You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
thanks for the codes. the idea that using ResNet to rectify the results is wonderful!
i have two problems.
in 'correlation 1d layer', have you try to set the "single_direction" -1 ? due to the samples all have the positive disparity. i do the test and find the result does not achieve expectation.
have you try to generate the 'cost_volume' using caffe? i found it is very hard using caffe. if you do a success trial . Please let me know! Thx~
The text was updated successfully, but these errors were encountered:
thanks for the codes. the idea that using ResNet to rectify the results is wonderful!
i have two problems.
The text was updated successfully, but these errors were encountered: