picasso icon indicating copy to clipboard operation
picasso copied to clipboard

specifying multiple feeds in BACKEND_TF_INPUT_VAR

Open kumarsameer opened this issue 8 years ago • 1 comments

  • picasso version: 0.1.1
  • Python version: 3.5.2
  • Operating System: ubuntu 16.04

Description

The documentation specifies that to use tensorflow backend, we need to specify the feeds via BACKEND_TF_INPUT_VAR variable. But how do I specify multiple inputs ?

What I Did

will the following work ?

BACKEND_TF_INPUT_VAR='convolution2d_input_1:0'
BACKEND_TF_INPUT_VAR='convolution2d_input_2:0'

kumarsameer avatar May 30 '17 14:05 kumarsameer

It's not supported yet, since the web frontend only supports uploading single images. We would like to add this, however. In the meantime, is it possible to fix your other inputs?

@jwayne have you encountered this problem modifying picasso for other models?

rhsimplex avatar May 30 '17 16:05 rhsimplex