-
Notifications
You must be signed in to change notification settings - Fork 70
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
Add way to do cross check implementations #133
Comments
Make tensorflow operations for each layer:
Networks to use:
Add more for other layers. |
We checked bullet items as below with tensorflow. |
@jijoongmoon conv2d backward verification is still to be done as per https://github.com/nnstreamer/nntrainer/blob/master/test/unittest/unittest_nntrainer_layers.cpp#L949 |
Adam validation is done with #376 |
Can this be closed now? |
Once MNIST comparison with tensorflow is done, add this to a test which is run in every PR. |
Update tensorflow training example for mnist application This with same initialization as nntrainer (shown with zero initialization) matches the final accuracy as well as loss with nntrainer See also nnstreamer#133 **Self evaluation:** 1. Build test: [x]Passed [ ]Failed [ ]Skipped 2. Run test: [x]Passed [ ]Failed [ ]Skipped Signed-off-by: Parichay Kapoor <pk.kapoor@samsung.com>
Update tensorflow training example for mnist application This with same initialization as nntrainer (shown with zero initialization) matches the final accuracy as well as loss with nntrainer See also nnstreamer#133 **Self evaluation:** 1. Build test: [x]Passed [ ]Failed [ ]Skipped 2. Run test: [x]Passed [ ]Failed [ ]Skipped Signed-off-by: Parichay Kapoor <pk.kapoor@samsung.com>
We are close this issues. We reopen or make new issue after implementing Transpose convolution. |
Update tensorflow training example for mnist application This with same initialization as nntrainer (shown with zero initialization) matches the final accuracy as well as loss with nntrainer See also nnstreamer#133 **Self evaluation:** 1. Build test: [x]Passed [ ]Failed [ ]Skipped 2. Run test: [x]Passed [ ]Failed [ ]Skipped Signed-off-by: Parichay Kapoor <pk.kapoor@samsung.com>
Update tensorflow training example for mnist application This with same initialization as nntrainer (shown with zero initialization) matches the final accuracy as well as loss with nntrainer See also #133 **Self evaluation:** 1. Build test: [x]Passed [ ]Failed [ ]Skipped 2. Run test: [x]Passed [ ]Failed [ ]Skipped Signed-off-by: Parichay Kapoor <pk.kapoor@samsung.com>
Currently there is limited way to check if the current implementation is logically correct. (Aside from it is running and have pretty good result.)
Cross checking the implementation with other frameworks is needed.
The text was updated successfully, but these errors were encountered: