I'm using tensorboard to visualize the image(CelebA) generated by dcgan
Specifically, I created a writer and image summary with:
tf.summary.image('generated', image_output)
summary_op = tf.summary.merge_all()
writer = tf.summary.FileWriter(logdir, graph)
summary = sess.run(summary_op)
after each 100 step I would add a image summary with:
writer.add_summary(image, step)
I think the event file of tensorboard save all the images generated at each step since the event file keeps growing larger. But when I run tensorboard I can only see the latest image.
Is there any way to see the former images? Or they are not saved in the event file and I can't see them.
The issue that you have raised was common enough to warrant a feature request several months ago that was rolled into TensorFlow 1.1.0.
A small sliding bar appears below each image summary in Tensorboard with which you can scroll through the summary steps if you upgrade to TensorFlow 1.1.0+.
Related
There are 2 things i want to clarify
Why is it that when i open tensorboard for eval file after training using object detection api, i see 2 of the same test image side by side? Does it mean that each picture goes through detection twice? I have attached an example below.
Test image
I only can view 9 test images in tensorboard but I have 16 test images. Is there a way i can view all 16 test images?
You see two images side by side. The right image is your labeled image as groundtruth. The left image is your prediction result. Why only 9 test images is visible I can't tell you for now.
The ILSVRC 2012 small classification dataset is not separated by folder and don't have a labels file. How get the labels for the training set?
I tried on nonpub downloads page but does not exist anymore, and i tried by the filenames but their don't have the synset id on it.
I've been having the same issue today following this tutorial on reproducing ImageNet Validation results. I think I've found an answer, even if partial
In the article they point out to this link to get the validation set for object detection. I downloaded it and had the same issue as yourself, it only contains images without labels. What I've found is that this same website had this other link for the bounding boxes. I've downloaded it and alongside with the bboxes it comes with the proper class for each image
Hope this helps!
I am new to object detection.
Can I save the output images of my tensorflow object detection API in a folder after detection is completed in the Jupyter notebook?
It was intuitive to right click and save the images when detecting a few images, but it is impractical to do same when I have over 1000 images to run detection on.
is the output already stored in a subfolder that I am unaware of??
Can I store the output in a folder or directory?
Any suggestions will be appreciated.
I have checked differnt means but found nothing helpful
this is what my detection cell looks like
Thanks
Use PIL you can save your image to disk. First you import the module
from PIL import Image
And then you can comment out the plot lines, add the save lines.
#plt.figure(figsize=IMAGE_SIZE)
#plt.imshow(image_np)
#save to same folder as data input
img = Image.fromarray(image_np)
img.save(image_path[:-4]+'output_'+'.png')
I have a simple MNIST model from the tensorflow tutorial. I want to see how the first convolutional layer's filters changes with time. When I use tf.summary.image, only one of the steps is displayed, and the rest is ignored. Is there any way to work this around?
TF does not have videos, but you can generate image at each step, save them in some directory and then create a video from them.
Is it possible to view the images for all steps in the Image viewer of tensorboard?
Only the images corresponding to the last step are shown in the Images tab, though they are accessible changing the index of the the url of the image:
(Changing the 0 for the required step in the url: localhost:6006/data/individualImage?index=0&tag=777_mask_output_output_1%2Fimage%2F0&run=tensorboard_logs&ts=1487721119.455751)
As I understood, an additional slider was added at TensorFlow v. 1.1.0: