TensorFlow - Session 02

Can someone help me with this issue?

This comment has been removed.

Hi Team,

IFrame not able to embedd html on my jupyter. It is giving me a blank page. Is there any issue with my browser which is stopping it.Please let me know the solution, I am not abe to visualize tensor graphs on jupyter.

Regards,

Birendra Singh

Hi, is the Modularity(last topic of this video) part is over ? or their is another video for that.

At 50:14

I did not understand.

tf,Variable(0.2, name="b")

tf.Variable=(0.3,name="c")

What are we calculating here??

i did not understand.

Tensor Board

Why are we creating summary in MSE in step2 ??

This comment has been removed.

Hi Team,
I am facing below error while running tensorflow.ipynb
module 'tensorflow' has no attribute 'reset_default_graph'

I could not under this piece of coce in the deeplearning notebook <b>tensorflow.ipynb</b>. Maybe I have missed some basic python syntax.

def differentiate(f, x):
dx = 0.00001
return (f(x+dx) - f(x))/dx


def sq(x):
    return x*x

def cube(x):
    return x*x*x

differentiate(sq, 5)

when the 'differentiate' function is called, first paraemeter is just 'sq', How does that work when 'sq' function itself takes a parameter 'x'? If no parameter is supplied at the time

of the fucntion call, is there any default value that is assumed for parameter 'x'?

Thanks in advance

Thank you sir  but i could not able to  see  tensor board with following URL

http://g.cloudxlab.com.ec2.internal :6006

Help Me how to do that 

e

Help Me,  i am stuck in the first step  of starting   Tensor board  , 

3rd sreenshot

___________________________

2nd sreenshot

________________________

Hello,

I need your help in understanding few issues reproducing the Tensorflow graph on browser using web console. Please refer to the 3 images below and answer the queries below:

1) Why am I able to see only 2 runs /graph? Although when I refer to your video TensorFlow- session 02 - 2:24:13 part, I see that there are multiple runs inside the TensorBoard

2) Even the folder 'ml/deep_learning/tf_logs' does not contain multiple logs. Am I missing something here?

3) What steps do we need to take to visualize multiple logs and graphs?

4) Or if I am able to see only one log after executing the code once in Jupyter notebook, there will be only one graph? Please advise if I am wrong or else please guide.

Thanks

One more question -

What does to, msev mean here?

>>  if epoch % 100 == 0:
                to, msev = sess.run([training_op, mse], feed_dict={X: X_batch, y: y_batch})
                erros.append(msev)

Thanks

__________________________________________________________________

# This is execution phase
# Fetch the mini-batches one by one then provide the value of x and y via the feed_dict parameter

erros = []
with tf.Session() as sess:
    sess.run(init)
    for epoch in range(n_epochs):
        for batch_index in range(n_batches):
            X_batch, y_batch = fetch_batch(epoch, batch_index, batch_size)
            if epoch % 100 == 0:
                to, msev = sess.run([training_op, mse], feed_dict={X: X_batch, y: y_batch})
                erros.append(msev)
            else:
                sess.run(training_op, feed_dict={X: X_batch, y: y_batch})
    best_theta = theta.eval()
print(best_theta)

Hello,

Please let me know why are we using the this code below?

>> init = tf.global_variables_initializer()

Which global variables are we initializing and for what o/p will this function be used? 

When I look at the help it says: " An Op that initializes global variables in the graph.". Which graph and variables are we talking about here?

Thanks

______________________________________________________________________

#Let's implement mini-batch using placeholder nodes

reset_graph()

learning_rate = 0.01

# Let's make X and y as placeholder nodes

X = tf.placeholder(tf.float32, shape=(None, n + 1), name="X")
y = tf.placeholder(tf.float32, shape=(None, 1), name="y")

theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0, seed=42), name="theta")
y_pred = tf.matmul(X, theta, name="predictions")
error = y_pred - y
mse = tf.reduce_mean(tf.square(error), name="mse")
optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)
training_op = optimizer.minimize(mse)

init = tf.global_variables_initializer()

This comment has been removed.

This comment has been removed.

Hi Sir,

       Kindly, provide the tensor python code book which is using the latest tensorflow 2.0. Since, I have install the tensorflow 2.0. It is giving the error in the notebook that

AttributeError: module 'tensorflow' has no attribute 'reset_default_graph'. Kindly, help.
 

Hi

I am not able to get the graph ,Its not giving any error. I am  running the code locally and I have created a new environment where tensorflow 1.10 and python 3.6.9  is installed .

Hi,

Graph is not being populated in the notebook "tenserflow.ipynb" in the section "9  Visualizing the graph - 9.1  inside Jupyter and 9.2  Using TensorBoard"

Regards,

Punit

Hi,

I am getting issue with Tensorboard. I got error for "source activate py36". Could not find py36 env. with "conda info --envs". I was still able to start the TensorBoard but not able to open it in the browser.

f.cloudxlab.com:6006

Hi,

Why we are calacualting Gradient ? Only to know the value of Theta or going to use it to Minimize the cost function i.e. MSE.

Sir i use tensorflow 2 so there are certain things not applicable from here when i try myself. I've tried using tf.compat.v1 in some cases, but got stuck here in summary.filewriter. could you please tell me what can i do to fix this problem?

sess=tf.session()

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-10-6c7cd0cabe5e> in <module>
----> 1 sess=tf.session()

AttributeError: module 'tensorflow' has no attribute 'session'

In manually calculating method, we take initial value of theta as: tf.Variable(tf.random_uniform([n+1,1],-1.0,1.0,seed=42).

Could you please expain what is n+1 here and -1.0,1.0?

Also, is theta too a global variable?

Thanks.

In the above video at 1.28.59, the following equation is being shown. It is about converting the data to real-time data. But I didn't got the concept of these equations. like, epocj is multiplied by n-batches and added to batch_index. why multiplied to n_batches. 

then again, nbatches is calculated as int(np.ceil(m/batch_size)). what is m. why we dividing by batch size.

Please explain.

Hi,
Can you please provide me a tutorial or simplified code for these 2 projects

cross attenation for image text
2)text to image conversion

please reply sir it is an urgent need.

Thank you

Hi,

I am running this locally in tensorflow2.2.0 and successfully runned all the above codes and successfully showing the results.
But just on running "show_graph(tf.compat.v1.get_default_graph())" it is not showing and graph but it is also not showing any error. Please help me out. 
Thanks

Hi,

How can we run tensorboard locally and not in cloudxlab?? as i am unable to proceed fiurther now. 

I have tensorflow version- (in my local device)

import tensorflow
print(tensorflow.__version__)

2.0.0

but I am getting this error while executing reset default graph. please help me sir

(error:please see below and 
 the picture i have uploaded)

AttributeError: module 'tensorflow' has no attribute 'reset_default_graph'

AttributeError: module 'tensorflow' has no attribute 'Session'

AttributeError: module 'tensorflow' has no attribute 'global_variables_initializer'

Hi,

for epoch in range(n_epochs):
        if epoch % 10 == 0:
            print("Epoch", epoch, "MSE =", mse.eval())
        x = sess.run(training_op)
    best_theta = theta.eval()
In this loop what is the significance of this line???

Hi,

theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0, seed=42), name="theta") 

in the above code why we are writing [n + 1, 1 and  -1.0, 1.0????

Hi,
I have come upto here(screenshot). then in another browser I do: f.cloudxlab.com/6011
--> error: Site cant be reached. Page took too long to respond.
Any suggestions? can't be the internet. I am able to watch video without buffer.

After launching tensorboard on terminal, i m not able to open it in browser (screenshot attached

Hi,

I am unable to open tensorboard in the browser after launching it from the terminal. Screenshot attached

Please help.

Hi
I am not able to open tensorboard in browser as shown ion screenshot

Can we not use the packages like we used to do in sklearn for applying different kinds of models?

Hi
I am not able to activate python3 in console for running tensorboard. Plz help
Thanks

Hi
In Visualizuing the graph portion, code is running without any error but graph is not displayed.Plz help.
Thanks
Prachi

Hi
How we would decide momentum factor value in momentum gradient descent optimizer?
What is better approach out of gradient descent optimizer and momentum gradient descent optimizer ?

Hello,

when i am trying to get grap inside jupyter using the same code. it is just showing blank.. no graph..
please help me out with it...
i have pasted it at the end of code after eveluating...
do we have to draw just after graphing code?

housing_data_plus_bias = np.c_[np.ones((m, 1)), housing.data]

here what is np.c_ ??

Hi
I'm getting this error.

Another question : The function: gradients = Gradient(ops,[theta]) [0]

Will the elements in the gradients list be the same as the number of elements in the theta list?
I didn't see you define the theta list in the video , if you did , please tell me where it is,

A question : what is the metaphor meaning for the tesorflow session?

When tried to activate py36 below issue came

Please help me with below issue.

Once we created a session, we need to close the session right..!!, but why do you not closing the session after it's use?

for mini batch gradient descent, X is having 8 features, for place holder of X, why do we need to use n + 1?

While working on the terminal i am not able able to import tensorflow as the module is absent. I am not able to install it as well due to some permission issues. I believe that is why i am not able to use tensorboard. Can you please help?

source activate py36 gives the error: "bash: activate: No such file or directory"