Skip to content
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 specify the validation data during training and output test accuracy #30

Open
yghlc opened this issue Apr 24, 2017 · 1 comment

Comments

@yghlc
Copy link

yghlc commented Apr 24, 2017

Hello,
There is no test net accuracy in the log file even I specified the test data layer in the train.prototxt file:

layers {
  name: "data"
  type: IMAGE_SEG_DATA
  top: "data"
  top: "label"
  image_data_param {
    root_folder: ""
    source: "./list/test_aug.txt"
    label_type: PIXEL
    batch_size: 100
  }
  transform_param {
    mean_value: 75.209
    mean_value: 85.950
    mean_value: 95.685
    crop_size: 663
    mirror: false
  }
  include: { phase: TEST }
}

Only loss and train net accuracy in the log file, such as:

I0424 09:32:20.303715 32660 solver.cpp:209] Iteration 50, loss = 0.270298
I0424 09:32:20.303766 32660 solver.cpp:224]     Train net output #0: accuracy = 0.883158
I0424 09:32:20.303774 32660 solver.cpp:224]     Train net output #1: accuracy = 0.780627
I0424 09:32:20.303809 32660 solver.cpp:224]     Train net output #2: accuracy = 0.70205

How do I get the test accuracy during training?
What the meaning of Train net output #0, Train net output "#1", Train net output "#2"? "#0", "#1", "#2" means something?

Thanks for any suggestion!

@mwangaz
Copy link

mwangaz commented Feb 24, 2019

Hey. Did you figure out how to get the validation accuracy or loss? I also want to do that.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants