-
Notifications
You must be signed in to change notification settings - Fork 242
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
[tests] Fix bug in weights tensor layout in solver test #1950
Conversation
@xinlipn please add an issue where this problem is found, and add a test :) |
|
Local test seems to be ok. Will try to reproduce on different nodes. 130/136 Test #130: LOG_TEST.AssertLogCmdOutput_Neg ........................................................................................................... Passed 0.08 sec 100% tests passed, 0 tests failed out of 85 |
Tested on gfx908 node and also passed.
133/138 Test #133: LOG_TEST.AssertLogFindCmdOutput_Neg ....................................................................................................... Passed 0.10 sec 138/138 Test #138: ConvFwdTest/ConvFwdSolverTestFloat.ConvASM3x3UFwd/(1,N:64C:1024H:14W:14k:1024y:3x:3pad_y:1pad_x:1stride_y:1dilation_y:1conv_mode:0) ....... Passed 178.18 sec 100% tests passed, 0 tests failed out of 87 Total Test time (real) = 393.76 sec |
@xinlipn Please format logs like this:
When logs are very long, either attach as text files or consult https://github.com/ROCmSoftwarePlatform/MIOpen/wiki/How-to-insert-console-logs-into-github-pages |
@xinlipn Please add |
test/gtest/solver.hpp
Outdated
@@ -99,7 +99,7 @@ struct ConvFwdSolverTest | |||
test_skipped = false; | |||
std::tie(algo, conv_config) = GetParam(); | |||
input = tensor<T>{conv_config.N, conv_config.C, conv_config.H, conv_config.W}; | |||
weights = tensor<T>{1, conv_config.k, conv_config.x, conv_config.y}; | |||
weights = tensor<T>{conv_config.k, conv_config.C, conv_config.x, conv_config.y}; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
weights = tensor<T>{conv_config.k, conv_config.C, conv_config.x, conv_config.y}; | |
weights = tensor<T>{conv_config.k, conv_config.C, conv_config.y, conv_config.x}; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@JehandadKhan done
The original file was renamed. A new PR had been created, reviewed and merged. Thus close this PR |
Weight tensor should be initilized as k, C, y, x instead of k, C, x, y