New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
How to build graph once and run multi images/masks with different resolutions? #194
Comments
Hey @jtetc . That is a very good question. And I appreciate that you have searched over issues before asking. In principle our trained models support different image resolutions. You could have a try by setting input shapes as (1, None, None, 3). However, as I tried last year, I failed. The major obstacle lies in the API of |
Thanks for your instant response @JiahuiYu Thanks again. |
@jtetc Great! Thanks for letting me know. |
Hi, @jtetc
where are the And I want to set input shape as (1, None, None, 3), but I got an error like this:
Here is my test code, I want to load model only once and test multiple images, but I failed,
@JiahuiYu could we build graph once and feed images/masks with different resolutions? Thanks |
@jtetc If possible, please help @ccc013. Thanks! |
@JiahuiYu @jtetc Thanks for your fast response! But, when I set the option
Here is my test code, and I use tensorflow 1.5 version, @jtetc I don't know if my code has some wrong place?
|
@ccc013 I didn't use
I am a green hand in tensorflow. I guess we cannot get the shape of x since input shape is |
You modification should be fine. |
@jtetc Thanks you very much! I change the code like you do, and actually I also change other code which contains particularly,I change the line
to
If I don't change this code, I stil got an error. And my test code works after changing above function codes! Hi, @JiahuiYu is my change above code in |
@ccc013 I have same modifications. They are same issues. |
@ccc013 FYI
|
@jtetc Thanks for your contribution! |
@jtetc Thanks! |
@jtetc I have trained my model already. If I want build graph once when testing, do I need to retrain the model? |
@awsssix If you didn't modify the model, you don't need to retrain. |
@jtetc Thanks for your replay. |
You can take a try |
Really great works your deepfill!
Is it possible to build graph once then feed it with multi images/masks with different resolutions?
Similar to #12 (comment)
a loop was created to process multi images but he built graph for every input image.
In your code #12 (comment) graph only has to be built once but all image/mask should be resized to same size then fed to the graph.
Can we build graph once and feed images/masks with different resolutions?
Thanks.
The text was updated successfully, but these errors were encountered: