Sometimes I really just want to interactively experiment with things like softmax(), or sigmoid() just to get a sense of how they behave. I'm struggling to be able to see the answer. Maybe I need to rewrite everything in numpy, but I hope not.
Example:
v = tf.sigmoid(tf.convert_to_tensor([0.123, 0.345]))
Now I have v, but heck if I can figure out how to see the values inside it. How can it be done?
In case you are running Tensorflow 2.0 -
v = tf.sigmoid(tf.convert_to_tensor([0.123, 0.345]))
v.numpy()
The answer is -
array([0.5307113, 0.5854046], dtype=float32)
If you are running Tensorflow 1.0 -
with tf.Session() as sess:
print(v.eval())
It gives the following answer -
[0.5307113 0.5854046]
Related
I'm teaching a course in Multivariate Calculus and decided to convert my notes from Sage to Jupyter using SymPy. I have rewritten nearly all my notes as Jupyter Notebooks and am very impressed how I can use multiple cells like Mathematica and I can use MarkDown cells with LaTeX as well as all the great features of matplotlib, NumPy and SymPy.
I'm nearly done converting my sagelets to Python scripts on Colab and found a discrepancy.
This Sage code resolves as pi:
integral(integral(integral(1, z, x^2+y^2, 2-x^2-y^2),
y, -sqrt(1-x^2), sqrt(1-x^2)),
x, -1, 1)
but this this SymPy code resolves as -pi/2:
Integral(1,
(z, x**2+y**2, 2-x**2-y**2),
(y, -sqrt(1-x**2), sqrt(1-x**2)),
(x, -1, 1)
).doit()
See #3 in this Jupyter Notebook:
https://colab.research.google.com/drive/1OlT9nfPG8TzoR_WpDavx-SAa07HLg3hV?usp=sharing
Shouldn't these be equal? What am I missing? Any help would be greatly appreciated as I've done A LOT of work on this course using SymPy and would like to use it in class this summer session!
Please help,
A. Jorge Garcia
Applied Math & CS
Nassau Community College
http://shadowfaxrant.blogspot.com
PS: Here's the SageCell version,
https://sagecell.sagemath.org/?z=eJzFk8FuwjAMhu9IvIMFB9qRTk1A07TrtE07cNuuSBkEGlGS4qRA-_RLQ1uQuDCGtlMsx_ns_5fTn-AbsJg-3scPoxg-J89sDB8os1TAu7JiiTw13U6fhvCy5WnOrQCu5v4OMxT2CWRdFpwHlJQkJiwkhTtGIdm7Yxx2OybRu6B3U2jPYbccg0FBykHT4seUdrYK52SzX0xIo-KArZvQo_DbYnsXyz1_e5A5unqe_ZQNiykjLHJR5KIKHJkN2oBWqZCcxFXviJ4a8deNL7XqOvrBzHEIr9JpsYmANTcG9MLHuZIWZvmXcNJ8YiHRWNAzy5WFXSIUYKJBGshQZxqt1IqnYLUvNpuco2hYc2ncq5ljoF77jEa5lKo19j-HaL_i6pKPuLoLareHZWWs39NmPf2yOmO_Adu6eJ4=&lang=sage
I want to compare two images for similarity. Since my purpose is to match a given image against a massive collection of images, I want to run the comparisons on GPU.
I came across tf.image.ssim and tf.image.psnr functions but I am unable to find and working examples only. The solutions in PyTorch is also appreciated. Since I don't have a good understanding of CUDA and C language, I am hesitant to try kernels in PyCuda.
Will it be helpful in terms of processing if I read the entire image collection and store as Tensorflow Records for future processing?
Any guidance or solution, greatly appreciated. Thank you.
Edit:- I am matching images of same size only. I don't want to do mere histogram match. I want to do SSIM or PSNR implementation for image similarity. So, I am assuming it would be similar in color, content etc
Check out the example on the tensorflow doc page (link):
im1 = tf.decode_png('path/to/im1.png')
im2 = tf.decode_png('path/to/im2.png')
print(tf.image.ssim(im1, im2, max_val=255))
This should work on latest version of tensorflow. If you use older versions tf.image.ssim will return a tensor (print will not give you a value), but you can call .run() to evaluate it.
There is no implementation of PSNR or SSIM in PyTorch. You can either implement them yourself or use a third-party package, like piqa which I have developed.
Assuming you already have torch and torchvision installed, you can get it with
pip install piqa
Then for the image comparison
import torch
from torchvision import transforms
from PIL import Image
im1 = Image.open('path/to/im1.png')
im2 = Image.open('path/to/im2.png')
transform = transforms.ToTensor()
x = transform(im1).unsqueeze(0).cuda() # .cuda() for GPU
y = transform(im2).unsqueeze(0).cuda()
from piqa import PSNR, SSIM
psnr = PSNR()
ssim = SSIM().cuda()
print('PSNR:', psnr(x, y))
print('SSIM:', ssim(x, y))
Using statsmodel's GLM, the tweedie deviance is included in the summary function, but I don't know how to do this for xgboost. Reading the API didn't help either.
In Python this is how you do it. Suppose predictions is the result of your gradient boosted tree and real are the actual numbers. Then using statsmodels you would run this:
import statsmodels as sm
dev = sm.families.Tweedie(pow_var=1.5).deviance(predictions, real)
Trying debug statements in Python/tensorflow1.0 using jupyter , but does not get any output printed from tf.Print
Thought sess.run(during training in below code) should have evaluated db1 tensor and print output which did not happen
However db1.eval in evaluate phase , printing entire tensor X with out "message X:".
def combine_inputs(X):
db1=tf.Print(X,[X],message='X:')
return (tf.matmul(X, W) + b,db1)
<<training code>>
_,summary=sess.run([train_op,merged_summaries])
## merged_summaries tensor triggers combine_inputs function. There are
## other tensor functions/coding in between , not giving entire code to keep
## it simple; code works as expected except tf.Print
<<evaluate code>>
print(db1.eval())
Confused on following
a) Why tf.Print is not printing during sess.run during training?
b) Why explicit db1.eval is necessary , expected tf.Print to trigger with
sess.run. If eval is required , could copy tensor X in my code to db1
and evaluate it with out tf.Print. Correct?
Tried going through other questions (like below one). Suggested to implement memory_util or predefined function. As learner could not understand why tf.Print does not work in my scenario
If anyone encountered similar issues , please assist. Thanks!
Similar question in stackoverflow
According to the documentation, tf.Print prints to standard error (as of version 1.1), and it's not compatible with jupyter notebook. That's why you can't see any output.
Check here:
https://www.tensorflow.org/api_docs/python/tf/Print
You can check the terminal where you launched the jupyter notebook to see the message.
import tensorflow as tf
tf.InteractiveSession()
a = tf.constant(1)
b = tf.constant(2)
opt = a + b
opt = tf.Print(opt, [opt], message="1 + 2 = ")
opt.eval()
In the terminal, I can see:
2018-01-02 23:38:07.691808: I tensorflow/core/kernels/logging_ops.cc:79] 1 + 2 = [3]
When I invoke tf.contrib.layers.convolution2d the tensorflow execution terminates with an error about one of the parameters used
got an unexpected keyword argument 'weight_init'
The parameter passed are the follows:
layer_one = tf.contrib.layers.convolution2d(
float_image_batch,
num_output_channels=32,
kernel_size=(5,5),
activation_fn=tf.nn.relu,
weight_init=tf.random_normal,
stride=(2, 2),
trainable=True)
That is exactly as described in the book that I'm reading. I suspect a possible syntax problem with weight_init=tf.random_normal written directly inside the call, but I don't know how to fix. I'm using Tensorflow 0.12.0
The book that you are reading (You didn't mention which one) might be using an older version of TensorFlow when the initial values for the weight tensor was passed through the weight_init argument. In the TensorFlow library version you are using (You didn't mention your TF version), probably that argument is replaced with weight_initializer. The latest (TensorFlow v0.12.0) documentation for tf.contrib.layers.convolution2d is here.
To fix your problem, you can change the following line in your code:
weight_init=tf.random_normal
to
weight_initializer=tf.random_normal_initializer()
According to the documentation, by default, tf.random_normal_initialier uses a 0.0 mean, a standard deviation of 1.0 and the datatype to be tf.float32. You may change the arguments as per your need using this line instead:
weight_initializer=tf.random_normal_initializer(mean=0.0, stddev=1.0, seed=None, dtype=tf.float32)