The main difference is that TensorFlow is a framework for machine learning, and OpenCV is a library for computer vision. It can be a good start to check the link below to get a grasp for the difference between framework and library: https://stackoverflow.com/questions/148747/what-is-the-difference-between-a-framework-and-a-library 

You can do image recognition with TensorFlow. Though it is suited for more general problems as well, such as: classification, clustering and regression.

I guess people downvoted because this question might be more relevant to: https://datascience.stackexchange.com/ 

 

     import { Observable, of } from &quot;rxjs&quot;;
    
    // And if I try to return like this
      return Observable.of(this.purposes);

I am getting an error stating, Property &#39;of&#39; does not exist on type &#39;typeof Observable&#39;

Could not use Observable.of in RxJs 6 and Angular 6

I need to stack widgets like this:

[![coins][1]][1]

I wrote the code below. However the coins are coming one after another with some default padding. How can I get something like the image above?

```dart
Row(
  children: &lt;Widget&gt;[
    Icon(
      Icons.monetization_on, size: 36.0, 
      color: const Color.fromRGBO(218, 165, 32, 1.0),
    ),
    Icon(
      Icons.monetization_on, size: 36.0,
      color: const Color.fromRGBO(218, 165, 32, 1.0),
    ),
  ],
),
```


[1]: https://i.stack.imgur.com/4AaeT.png

How do I stack widgets overlapping each other in flutter

I&#39;m new into the AI world, I&#39;ve start doing some stuff using Python &amp; OpenCV for face detection and so on. I know that with the implementation of some algorithms I can develop AI system using Python &amp; OpenCV. So my question is : What is the position of Tensorflow here? Can I say Tensorflow is an alternative to OpenCV? as I can say Python is an alternative programming language to Java (for example).

Tensorflow vs OpenCV

<p>I'm new into the AI world, I've start doing some stuff using Python &#x26; OpenCV for face detection and so on. I know that with the implementation of some algorithms I can develop AI system using Python &#x26; OpenCV. So my question is : What is the position of Tensorflow here? Can I say Tensorflow is an alternative to OpenCV? as I can say Python is an alternative programming language to Java (for example).</p>


I am running pylint on an opencv project and I am getting many pylint errors in VS code about  members not being present.

Example code:

    import cv2
    cv2.imshow(....)

Errors obtained: 

[![enter image description here][1]][1]


  [1]: https://i.stack.imgur.com/4Kbdh.png

However , the code runs correctly without any errors.

Versions : pylint 1.8.1 , astroid 1.6.0

PyLint not recognizing cv2 members

As I&#39;m lead to believe, OpenCV reads images in BGR &lt;strike&gt;colorspace&lt;/strike&gt; ordering and we usually have to convert it back to RGB like this:

    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

But when I try to simply read an image and show it, the coloring seems fine (without the need to convert BGR to RGB):

    img_bgr = cv2.imread(image_path)
    cv2.imshow(&#39;BGR Image&#39;,img_bgr)
    img_rgb = cv2.cvtColor(img_bgr, cv2.COLOR_BGR2RGB)
    cv2.imshow(&#39;RGB Image&#39;,img_rgb )
    cv2.waitkey(0)

So is `imshow()` changing the ordering within the function automatically (from BGR to RGB) or the ordering has been BGR all along?



Python OpenCV - imshow doesn&#39;t need convert from BGR to RGB

I&#39;ve just begun using the OpenCV library for Python and came across something I didn&#39;t understand.

    cap = cv2.VideoCapture(0)

    while True:
          ret, frame = cap.read() #returns ret and the frame
          cv2.imshow(&#39;frame&#39;,frame)

          if cv2.waitKey(1) &amp; 0xFF == ord(&#39;q&#39;):
              break

When I use `cv2.waitKey(1)`, I get a continuous live video feed from my laptops webcam. However when I use `cv2.waitKey(0)`, I get still images. Every time I close the window, another one pops up with another picture taken at the time. 
Why does it not show as a continuous feed?

Difference in output with waitKey(0) and waitKey(1)

I need to install `cv2` for a script that has been written for me. I tried  `pip install cv2` and `pip install open_cv` and got the same problem - a warning message from `dist.py` and complains about `zlib` being not found. No cv2 installed. I also tried `pyopenvc` and `pip install opencv-python`.

So, I went to the opencv site and downloaded the relevant exe. Ran it - generated a heap of subdirectories and a make file and stuff. 

What do I do now?

How do I install opencv using pip?

So I am trying to use:

    sift = cv2.xfeatures2d.SIFT_create()

and it is coming up with this error:

 

    cv2.error: OpenCV(3.4.3) C:\projects\opencv-python\opencv_contrib\modules\xfeatures2d\src\sift.cpp:1207: error: (-213:The function/feature is not implemented) 
    This algorithm is patented and is excluded in this configuration; Set OPENCV_ENABLE_NONFREE CMake 
    option and rebuild the library in function &#39;cv::xfeatures2d::SIFT::create&#39;

I am using `Python 3.5.0` and `opencv(3.4.3)` and I am just using idle. This occured after I tried to install TensorFlow and I have tried looking around and have installed opencv-contrib-python but I am still getting the same error. Thank you in advance and I apologise if I have not included enough info




sift = cv2.xfeatures2d.SIFT_create() not working even though have contrib installed

In `Keras`, the high-level deep learning library, there are multiple types of recurrent layers; these include `LSTM` (Long short term memory) and `CuDNNLSTM`. According to the [Keras documentation](https://keras.io/layers/recurrent/#cudnnlstm), a `CuDNNLSTM` is a:

&gt; Fast LSTM implementation backed by CuDNN.
 Can only be run on GPU, with the TensorFlow backend.

It is my belief that Keras automatically uses the GPU wherever possible. According to the [TensorFlow build instructions](https://www.tensorflow.org/install/install_sources), to have a working TensorFlow GPU backend, you will need CuDNN:

&gt; The following NVIDIA software must be installed on your system:

&gt;   - NVIDIA&#39;s Cuda Toolkit (&gt;= 7.0). We recommend version 9.0. For details, see NVIDIA&#39;s documentation. Ensure that you append the relevant Cuda pathnames to the LD_LIBRARY_PATH environment variable as described in the NVIDIA documentation.
&gt;   - The NVIDIA drivers associated with NVIDIA&#39;s Cuda Toolkit.
&gt;   - __cuDNN__ (&gt;= v3). We recommend version 6.0. For details, see NVIDIA&#39;s documentation, particularly the description of appending the appropriate pathname to your LD_LIBRARY_PATH environment variable.


Therefore, how would a `CuDNNLSTM` differ in any way from a normal `LSTM` using a TensorFlow GPU backend? Will `CuDNNLSTM` be automatically selected and replace the normal `LSTM` when an available TensorFlow GPU backend is found?


What is the difference between CuDNNLSTM and LSTM in Keras?

I got this error when I tried to modify the learning rate parameter of SGD optimizer in Keras. Did I miss something in my codes or my Keras was not installed properly? 

Here is my code:

    from tensorflow.python.keras.models import Sequential
    from tensorflow.python.keras.layers import Dense, Flatten, GlobalAveragePooling2D, Activation
    import keras
    from keras.optimizers import SGD
    
    model = Sequential()
    model.add(Dense(64, kernel_initializer=&#39;uniform&#39;, input_shape=(10,)))
    model.add(Activation(&#39;softmax&#39;))
    model.compile(loss=&#39;mean_squared_error&#39;, optimizer=SGD(lr=0.01), metrics= [&#39;accuracy&#39;])*

and here is the error message:

&gt; Traceback (most recent call last):   File
&gt; &quot;C:\TensorFlow\Keras\ResNet-50\test_sgd.py&quot;, line 10, in &lt;module&gt;
&gt;     model.compile(loss=&#39;mean_squared_error&#39;, optimizer=SGD(lr=0.01), metrics=[&#39;accuracy&#39;])   File
&gt; &quot;C:\Users\nsugiant\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\keras\_impl\keras\models.py&quot;,
&gt; line 787, in compile
&gt;     **kwargs)   File &quot;C:\Users\nsugiant\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\keras\_impl\keras\engine\training.py&quot;,
&gt; line 632, in compile
&gt;     self.optimizer = optimizers.get(optimizer)   File &quot;C:\Users\nsugiant\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\keras\_impl\keras\optimizers.py&quot;,
&gt; line 788, in get
&gt;     raise ValueError(&#39;Could not interpret optimizer identifier:&#39;, identifier) ValueError: (&#39;Could not interpret optimizer identifier:&#39;,
&gt; &lt;keras.optimizers.SGD object at 0x000002039B152FD0&gt;)

&quot;Could not interpret optimizer identifier&quot; error in Keras

Using ResNet50 pre-trained Weights I am trying to build a classifier. The code base is fully implemented in Keras high-level Tensorflow API. The complete code is posted in the below GitHub Link. 

Source Code: [Classification Using RestNet50 Architecture][1]

The file size of the pre-trained model is **94.7mb**.

I loaded the pre-trained file 

    new_model = Sequential()
    
    new_model.add(ResNet50(include_top=False,
                    pooling=&#39;avg&#39;,
                    weights=resnet_weight_paths))

and fit the model 

    train_generator = data_generator.flow_from_directory(
        &#39;path_to_the_training_set&#39;,
        target_size = (IMG_SIZE,IMG_SIZE),
        batch_size = 12,
        class_mode = &#39;categorical&#39;
        )
    
    validation_generator = data_generator.flow_from_directory(
        &#39;path_to_the_validation_set&#39;,
        target_size = (IMG_SIZE,IMG_SIZE),
        class_mode = &#39;categorical&#39;
        )
    
    #compile the model
    
    new_model.fit_generator(
        train_generator,
        steps_per_epoch = 3,
        validation_data = validation_generator,
        validation_steps = 1
    )

and in the Training dataset, I have two folders dog and cat, each holder almost 10,000 images. When  I compiled the script, I get the following error

&gt; Epoch 1/1 2018-05-12 13:04:45.847298: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 38535168
&gt; exceeds 10% of system memory. 2018-05-12 13:04:46.845021: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 37171200
&gt; exceeds 10% of system memory. 2018-05-12 13:04:47.552176: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 37171200
&gt; exceeds 10% of system memory. 2018-05-12 13:04:48.199240: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 37171200
&gt; exceeds 10% of system memory. 2018-05-12 13:04:48.918930: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 37171200
&gt; exceeds 10% of system memory. 2018-05-12 13:04:49.274137: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 19267584
&gt; exceeds 10% of system memory. 2018-05-12 13:04:49.647061: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 19267584
&gt; exceeds 10% of system memory. 2018-05-12 13:04:50.028839: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 19267584
&gt; exceeds 10% of system memory. 2018-05-12 13:04:50.413735: W
&gt; tensorflow/core/framework/allocator.cc:101] Allocation of 19267584
&gt; exceeds 10% of system memory.

Any ideas to optimize the way to load the pre-trained model (or) get rid of this warning message?

Thanks!

  [1]: https://gist.github.com/Madhivarman/676650f71ec35a5f2802631fcfa0ff73


Tensorflow Allocation Memory: Allocation of 38535168 exceeds 10% of system memory

I have created a custom tensorflow lite model using retrain.py from https://github.com/tensorflow/hub/blob/master/examples/image_retraining/retrain.py using the following command

    python retrain.py --image_dir newImageDirectory --tfhub_module https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/feature_vector/1

Then I convert using `toco` the `output_graph.pb` file to a lite file. Using the below command:

    bazel run tensorflow/contrib/lite/toco:toco -- --input_file=/tmp/output_graph.pb --output_file=/tmp/optimized.lite --input_format=TENSORFLOW_GRAPHDEF --output_format=TFLITE --inpute_shape=1,224,224,3 --input_array=input --output_array=final_result --inference_type=FLOAT --input_data_type=FLOAT

Then I take the new lite file and the labels.txt file and put them in tensorflow for poets 2 https://github.com/googlecodelabs/tensorflow-for-poets-2 to see if I can have it start to classify new categories. When the application launches I receive the following error:

    Caused by: java.lang.NullPointerException: Can not allocate memory for the interpreter                                                                                            at org.tensorflow.lite.NativeInterpreterWrapper.createInterpreter(Native Method)
                                                                          at org.tensorflow.lite.NativeInterpreterWrapper.&lt;init&gt;(NativeInterpreterWrapper.java:63)
                                                                                            at org.tensorflow.lite.NativeInterpreterWrapper.&lt;init&gt;(NativeInterpreterWrapper.java:51)
                                                                                            at org.tensorflow.lite.Interpreter.&lt;init&gt;(Interpreter.java:90)
                                                                                            at com.example.android.tflitecamerademo.ImageClassifier.&lt;init&gt;(ImageClassifier.java:97)



Tensorflow Custom TFLite java.lang.NullPointerException: Cannot allocate memory for the interpreter

I have noticed that some newer TensorFlow versions are incompatible with older CUDA and cuDNN versions. Does an overview of the compatible versions or even a list of officially tested combinations exist? I can&#39;t find it in the TensorFlow documentation. 

Which TensorFlow and CUDA version combinations are compatible?

I installed TensorFlow 1.10.1 but when I tried to import TensorFlow it said that I need TensorFlow version 1.10.0. Thus, I installed it and now I get the following warnings:

```
&gt;&gt;&gt; import tensorflow
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorflow\python\framework\dtypes.py:516: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_qint8 = np.dtype([(&quot;qint8&quot;, np.int8, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorflow\python\framework\dtypes.py:517: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_quint8 = np.dtype([(&quot;quint8&quot;, np.uint8, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorflow\python\framework\dtypes.py:518: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_qint16 = np.dtype([(&quot;qint16&quot;, np.int16, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorflow\python\framework\dtypes.py:519: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_quint16 = np.dtype([(&quot;quint16&quot;, np.uint16, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorflow\python\framework\dtypes.py:520: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_qint32 = np.dtype([(&quot;qint32&quot;, np.int32, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorflow\python\framework\dtypes.py:525: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  np_resource = np.dtype([(&quot;resource&quot;, np.ubyte, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorboard\compat\tensorflow_stub\dtypes.py:541: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_qint8 = np.dtype([(&quot;qint8&quot;, np.int8, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorboard\compat\tensorflow_stub\dtypes.py:542: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_quint8 = np.dtype([(&quot;quint8&quot;, np.uint8, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorboard\compat\tensorflow_stub\dtypes.py:543: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_qint16 = np.dtype([(&quot;qint16&quot;, np.int16, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorboard\compat\tensorflow_stub\dtypes.py:544: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_quint16 = np.dtype([(&quot;quint16&quot;, np.uint16, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorboard\compat\tensorflow_stub\dtypes.py:545: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  _np_qint32 = np.dtype([(&quot;qint32&quot;, np.int32, 1)])
C:\Users\PC\Anaconda3\envs\tut\lib\site-packages\tensorboard\compat\tensorflow_stub\dtypes.py:550: FutureWarning: Passing (type, 1) or &#39;1type&#39; as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.
  np_resource = np.dtype([(&quot;resource&quot;, np.ubyte, 1)])

&quot;synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / &#39;(1,)type&#39;.&quot; problem in TensorFlow

Content Type	Original Author	Original Content on Stackoverflow
Question	binmosa	View Question on Stackoverflow
Solution 1 - Opencv	Emmet B	View Answer on Stackoverflow

Tensorflow vs OpenCV

Opencv Problem Overview

Opencv Solutions

Solution 1 - Opencv

How do I stack widgets overlapping each other in flutter

Could not use Observable.of in RxJs 6 and Angular 6

Attributions