Tensorflow: Add small number before division for numerical stabilityHow to matmul a 2d tensor with a 3d tensor in tensorflow?VBScript: Expanding precision to 16 decimals to circumvent scientific notation?How do I make exponents not show leading zeros in vb.net?python scientific notation with forced leading zerobcdiv using very small float with scientific notation cause “Division by zero” errorphp str_replace casting integer to scientific numberTensorflow: NaN's without any divisionTensorflow results have the following form <tf.Tensor 'Add:0' shape=() dtype=int32>python decimal module without expHow small can a JavaScript number be without displaying in scientific notation?Matlab: apply scientific notation in textbox
How can I get a job without pushing my family's income into a higher tax bracket?
How do I calculate how many of an item I'll have in this inventory system?
Is the book wrong about the Nyquist Sampling Criterion?
Is 'contemporary' ambiguous and if so is there a better word?
Removing racism on a multi raced world
Endgame puzzle: How to avoid stalemate and win?
As a GM, is it bad form to ask for a moment to think when improvising?
How to ask systemd to not start a system service on boot?
Is any special diet a treatment of autism?
What are the advantages of luxury car brands like Acura/Lexus over their sibling non-luxury brands Honda/Toyota?
getline() vs. fgets(): Control memory allocation
Why did the Apollo 13 crew extend the LM landing gear?
Why does sound not move through a wall?
Should homeowners insurance cover the cost of the home?
Should I simplify my writing in a foreign country?
Importing a Part of the JSON
Extra space in cells when using token lists to build tabular content
Is Soreness in Middle Knuckle of Fretting Hand Index Finger Normal for Beginners?
Nested loops to process groups of pictures
How can I get people to remember my character's gender?
Prove that a definite integral is an infinite sum
Start job from another SQL server instance
Where are the "shires" in the UK?
Typeset year in old-style numbers with biblatex
Tensorflow: Add small number before division for numerical stability
How to matmul a 2d tensor with a 3d tensor in tensorflow?VBScript: Expanding precision to 16 decimals to circumvent scientific notation?How do I make exponents not show leading zeros in vb.net?python scientific notation with forced leading zerobcdiv using very small float with scientific notation cause “Division by zero” errorphp str_replace casting integer to scientific numberTensorflow: NaN's without any divisionTensorflow results have the following form <tf.Tensor 'Add:0' shape=() dtype=int32>python decimal module without expHow small can a JavaScript number be without displaying in scientific notation?Matlab: apply scientific notation in textbox
.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty height:90px;width:728px;box-sizing:border-box;
In order to prevent divisions by zero in TensorFlow, I want to add a tiny number to my dividend. A quick search did not yield any results. In particular, I am interested in using the scientific notation, e.g.
a = b/(c+1e-05)
How can this be achieved?
tensorflow scientific-notation numerical-stability
add a comment |
In order to prevent divisions by zero in TensorFlow, I want to add a tiny number to my dividend. A quick search did not yield any results. In particular, I am interested in using the scientific notation, e.g.
a = b/(c+1e-05)
How can this be achieved?
tensorflow scientific-notation numerical-stability
Assuming a,b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c.
– Souradeep Nanda
Mar 23 at 3:36
how can 1e-05 directly be specified as a tensor?
– Fábio
Mar 23 at 5:03
1
Tensorflow automatically typecasts1e-5totf.constant(1e-5). You can write it manually if you want.
– Souradeep Nanda
Mar 23 at 5:05
1
Thank you. Do you mind posting this as an answer to the question below?
– Fábio
Mar 23 at 10:01
add a comment |
In order to prevent divisions by zero in TensorFlow, I want to add a tiny number to my dividend. A quick search did not yield any results. In particular, I am interested in using the scientific notation, e.g.
a = b/(c+1e-05)
How can this be achieved?
tensorflow scientific-notation numerical-stability
In order to prevent divisions by zero in TensorFlow, I want to add a tiny number to my dividend. A quick search did not yield any results. In particular, I am interested in using the scientific notation, e.g.
a = b/(c+1e-05)
How can this be achieved?
tensorflow scientific-notation numerical-stability
tensorflow scientific-notation numerical-stability
asked Mar 23 at 0:55
FábioFábio
1,79021129
1,79021129
Assuming a,b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c.
– Souradeep Nanda
Mar 23 at 3:36
how can 1e-05 directly be specified as a tensor?
– Fábio
Mar 23 at 5:03
1
Tensorflow automatically typecasts1e-5totf.constant(1e-5). You can write it manually if you want.
– Souradeep Nanda
Mar 23 at 5:05
1
Thank you. Do you mind posting this as an answer to the question below?
– Fábio
Mar 23 at 10:01
add a comment |
Assuming a,b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c.
– Souradeep Nanda
Mar 23 at 3:36
how can 1e-05 directly be specified as a tensor?
– Fábio
Mar 23 at 5:03
1
Tensorflow automatically typecasts1e-5totf.constant(1e-5). You can write it manually if you want.
– Souradeep Nanda
Mar 23 at 5:05
1
Thank you. Do you mind posting this as an answer to the question below?
– Fábio
Mar 23 at 10:01
Assuming a,b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c.
– Souradeep Nanda
Mar 23 at 3:36
Assuming a,b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c.
– Souradeep Nanda
Mar 23 at 3:36
how can 1e-05 directly be specified as a tensor?
– Fábio
Mar 23 at 5:03
how can 1e-05 directly be specified as a tensor?
– Fábio
Mar 23 at 5:03
1
1
Tensorflow automatically typecasts
1e-5 to tf.constant(1e-5). You can write it manually if you want.– Souradeep Nanda
Mar 23 at 5:05
Tensorflow automatically typecasts
1e-5 to tf.constant(1e-5). You can write it manually if you want.– Souradeep Nanda
Mar 23 at 5:05
1
1
Thank you. Do you mind posting this as an answer to the question below?
– Fábio
Mar 23 at 10:01
Thank you. Do you mind posting this as an answer to the question below?
– Fábio
Mar 23 at 10:01
add a comment |
1 Answer
1
active
oldest
votes
Assuming a, b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c. Tensorflow automatically typecasts the 1e-5 to tf.constant(1e-5).
Tensorflow however has some limitations with non-scalar broadcasts. Take a look at my other answer.
add a comment |
Your Answer
StackExchange.ifUsing("editor", function ()
StackExchange.using("externalEditor", function ()
StackExchange.using("snippets", function ()
StackExchange.snippets.init();
);
);
, "code-snippets");
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "1"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f55309585%2ftensorflow-add-small-number-before-division-for-numerical-stability%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
Assuming a, b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c. Tensorflow automatically typecasts the 1e-5 to tf.constant(1e-5).
Tensorflow however has some limitations with non-scalar broadcasts. Take a look at my other answer.
add a comment |
Assuming a, b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c. Tensorflow automatically typecasts the 1e-5 to tf.constant(1e-5).
Tensorflow however has some limitations with non-scalar broadcasts. Take a look at my other answer.
add a comment |
Assuming a, b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c. Tensorflow automatically typecasts the 1e-5 to tf.constant(1e-5).
Tensorflow however has some limitations with non-scalar broadcasts. Take a look at my other answer.
Assuming a, b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c. Tensorflow automatically typecasts the 1e-5 to tf.constant(1e-5).
Tensorflow however has some limitations with non-scalar broadcasts. Take a look at my other answer.
answered Mar 24 at 3:54
Souradeep NandaSouradeep Nanda
1,09411625
1,09411625
add a comment |
add a comment |
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f55309585%2ftensorflow-add-small-number-before-division-for-numerical-stability%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Assuming a,b and c are tensors. The formula you have written will work as expected. 1e-5 will be broadcasted and added on the tensor c.
– Souradeep Nanda
Mar 23 at 3:36
how can 1e-05 directly be specified as a tensor?
– Fábio
Mar 23 at 5:03
1
Tensorflow automatically typecasts
1e-5totf.constant(1e-5). You can write it manually if you want.– Souradeep Nanda
Mar 23 at 5:05
1
Thank you. Do you mind posting this as an answer to the question below?
– Fábio
Mar 23 at 10:01