Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jan 18 09:28
    ghylander opened #84
  • Nov 15 2020 14:33
    tomassams opened #83
  • Sep 04 2020 00:19
    dependabot[bot] labeled #82
  • Sep 04 2020 00:19
    dependabot[bot] opened #82
  • Sep 04 2020 00:19

    dependabot[bot] on npm_and_yarn

    build(deps): bump node-sass in … (compare)

  • Jul 21 2020 12:36

    dependabot[bot] on npm_and_yarn

    build(deps-dev): bump codecov i… (compare)

  • Jul 21 2020 12:36

    dependabot[bot] on npm_and_yarn

    (compare)

  • Jul 21 2020 12:36
    dependabot[bot] closed #79
  • Jul 21 2020 12:35
    dependabot[bot] labeled #81
  • Jul 21 2020 12:35
    dependabot[bot] opened #81
  • Jun 18 2020 16:00
    Shameendra opened #80
  • Feb 19 2020 17:47
    dependabot[bot] labeled #79
  • Feb 19 2020 17:47
    dependabot[bot] opened #79
  • Feb 19 2020 17:47

    dependabot[bot] on npm_and_yarn

    build(deps-dev): bump codecov i… (compare)

  • Nov 07 2019 12:30
    ashishpatel26 opened #78
  • Sep 20 2019 20:56
    tirthajyoti opened #77
  • Mar 25 2019 04:38
    rushic24 opened #76
  • Feb 14 2019 00:29
    ruslangrimov opened #75
  • Nov 13 2018 20:35
    sudheerExperiments opened #74
  • Oct 26 2018 00:47
    jorgimello opened #73
isaacgerg
@isaacgerg
my quiver is broke right now for my grayscale model.
but, what do is your value for single_input_shape?
it looks like when keras loads an image you must provide this. We need to make sure the corret value is getting in there.
also, are you using theano or tf for backend? This will change the ordering of dimensions.
Tasos Varoudis
@varoudis
hmm I hacked the util.py and I forced greyscale=True on image.load but I still get the "Exception: Error when checking : expected input_1 to have shape (1, 128, 128, 1) but got array with shape (1, 128, 128, 3)"
Jake Bian
@jakebian
@varoudis confused about your last message, if your array has shape (1, 128, 128, 3) it has 3 channels, not 1.
oh I see what you're trying to do
Try grayscale=True
gray with an "a"
isaacgerg
@isaacgerg
any luck with getting a model of model's working?
@jakebian just saw you closed my ticket. i didnt realize the other ticket was the same thing. sorry!
jalFaizy
@faizankshaikh
I'm having issues when loading image into quiver
When I select an image and layer, it still shows "No data for this layer"
in traceback it shows this IOError: [Errno 2] No such file or directory: u'trial.jpg'
How can the image not be found if it is displayed on the home page?
Screenshot (32).png
Tasos Varoudis
@varoudis
@jakebian as I posted on gh, the spelling on the code is correct but I didn't get it to work :S not sure if I was missing something else.
Tasos Varoudis
@varoudis
@jakebian processing by mean the 3 channels in all stages need to change too
planaria158
@planaria158
Started quiver and it appears to launch normally. My model is visible in the left pane. However, no files visible on the right pane (path is correctly set and there are legit image files present). Any thoughts? I'm using Chrome on a Mac; keras 1.1.1; installed quiver via pip. Perhaps its a browser issue?
Jake Bian
@jakebian
@/all Hey guys, thanks for being patient. Made a batch of fixes just now, python 3 issues and file path issues should be fixed. Run a pip inistall --upgrade and keep bothering me if issues persist.
planaria158
@planaria158
Thanks Jake! Just entered a follow up comment to this issue in jakebian/quiver#26
Marouane Ben Romdhane
@marouanebr
Hi Jake! I have the same problem as @planaria158 : the browser does not display any pictures!
Can you please share a demo code, to check any mistakes in mine. Thanks :)
Lalit7Jain
@Lalit7Jain
Hi Jake, There seems to be an issue with the installation of quiver
Can anybody help me here?
RichardAnthoyFalzini
@RichardAnthoyFalzini
hi same problem as @planaria158, non error shown, just updated the package and still no picture visible on right pane. any suggestions?
sudheerExperiments
@sudheerExperiments
@RichardAnthoyFalzini, I am have same issue, any solution?
Any solution for 'No data for this layer' issue??
ghylander
@ghylander
hi all
any activity here?
ghylander
@ghylander
if anyone is interested, you can define get a nested model by using: submodel = training_model.get_layer("model").get_layer("MobilenetV3large")
(in my case, i have a training model built upon an inference model. The latter loads an h5 model. To access the doubly nested model i first have to access the nested inference model and later the collapsed model. Also in my case, the collapsed model is called MobilenetV3large)
once i create this submodel from the nested model, i can do submodel.summary() and the whole model is dispplayed
doing all of this has the advantage that i can load my weights and then visualize the output of my trained ntwork