 3001 iterations and a batch size of 128, I got 89.3% accuracy on the test set.

    Step: 3000
    Minibatch accuracy: 86.7%
    Validation accuracy: 82.6%

    Finish (after Step 3001):
    Test accuracy: 89.3%

    ReplyDelete
  18. Minibatch loss at step 3000: 55.872269
    Minibatch accuracy: 79.7%
    Validation accuracy: 84.4%
    Test accuracy: 90.6%

    With a neural network with a single hidden layer (1024 nodes), Relu and l2 regularization.

    ReplyDelete
  19. Minibatch loss at step 10000: 123.963661
    Minibatch accuracy: 45.3%
    Validation accuracy: 85.3%
    Test accuracy: 91.4%

    With a dropout and relu and l2 regularizer, single hidden layer 1024 node.

    ReplyDelete
  20. Yaroslav Bulatov,

    Thank you for the fun and challenging dataset.

    How were the names of the files chosen?

    I'm working on renaming each one to the phash value of the image. It looks like the names might already be the result of a hash.

    ReplyDelete
  21. This comment has been removed by the author.

    ReplyDelete
  22. This comment has been removed by the author.

    ReplyDelete
  23. Check out the Udacity course in deep learning, made by Google. They use this dataset extensively and show some really powerful techniques. The goal of the last assignment was to experiment with this techniques to find the best accuracy using a regular multi-layer perceptron. I have a pretty beefy machine: 6600K OC, 2x GTX 970 OC, 16gb DDR4, Samsung 950 Pro; so I set up a decent sized network and let it train for a while.

    My best network gets:

    Test accuracy: 97.4%
    Validation accuracy: 91.9%
    Minibatch accuracy: 97.9%

    First I applied a Phash to every image and removed any with direct collisions. Then I split the large folder into ~320k training and ~80k validation. I used ~17k in the small folder for testing. Trained on mini-batches using SGD on the cross-entropy, dropout between each layer and an exponentially decaying learning rate. The network has three hidden layers with RELU units, plus a standard softmax output layer.

    Here are the parameters:
    Mini-batch size: 1024
    Hidden layer 1 size: 4096
    Hidden layer 2 size: 2048
    Hidden layer 3 size: 1024
    Initial learning rate: 0.1
    Dropout probability: 0.5

    I ran this for 150k iterations, took an hour and half using one GPU. Learning pretty much stopped at 60k, but the model never began to overfit. I believe that is because the dataset is so large and the dropout. Even at the end of all that training with a good size network the mini-batch accuracy still did not reach 100% so learning could continue, albeit slowly.

    The next assignment is to use a convolutional network, which looks promising. I'll try to post those results too.

    ReplyDelete
  24. Anonymous6:39 AM

    Could you make your code available? Or at least say which parameters you have use to the exponentially decaying learning rate? Did you use l2 regularization (if yes, with which regularization factor?) I tried to use the same network as you did and it simply doesn't converge.

    ReplyDelete
  25. Test accuracy: 98.09%

    With a CNN layout as follows:

    3 x convolutional (3x3)
    max pooling (2x2)
    dropout (0.25)

    3 x convolutional (3x3)
    max pooling (2x2)
    dropout (0.25)

    dense (4*N)
    dropout (0.5)
    dense (2*N)
    dropout (0.5)
    dense (N)
    dropout (0.5)
    softmax (10)

    N is the number of pixel in the images. All layers use relu activation. I also used some zero padding before each convolutional layer. The network was trained with Adadelta. It took ~45 iterations with an early stopping at patience 10. As a final step I ran SGD with the same early stopping and decaying learning rate starting at 0.1. It ran about 15 iterations. Evaluating the network on the training set, the accuracy was 99.07% and 94.25% on the validation set.

    ReplyDelete
  26. This comment has been removed by the author.

    ReplyDelete
  27. Minibatch loss at step 4999: 0.901939
    Minibatch accuracy: 75.0%
    Validation accuracy: 87.3%
    Test accuracy: 93.3% @step=4999
    Model saved in file: save/myconvnet_5000

    I used a architecture similar to LeNet, and it seems to be better as step get larger.

    ReplyDelete
  28. Where can I download notMNIST? The link above goes to an account that has been suspended.

    ReplyDelete
  29. Not sure if this is the complete dataset, but the Udacity course on Deep Learning using notMNIST provides the following links:

    http://commondatastorage.googleapis.com/books1000/notMNIST_large.tar.gz
    http://commondatastorage.googleapis.com/books1000/notMNIST_small.tar.gz

    ReplyDelete
  30. Test accuracy: 96.98%

    With a CNN layout with following configurations, which is similar to [LeNet5](http://culurciello.github.io/tech/2016/06/04/nets.html)
    However there is little difference

    convolutional (3x3x8)
    max pooling (2x2)
    dropout (0.7)
    relu

    convolutional (3x3x16)
    max pooling (2x2)
    dropout (0.7)
    relu

    convolutional (3x3x32)
    avg pooling (2x2): according to above article
    dropout (0.7)
    relu

    fully-connected layer (265 features)
    relu
    dropout (0.7)

    fully-connected layer (128 features)
    relu
    dropout (0.7)

    softmax (10)

    decaying learning rate starting at 0.1
    batch_size: 128

    Training accuracy: 93.4%
    Validation accuracy: 92.8%

    ReplyDelete
  31. Accuracy: 96.1 without convolution (assignment 3 in TensorFlow course)
    Using Xavier initialization significantly boosted my results. Network specifications:
    1. Batch size = 2048
    2. Hidden units: 4096, 2048, 1024
    3. Adam optimizer with 0.0001 learning rate
    4. Dropout on each hidden layer
    5. Xavier initialization

    ReplyDelete
  32. Hi

    I'm trying to use tensorflow to do character recognition. I am able to use your dataset(A-J) and get some data from char74k dataset (from K to Z) to train character data and predict. but the char74k set is a pretty limited set and is not enough to get a good accuracy. Have you posted anything similar for characters from K to Z?

    ReplyDelete
  33. This comment has been removed by the author.

    ReplyDelete
  34. no convolution, 1 hidden layer 94.4 % with test set

    batch size 128
    L2 regularization beta 0 (no L2 regularization)
    initialize w with deviation 0.03
    initialize bias with all 0
    Learning rate 0.5 (fix, not decay)
    single hidden layer unit # 1024
    dropout_keepratio 1 (no dropout)

    I'm following udacity tutorial.
    It's strange that whenever i put L2 regularization, dropout, Learning rate decay, the test accuracy falls. I can't figure out why.

    ReplyDelete
  35. Anonymous10:06 AM

    The test accuracy will fall if you choose a wrong value of regularization parameters. A beta of .005 gives good results.

    ReplyDelete
  36. Anonymous2:33 PM

    My accuracy falls slightly after using dropout. Is there a possibility of wrong implementation of tf.nn.dropout() or is it a possible scenario?

    ReplyDelete
  37. This comment has been removed by the author.

    ReplyDelete
  38. Multi Layer Neural Net without convolution - Test Accuracy = 94.4%

    Architeture
    3 Layer Neural Network(No convolution) = input-784, hidden-526, output=10
    L2- Regularization with lambda(regularization parameter) = .001
    Number of steps = 3000
    Batch size = 500

    ReplyDelete
  39. 2 hidden Layers ( Toal 4 layers ) - without convolution - Test Accuracy = 95.8 %

    Architecture
    3 Layer Neural Network(No convolution) = input-784, hidden1-960, hidden2=650 output=10
    L2- Regularization with lambda(regularization parameter) = .0005
    Number of steps = 75000
    Batch size = 1000

    ReplyDelete
  40. Minibatch accuracy: 93.2%
    Validation accuracy: 91.2%
    Test accuracy: 96.3%
    After 10000 steps.

    Architecture:
    Two hidden layers:
    num_hidden_nodes = 1024
    num_hidden_nodes_2 = 100

    Both with Relu inputs. Cross entropy + L2 regularization (beta = 1.3e-4).
    SGD, batch size 400.
    Most importantly, weights were initialized with truncated normal distro. with sigma = 0.01.

    Exponential decay starting at 0.5, 0.65 decay_rate every 1000 steps.

    ReplyDelete
  41. Using Keras on an average gaming laptop with moderate GPU, training took less than 2' on the full (udacity) training set of 200.000 samples, using 10.000 validation samples and measuring accuracy on separate test set of 10.000 samples.
    With a simple multilayer network, I reached 96.66%

    With KERAS, the code for the network itself is really simple:

    batch_size = 128
    nb_classes = 10
    nb_epoch = 20

    model = Sequential()

    model.add(Dense(1024, input_shape=(784,)))
    model.add(Activation('relu'))
    model.add(Dropout(0.2))

    model.add(Dense(512))
    model.add(Activation('relu'))
    model.add(Dropout(0.2))

    model.add(Dense(256))
    model.add(Activation('relu'))
    model.add(Dropout(0.2))

    model.add(Dense(10))
    model.add(Activation('softmax'))

    model.summary()

    model.compile(loss='categorical_crossentropy',
    #optimizer=RMSprop(),
    optimizer='adagrad',
    #optimizer='adadelta',
    metrics=['accuracy'])

    history = model.fit(train_dataset, train_labels,
    batch_size=batch_size, nb_epoch=nb_epoch,
    verbose=1, validation_data=(valid_dataset, valid_labels))
    score = model.evaluate(test_dataset, test_labels, verbose=0)

    ReplyDelete
  42. If you want to generate your own dataset like notMNIST, you should try not_notMNIST

    ReplyDelete
  43. This comment has been removed by the author.

    ReplyDelete
  44. This comment has been removed by the author.

    ReplyDelete
  45. My final result is 96.23% accuracy. Network architecture (built with Keras):

    conv(3x3x32)
    maxp(2x2)
    dropout(0.05)

    conv(3x3x16)
    maxp(2x2)
    dropout(0.05)

    dense(128, relu)
    dense(64, relu)
    dense(10, softmax)

    I used SGD with default params. Also got 92.03% on valid dataset, 92.24% on train dataset. Seems that it is global tendency that test score is higher,

    ReplyDelete
  46. This comment has been removed by the author.

    ReplyDelete
  47. 97.2% on a fully connected net.

    At last iteration, 100k:
    Minibatch accuracy: 99.0%
    Validation accuracy: 92.2%
    Test accuracy: 97.2%

    Architecture:
    3 hidden layers, 4096 - 3072 - 1024, with relu and 0.5 dropout
    Xavier weight init
    Batch size 200
    Data sets original (200k train, 10k valid, 10k test), no further preprocessing
    Loss: softmax_cross_entropy_with_logits + L2 regularization on weights with weight of 1e-4
    Learning rate 0.3 with decay of 0.96 every 1000 iterations
    Total 100k iterations
    [edit - I forgot the dropout on first post]

    ReplyDelete
  48. Test accuracy: 96.12% with only 5000 iterations on a convolutional network with two conv layers and a final fully connected layer.
    Minibatch of 50 images was used.

    ReplyDelete
  49. On a very simple 1 hidden layer network without regularization I also get:

    Minibatch accuracy: 89.8%
    Validation accuracy: 82.9%
    Test accuracy: 89.8%

    I've seen many other users reporting Test accuracy which is significantly higher than validation accuracy.
    Validation and Test are the same size in my case. Is a higher test score reasonable or is it just chance? Should I consider the worst between test and validation as the expected performance of my network?

    Gianni

    ReplyDelete
  50. Anonymous1:58 AM

    I use 2 hidden layers and GradientDescentOptimizer, but the loss is nan. Why?

    ReplyDelete
  51. Try to reduce your learning rate

    ReplyDelete
  52. 96.2% on a fully connected net.

    setps, 200000:
    batch 200 accuracy: 94.0%
    test accuracy: 96.2%
    https://github.com/ms03001620/NotMnist

    ReplyDelete
  53. test acc 98.3%
    mini-batch train acc 95.7%
    val acc 94.2%

    Techniques: "shallow" resnet (used val set to select arch), dropout, horizontal + vertical shift data augmentation, reduce lr on plateaus.

    implementation

    ReplyDelete
  54. This comment has been removed by the author.

    ReplyDelete
  55. Test Accuracy 95.5%
    with batch size = 128, number of iterations = 10k
    Three 5x5 convoution layers of depth 16, 32, 64 respectively
    Three hidden layers with number of hidden nodes 256, 128 and 64 respectively
    Dropout 0.7
    Learning decay starting with 0.2 learning rate

    https://sandipanweb.wordpress.com/2017/08/03/deep-learning-with-tensorflow-in-python-2/

    ReplyDelete
  56. This comment has been removed by the author.

    ReplyDelete
  57. Test accuracy: 97.2%

    Implementation:
    2 CNNs with max pooling followed by a 1 layer fully-connected NN:
    Patch size = 7x7
    Stride for CNN = 1
    Size of pooling size = 2x2
    Stride for Pooling = 2
    Depth = 50, 100
    Final layer nodes = 512
    Dropout_keep_probe = 0.7

    ReplyDelete
  58. L Taylor8:52 AM

    Managed to achieve 97.3% test score!

    Used 5 hidden layers, batch_size = 256, adam optimisation with initial learning rate of 1e-4, 200,000 steps. Would be happy to share details, code etc. if anybody is interested.

    ReplyDelete
  59. Test accuracy: 98.0%

    Implementation:
    2 CNNs with max pooling followed by a 1 layer fully-connected NN:
    Patch size = 5x5
    Stride for CNN = 1
    Size of pooling size = 2x2
    Stride for Pooling = 2
    Depth = 50, 100
    Hidden layer Nodes in FCNN = 512
    Dropout_keep_probe = init: 95%, decaying to 70%

    ReplyDelete
  60. Udacity Deep Learning course challenged me to get as high accuracy as I can using only dense layers, without any convolutions.

    09-16 04:03:04.724 assignment_03_regularization.py:470 INFO Train loss: 0.167651, train accuracy: 97.99%
    09-16 04:03:04.725 assignment_03_regularization.py:473 INFO Test loss: 0.256166, TEST ACCURACY: 96.51% BEST ACCURACY 96.64% <<<<<<<

    Managed to achieve only as high as 96.6% with the following model:
    - 5 fully connected layers 2048-1024-1024-1024-512
    - 0.5 dropout
    - batch normalization
    - weight decay with 0.00001 scale
    - batch 128 images
    - Adam optimizer with starting LR=1e-4
    - Xavier weight initialization (this is critical!)

    This is on "sanitized" test dataset, where I removed all images that were identical or close to some images in training data. Without this sanitizing, it would've probably been a bit better.
    Trained that for 2 hours on GTX1060, it continued to climb higher, but slowly.
    Code is here: https://github.com/alex-petrenko/udacity-deep-learning/blob/master/assignment_03_regularization.py
    (function is called train_deeper_better())

    ReplyDelete
  61. Simple convnet achieved 98.19% on test, code here: https://github.com/alex-petrenko/udacity-deep-learning/blob/14714ee4151b798cde0a31a94ac65e08b87d0f65/assignment_04_convolutions.py#L39

    (5,5)->(5,5)->pool->(3,3)->(3,3)->pool->fc1024->fc1024->logits

    INFO Starting new epoch #121!
    INFO Minibatch loss: 0.150696, reg loss: 0.041653, accuracy: 96.88%
    INFO Train loss: 0.068505, train accuracy: 99.28%
    INFO Test loss: 0.118685, TEST ACCURACY: 98.05% BEST ACCURACY 98.19% <<<<<<<

    ReplyDelete
  62. jabdov9:53 AM

    Test accuracy: 96.7%

    Implementation:
    * 2 hidden layers of 1024 & 256 nodes
    * weight initialization using gaussian random distribution with stddev = 2/sqrt(size of layer input)
    * minibatch size of 128
    * 30 epochs -each epoch full pass over train data set, randomly shuffled by minibatches (200000 / 128 = 1563 steps per epoch)
    * learning rate 0.1
    * dropout with keep_prob = 0.9
    * no L2 regularization

    After epoch 30:
    train accuracy = 98.0%
    dev accuracy = 92.0%
    test accuracy = 96.7%

    ReplyDelete
  63. My best straight-forward CNN:
    C5x4-C19x8-C5x16-P2-C7x64-P2-C3x256-P3S-C1x1024-C1x512-F2048-F64-F10
    where
    C5x4 = convolution with 5x5 kernel and 4 maps output
    P3S = pooling with 3x3 size of type SAME

    ReLU
    initial weight SD: 0.05 for conv layers, Xavier for full layers
    max pooling
    full layer dropout 0,6
    conv layer dropout 0,1
    conv layer dropout before pooling
    shuffle train dataset after each epoch
    momentum optimizer with learning rate 0.05
    batch size 2048

    after 470 epochs (early stop):

    train accuracy 99.6%
    validation accuracy 94.6%
    test accuracy 98.2%

    This configuration was evolved looking for hyperparameter optimization, and came up at the 103rd configuration try. Further runs of another 34 configurations did not improve it.

    During training, I have seen as much as 98.4% on test set, but corresponding to lower validation accuracy. In general, several runs of the same configuration could end up with 0.3% difference in validation accuracy. So ultimately one could run several times the winner configuration until the lucky initial weights combination is reached.

    ReplyDelete
  64. I got 94.7% of accuracy following the udacity's deep learning course dataset partition:

    200000 for training
    10000 for validation and testing each.

    I took the lenet5 architecture as inspiration and added some tricks:

    Conv with depth 6
    Max pooling
    Conv with depth 16
    Max pooling
    Conv with depth 120
    Fully Connected layer with 84 units

    The kernel size was 5x5 for all the conv layers with stride of 1.

    I used L2 regularization in the fully connected weights with beta = 3 * 1e-3 and using softmax cross entropy in the loss function.

    Finally, I added dropout in the FC layer with prob = 0.5.

    The optimizer is SGD (minibatch size = 128) with exponential decay learning rate:

    initial Lr: 0.1
    decay steps: 100000
    decay rate: 0.96
    The learning rate is decayed at discrete intervals.

    This is a relatively small net for the current standards, so it runs pretty fast :)

    If it is run fore more steps it can achieve up to 96% in accuracy with around 5k steps. I suspect it can achieve even more, but I don't have either patience or a decent GPU :P

    ReplyDelete

  65. This information you provided in the blog that is really unique I love it!! Thanks for sharing such a great blog Keep posting..
    Learn How to Fix a Slow iPhone

    ReplyDelete
  66. I just released a similar dataset for Chinese characters.

    https://medium.com/@peterburkimsher/making-of-a-chinese-characters-dataset-92d4065cc7cc

    The notMNIST data is being used in CS231n, but I wanted to try something different. So I made my own! Please let me know if it's useful.

    ReplyDelete
  67. thanks for the great post.. this information are very useful to who is looking for the Cognos Online training institute. keep sharing

    ReplyDelete
  68. Thanks for the post got to more on Student Loans In India

    ReplyDelete
  69. These guys are really great job in the data sector by completing the whole data function and keep it in decorated way. visit site is the best option for the writing service help.

    ReplyDelete
  70. I am that kind of guy who love to do every sort of work by manually as if there has a mistake it let me know the facts for next time. see more details here and you'll be get and helpful ideas on the subject of academic papers writing.

    ReplyDelete
  71. In recent time this joint admission is really a serious matter where it could be a matter of good thinking on admission side. http://www.orthopedicresidency.com/why-our-orthopedic-residency-personal-statement/ for the students that is very helpful for the writing services.

    ReplyDelete
  72. After reading this blog i very strong in this topics and this blog really helpful to all.Ruby on Rails Online Course Bangalore

    ReplyDelete
  73. Anonymous5:19 AM

    Your blog is very informative. Explained Perfectly. Keep sharing this kind of information in your blog.

    IoT Training in Chennai | IoT Courses in Chennai

    ReplyDelete
  74. This notmist dataset is really important to make the whole site http://www.academicghostwriter.org/academic-writing-services/ this look valuable and make it happens for the rest of the time.

    ReplyDelete
  75. You might need these helpful link for the dataset of your works and you could literally get those things done easily with one hand.

    ReplyDelete
  76. This comment has been removed by the author.

    ReplyDelete
  77. Great article and helpful Information and Thanks for sharing this article and also visit we are leading the Best MATLAB Training in Jodhpur, web development & designing, Python, java, android, iPhone, PHP training institute in jodhpur

    ReplyDelete
  78. Thank you for sharing wonderful information with us to get some idea about that content. check it once through
    Best Workday Online Training From India
    Best Mule Esb Online Training From India
    Best Devops Online Training From India

    ReplyDelete
  79. Anonymous3:30 AM

    PCB Design Training in Bangalore offered by myTectra. India's No.1 PCB Design Training Institute. Classroom, Online and Corporate training in PCB Design
    pcb design training in bangalore

    ReplyDelete
  80. DIAC Automation Provide very good Automation training in Noida with the help of multiple brands of Industrial automation products. We Provide Advanced training in the layout, wiring, programming, troubleshooting and implementation of PLCs in an industrial enviornment, using hands on experimentation on real industrial PLCs and other industrial components such as: sensors, mechanical switches and relays to produce real programming experiments and control situations. Call @9953489987,,9310096831.

    ReplyDelete
  81. Thanks for Sharing,
    Keep Updating
    https://bitaacademy.com/

    ReplyDelete
  82. Thank you. it's a nice blog, keep posting.
    Click here:
    Data Science Online Training

    ReplyDelete
  83. I accept there are numerous more pleasurable open doors ahead for people that took a gander at your site. I wish to show thanks to you just for bailing me out of this particular trouble. As a result of checking through the net and meeting techniques that were not productive, I thought my life was done.
    Angularjs training in Chennai
    Angularjs training in Velachery

    ReplyDelete
  84. hi i am learning python programming.. i would like to be a Machine Learning Expert.. after learning python.. what should be the next framework i should learn before start learning machine language
    Digital Marketing Course in bangalore

    ReplyDelete
  85. Thank you for sharing this information it has helped me to know more about Educational Loan

    ReplyDelete
  86. The information which you have provided is very good. It is very useful who is looking for at machine learning online training Bangalore

    ReplyDelete
  87. nice blog, jangan lupa kunjungi juga Sedot WC Bojonegoro

    ReplyDelete
  88. I appreciate this informative blogpost. Thanks for sharing with us and find more details of Educational Trainers Exporter

    ReplyDelete
  89. This post is very good and helpfull and I must appreciate you for this.. Keep posting like this..
    Technoglobe is best for Machine Learning training in Jaipur

    ReplyDelete
  90. Laura8:16 PM

    That's really awesome, love it.
    vo genesis is one of the best voiceover products IMO.

    ReplyDelete
  91. Thank you for your post. This is excellent information. It is amazing and wonderful to visit your site.
    internships provider in hyderabad
    internship in hyderabad for cse

    ReplyDelete
  92. Check out wide range of designer brand iPhone case in a variety of shapes, colors and sizes at Uneek Inc.

    ReplyDelete
  93. This comment has been removed by the author.

    ReplyDelete
  94. Existing without the answers to the difficulties you’ve sorted out through this guide is a critical case, as well as the kind which could have badly affected my entire career if I had not discovered your website.
    nebosh course in chennai

    ReplyDelete
  95. It is really a great work and the way in which you are sharing the knowledge is excellent. Thanks for your informative article
    occupational health and safety course in chennai

    ReplyDelete
  96. This comment has been removed by the author.

    ReplyDelete
  97. Industries, construction companies, and manufacturers frequently hire the assistance of professionals to prepare their property for use or sale. Professionals who offer industrial cleaning Sydney use the equipment and tools required to clean heavy machinery, large workstations and even equipment used by industries.

    ReplyDelete
  98. Thank you for taking the time to write about this much needed subject. I felt that your remarks on this technology is helpful and were especially timely.

    Right now, DevOps is currently a popular model currently organizations all over the world moving towards to it. Your post gave a clear idea about knowing the DevOps model and its importance.

    devops course fees in chennai | devops training in chennai with placement | devops training in chennai omr | best devops training in chennai quora | devops foundation certification chennai

    ReplyDelete
  99. This comment has been removed by the author.

    ReplyDelete
  100. Anonymous8:15 AM

    Learn Free SELENIUM | JAVA | TestNG Tutorials using http://www.stqatools.com with Programs and Interview Questions and Answers.

    Thank You !!!

    http://stqatools.com/java/
    http://stqatools.com/selenium/
    http://stqatools.com/testng/

    ReplyDelete
  101. good job and thanks for sharing such a good blog You’re doing a great job.Keep it up !!

    machine learning training in jaipur


    python training in jaipur


    best data science training training in jaipur

    ReplyDelete
  102. This comment has been removed by the author.

    ReplyDelete
  103. Nice post!Everything about the future(học toán cho trẻ mẫu giáo) is uncertain, but one thing is certain: God has set tomorrow for all of us(toán mẫu giáo 5 tuổi). We must now trust him and in this regard, you must be(cách dạy bé học số) very patient.

    ReplyDelete
  104. nice course. thanks for sharing this post.
    Networking Training in Delhi

    ReplyDelete

  105. Very nice post here thanks for it .I always like and such a super contents of these post.Excellent and very cool idea and great content of different kinds of the valuable information's.

    Check out : best hadoop training in chennai
    hadoop big data training in chennai
    best institute for big data in chennai
    big data course fees in chennai

    ReplyDelete
  106. Full Stack Development Training in Chennai Searching for Full Stack Development training in chennai ? Bita Academy is the No 1 Training Institute in Chennai. Call for more details.

    ReplyDelete
  107. Really Happy to say your post is very interesting. Keep sharing your information regularly for my future reference. Thanks Again.
    Check Out:
    big data training in chennai chennai tamil nadu
    big data training in velachery
    big data hadoop training in velachery

    ReplyDelete
  108. Sungguh Konten yang sangat menarik.
    Saya tidak pernah melihat konten seperti yang ada pada semua blog Anda.
    Jangan lupa lihat juga konten saya disini
    situs casino online
    Terimakasih.. :)

    ReplyDelete
  109. Hey there! I know this is kind of off-topic, but I’d figured I’d ask. Would you be interested in exchanging links or maybe guest authoring a blog post or vice-versa?
    nebosh course in chennai
    offshore safety course in chennai

    ReplyDelete
  110. An astounding web diary I visit this blog, it's inconceivably magnificent. Strangely, in this current blog's substance
    Oracle Fusion Financials Online Training
    Oracle Fusion HCM Online Training
    Oracle Fusion SCM Online Training

    ReplyDelete
  111. Do not use all of these Private Money Lender here.They are located in Nigeria, Ghana Turkey, France and Israel.My name is Mrs.Emily Michael, I am from Canada. Have you been looking for a loan?Do you need an urgent personal or business loan?contact Fast Legitimate Loan Approval he help me with a loan of $95,000 some days ago after been scammed of $12,000 from a woman claiming to be a loan lender from Nigeria but i thank God today that i got my loan worth $95,000.Feel free to contact the company for a genuine financial contact Email:(creditloan11@gmail.com)

    ReplyDelete
  112. We offer real credit services to the general public, in order to improve mankind and save those who have bad credit. Credits We have certified and 100% reliable. We make sure that our customers are dealing with every encouragement with the best resources. Our service credit is very easy and effective. For more details about our credit, please contact us via E-mail: tivolifinancialhome@gmail.com WhatsApp us: +919873255004

    ReplyDelete
  113. I have found that this site is very informative, interesting and very well written. keep up the nice high quality writing
    Communication and Network Concepts

    ReplyDelete
  114. Thanks for the nice post and it is very useful for us and also very informative information to read.
    Apple iPhone Service Center in Chennai
    Oneplus Service Center in Chennai

    ReplyDelete

  115. This is an informative post and it is very useful and knowledgeable. therefore, I would like to thank you for the efforts you have made in writing this article.
    iphone app training course
    iphone training classes in bangalore
    iphone training

    ReplyDelete
  116. Nice post. Thanks for sharing this useful information.
    Best island resorts in kerala

    ReplyDelete


  117. Thank you for your post. This is excellent information. It is amazing and wonderful to visit your site.

    Scholarship preparation Classes at Cranbourne West

    ReplyDelete
  118. This is really good blog information thanks for sharing .I am really impressed with your writing abilities

    โปรโมชั่นGclub ของทางทีมงานตอนนี้แจกฟรีโบนัส 50%
    เพียงแค่คุณสมัคร Gclub กับทางทีมงานของเราเพียงเท่านั้น
    ร่วมมาเป็นส่วนหนึ่งกับเว็บไซต์คาสิโนออนไลน์ของเราได้เลยค่ะ
    สมัครสมาชิกที่นี่ >>> Gclub online

    ReplyDelete
  119. This comment has been removed by the author.

    ReplyDelete
  120. Really very interesting content.
    I have never seen content like everything on your blog.
    Don't forget to also see my content Thank you ... :)
    situs judi slot online terpercaya
    situs casino online terpercaya

    ReplyDelete
  121. SUMOBOLA - SITUS TARUHAN BOLA JALAN DAN MIX PARLAY TERBAIK DAN TERPERCAYA
    Bola Jalan
    Mix Parlay
    Sumo Bola
    Judi Bola

    ReplyDelete
  122. Truly commendable piece of information allocated by you. I am contented reading this worthwhile information here and I am sure this might be worthwhile for a majority of apprentices. Keep up with this tremendous work and continue updating.
    English practice App | English speaking app

    ReplyDelete
  123. Really very happy to say that your post is very interesting. I never stop myself to say something about it. You did a great job. Keep it up.
    We have an excellent IT courses training institute in Hyderabad. We are offering a number of courses that are very trendy in the IT industry. For further information, please once go through our site. DevOps Training In Hyderabad

    ReplyDelete
  124. Anonymous1:42 AM


    I was searching for loan to sort out my bills& debts, then i saw comments about Blank ATM Credit Card that can be hacked to withdraw money from any ATM machines around you . I doubted thus but decided to give it a try by contacting {blankatm156@gmail.com} they responded with their guidelines on how the card works. I was assured that the card can withdraw $5,000 instant per day & was credited with $50,000 so i requested for one & paid the delivery fee to obtain the card, after 24 hours later, i was shock to see the UPS agent in my resident with a parcel{card} i signed and went back inside and confirmed the card work's after the agent left. This is no doubts because i have the card & has made used of the card. This hackers are USA based hackers set out to help people with financial freedom!! Contact these email if you wants to get rich with this Via: blankatm156@gmail.com

    ReplyDelete
  125. A course in AI allows you to learn those skills. The advanced topics of convolutional neural networks, training deep networks, and recurrent neural networks are comprehended and mastered through the AI course. Many significant fields like customer service, financial services and the field of healthcare employ some of the major applications of Artificial Intelligence. The functioning of these AI applications is learned through an AI Course .

    ReplyDelete
  126. Nice blog with excellent information. Thank you. keep sharing.

    PERL Scripting Online Training

    ReplyDelete
  127. Thanks for sharing such a wonderful blog on Machine learning.This blog contains so much data about Machine learning ,like if anyone who is searching for the Machine learning data will easily grab the knowledge of Machine learning from this .Requested you to please keep sharing these type of useful content so that other can get benefit from your shared content.
    Thanks and Regards,
    Top institutes for machine learning in chennai
    best machine learning institute in chennai
    artificial intelligence and machine learning course in chennai

    ReplyDelete
  128. Excellent blogs!!!!you have for sharing them effect information..we developer very learning to easy.
    Apple iPhone Service Center in Chennai Anna Nagar


    ReplyDelete
  129. McAfee.com/Activate Since the world is developing each day with new computerized advances, digital dangers, malware, information, and harming diseases have additionally turned out to be increasingly more progressed with every day. These digital contaminations harm a gadget or documents in different ways.McAfee.com/Activate

    office.com/Setup is a software which is used by almost all company and business and even by individuals For all their office activities or for personal use. It has excels, word, and ppt as their constituent are most widely used apps. Install your office.com/Setup by downloading now.



    McAfee.com/ActivateMcafee is a antivirus software for laptop, PC, Mac for internet security form viruses and malware. Enter code to get started and protect while online surfing and downloading. McAfee.com/Activate

    ReplyDelete
  130. If you don"t mind proceed with this extraordinary work and I anticipate a greater amount of your magnificent blog entriesmachine learning training in bangalore

    ReplyDelete
  131. We are really grateful for your blog post. You will find a lot of approaches after visiting your post. Great work
    machine learning course in bangalore

    ReplyDelete
  132. . Our repair specialists are committed to serve you right the moment you contact us. LG Microwave Oven Service Center in Hyderabad Our repair center render quick and best repair services at door step.
    LG Microwave Oven Service Center in Hyderabad

    You can even appoint one of our service engineers to arrive at your place and carry out the repair and solve even more intricate issues of your microwave ovens. Our service engineer will reach you shortly upon your request or call and solve your problems.

    ReplyDelete
  133. www.mcafee.com/activate registered trademarks, company names, product names and brand names are the property of their respective owners, and mcafee.com/activate disclaims any ownership in such third-party marks. The use of any third party trademarks, logos, or brand names is for informational purposes only, and does not imply an endorsement by mfmcafee.com or vice versa or that such trademark owner has authorized mfmcafee.com to promote its products or services.


    www.office.com/setup is an independent support and service provider for the most secure remote technical services for all Office products. Our independent support services offer an instant support for all software related errors in the devices, laptops, desktops and peripherals. We have no link or affiliation with any of the brand or third-party company as we independently offer support service for all the product errors you face while using the Office. If your product is under warranty, then you may also avail our support services for free from manufacturer’s official website office.com/setup.

    mcafee activate is an independent support and service provider for the most secure remote technical services for all norton products. Our independent support services offer an instant support for all software related errors in the devices, laptops, desktops and peripherals. We have no link or affiliation with any of the brand or third-party company as we independently offer support service for all the product errors you face while using the norton. If your product is under warranty, then you may also avail our support services for free from manufacturer’s official website norton setup.

    ReplyDelete
  134. Anonymous1:54 AM

    Excellent Blog! I would like to thank for the efforts you have made in writing this post. I am hoping the same best work from you in the future as well. I wanted to thank you for this websites! Thanks for sharing. Great websites! Now please do visit our website which will be very helpful.
    machine learning course bangalore

    ReplyDelete
  135. UV Gullas College of MedicineUV Gullas College of Medicine- Do you Want to do MBBS in Philippines? Then make your decision with us.! Here no need any entrance examination.

    ReplyDelete
  136. This is very interesting article thanx for your knowledge sharing.this is my website is mechanical Engineering related and one of best site .i hope you are like my website .one vista and plzz checkout my site thank you, sir.
    mechanical engineering

    ReplyDelete
  137. Really very interesting content.
    I have never seen content like everything on your blog.
    Don't forget to also see my content Thank you ... :)

    situs judi poker

    ReplyDelete
  138. me project centers in chennai Real time projects centers provide for bulk best final year Cse, ece based IEEE me, mtech, be, BTech, MSC, mca, ms, MBA, BSC, BCA, mini, Ph.D., PHP, diploma project in Chennai for Engineering students in java, dot net, android, VLSI, Matlab, robotics, raspberry pi, python, embedded system, Iot, and Arduino . We are one of the leading IEEE project Center in Chennai.

    ReplyDelete
  139. Good Day Sir/Madam: Do you need an urgent loan to finance your business or in any purpose? We are certified and legitimate and international licensed loan lender we offer loans to Business firms. Individuals, companies firms, corporate bodies at an affordable interest rate of 3%. It might be a short or long term loan or even if you have poor credit. We shall process your loan as soon as we receive your application. We are an independent financial institution. We have built up an excellent reputation over the years in providing various types of loans to thousands of our customers. We offer Educational loan, Business loan, home loan, Agricultural loan, Personal loan, Auto loan with either a good or bad credit history. If you are interested in our above loan offer you are advice to fill the below information and return to us for more details. You can contact us with this email standardonlineinvestment@gmail.com we shall respond to you as soon as we receive your loan application details below.

    First name:
    Middle name:
    Date of birth (yyyy-mm-dd):
    Gender:
    Marital status:
    Total Amount Needed:
    Time Duration:
    Address:
    Currency Needed
    City:
    State/province:
    Zip/postal code:
    Country:
    Phone:
    Mobile/cellular:
    Monthly Income:
    Occupation:
    Which sites did you know about us.....
    standardonlineinvestment@gmail.com for immediate attention. Contact
    us now and get an urgent loan within two (2) days!!!
    Regards,
    Mr Abdul Muqse

    ReplyDelete
  140. Thank you for excellent article.You made an article that is interesting.
    Tavera car for rent in coimbatore|Indica car for rent in coimbatore|innova car for rent in coimbatore|mini bus for rent in coimbatore|tempo traveller for rent in coimbatore|kodaikanal tour package from chennai

    Keep on the good work and write more article like this...

    Great work !!!!Congratulations for this blog

    ReplyDelete
  141. Very gossipy post! I'm learning a lot from your articles. Keep us updated by sharing more such posts. it is helpful. Thanks for sharing.
    dot net training in chennai


    ReplyDelete
  142. UVGullasmedicalcollegeGet #MBBSinAbroad for low fees @ UV Gullas College of Medicine in Philippines. We are one of the best MBBS Universities in abroad.

    ReplyDelete
  143. vlsa global services providing best ece project in chennai.VLSA Global Services is the best ece projects in chennai , VLSA Global Services offers ece projects in Chennai and IEEE 2013 Final Year projects for Engineering students in JAVA, Dot Net, Android, Oracle, matlab, embedded system, python and PHP technologies

    ReplyDelete
  144. ipt training in chennai for your final year. DLK Career Development Center conduct national level inplant training programs.

    ReplyDelete
  145. final year project centers provides best projects for BE,ME,BCA,MCA and all other streams.we are best in doing all kinds of mca projects in chennai.we refer project centers students the great topic to select with and implement their ideas with new technologies.

    ReplyDelete
  146. I really like what you write in this blog, I also have some relevant Information about if you want more information. Thanks for sharing a piece of useful information.. we have learned so much information from your blog mtech project centers in chennai..... keep sharing

    ReplyDelete
  147. The explanation you given on machine learning is very useful. Thanks for sharing this innovative blog. Keep posting more in future.
    Interior Designers in Chennai
    Interior Decorators in Chennai
    Best Interior Designers in Chennai
    Home Interior designers in Chennai
    Modular Kitchen in Chennai

    ReplyDelete
  148. This is a decent post. This post gives genuinely quality data. I'm certainly going to investigate it. Actually quite valuable tips are given here. Much obliged to you to such an extent. Keep doing awesome. To know more information about
    Contact us :- https://www.login4ites.com/

    ReplyDelete
  149. phd projects in chennaicenters provide for bulk best final year Cse, ece based IEEE me, mtech, be, BTech, MSC, mca, ms, MBA, BSC, BCA, mini, Ph.D., PHP, diploma project in Chennai for Engineering students in java, dot net, android, VLSI, Matlab, robotics, raspberry pi, python, embedded system, Iot, and Arduino . We are one of the leading IEEE project Center in Chennai.

    ReplyDelete

  150. Just seen your Article, it amazed me and surpised me with god thoughts that eveyone will benefit from it. It is really a very informative post for all those budding entreprenuers planning to take advantage of post for business expansions. You always share such a wonderful articlewhich helps us to gain knowledge .Thanks for sharing such a wonderful article, It will be deinitely helpful and fruitful article.
    Thanks
    DedicatedHosting4u.com




    ReplyDelete
  151. UV Gullas Medical CollegeGet MBBS in Abroad for low fees @ UV Gullas College of Medicine in Philippines. We are one of the best MBBS Universities in abroad.
    Apply Now!

    ReplyDelete