Repository: systemml Updated Branches: refs/heads/master fb675b82c -> db9da2855
[SYSTEMML-2463] Fix paramserv tests (incorrect named argument usage) With the recently added support for named function arguments various places in SystemML check the validity of used named arguments. This makes the existing paramserv tests fail because they use incorrect name bindings that have been ignored so far. Project: http://git-wip-us.apache.org/repos/asf/systemml/repo Commit: http://git-wip-us.apache.org/repos/asf/systemml/commit/db9da285 Tree: http://git-wip-us.apache.org/repos/asf/systemml/tree/db9da285 Diff: http://git-wip-us.apache.org/repos/asf/systemml/diff/db9da285 Branch: refs/heads/master Commit: db9da28551bd85f234c196ac8fd7ea25cccc8543 Parents: fb675b8 Author: Matthias Boehm <[email protected]> Authored: Wed Jul 25 18:03:57 2018 -0700 Committer: Matthias Boehm <[email protected]> Committed: Wed Jul 25 18:03:57 2018 -0700 ---------------------------------------------------------------------- .../functions/paramserv/mnist_lenet_paramserv.dml | 18 ++++++------------ .../mnist_lenet_paramserv_minimum_version.dml | 18 ++++++------------ 2 files changed, 12 insertions(+), 24 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/systemml/blob/db9da285/src/test/scripts/functions/paramserv/mnist_lenet_paramserv.dml ---------------------------------------------------------------------- diff --git a/src/test/scripts/functions/paramserv/mnist_lenet_paramserv.dml b/src/test/scripts/functions/paramserv/mnist_lenet_paramserv.dml index 84095ec..bce4eea 100644 --- a/src/test/scripts/functions/paramserv/mnist_lenet_paramserv.dml +++ b/src/test/scripts/functions/paramserv/mnist_lenet_paramserv.dml @@ -157,14 +157,12 @@ gradients = function(matrix[double] features, [outc1, Houtc1, Woutc1] = conv2d::forward(features, W1, b1, C, Hin, Win, Hf, Wf, stride, stride, pad, pad) outr1 = relu::forward(outc1) - [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, 2, 2, 2, 2, 0, 0) ## layer 2: conv2 -> relu2 -> pool2 [outc2, Houtc2, Woutc2] = conv2d::forward(outp1, W2, b2, F1, Houtp1, Woutp1, Hf, Wf, stride, stride, pad, pad) outr2 = relu::forward(outc2) - [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, 2, 2, 2, 2, 0, 0) ## layer 3: affine3 -> relu3 -> dropout outa3 = affine::forward(outp2, W3, b3) outr3 = relu::forward(outa3) @@ -184,14 +182,12 @@ gradients = function(matrix[double] features, douta3 = relu::backward(doutr3, outa3) [doutp2, dW3, db3] = affine::backward(douta3, outp2, W3, b3) ## layer 2: conv2 -> relu2 -> pool2 - doutr2 = max_pool2d::backward(doutp2, Houtp2, Woutp2, outr2, F2, Houtc2, Woutc2, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + doutr2 = max_pool2d::backward(doutp2, Houtp2, Woutp2, outr2, F2, Houtc2, Woutc2, 2, 2, 2, 2, 0, 0) doutc2 = relu::backward(doutr2, outc2) [doutp1, dW2, db2] = conv2d::backward(doutc2, Houtc2, Woutc2, outp1, W2, b2, F1, Houtp1, Woutp1, Hf, Wf, stride, stride, pad, pad) ## layer 1: conv1 -> relu1 -> pool1 - doutr1 = max_pool2d::backward(doutp1, Houtp1, Woutp1, outr1, F1, Houtc1, Woutc1, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + doutr1 = max_pool2d::backward(doutp1, Houtp1, Woutp1, outr1, F1, Houtc1, Woutc1, 2, 2, 2, 2, 0, 0) doutc1 = relu::backward(doutr1, outc1) [dX_batch, dW1, db1] = conv2d::backward(doutc1, Houtc1, Woutc1, features, W1, b1, C, Hin, Win, Hf, Wf, stride, stride, pad, pad) @@ -314,14 +310,12 @@ predict = function(matrix[double] X, int C, int Hin, int Win, int batch_size, [outc1, Houtc1, Woutc1] = conv2d::forward(X_batch, W1, b1, C, Hin, Win, Hf, Wf, stride, stride, pad, pad) outr1 = relu::forward(outc1) - [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, 2, 2, 2, 2, 0, 0) ## layer 2: conv2 -> relu2 -> pool2 [outc2, Houtc2, Woutc2] = conv2d::forward(outp1, W2, b2, F1, Houtp1, Woutp1, Hf, Wf, stride, stride, pad, pad) outr2 = relu::forward(outc2) - [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, 2, 2, 2, 2, 0, 0) ## layer 3: affine3 -> relu3 outa3 = affine::forward(outp2, W3, b3) outr3 = relu::forward(outa3) http://git-wip-us.apache.org/repos/asf/systemml/blob/db9da285/src/test/scripts/functions/paramserv/mnist_lenet_paramserv_minimum_version.dml ---------------------------------------------------------------------- diff --git a/src/test/scripts/functions/paramserv/mnist_lenet_paramserv_minimum_version.dml b/src/test/scripts/functions/paramserv/mnist_lenet_paramserv_minimum_version.dml index aeec3df..a3677aa 100644 --- a/src/test/scripts/functions/paramserv/mnist_lenet_paramserv_minimum_version.dml +++ b/src/test/scripts/functions/paramserv/mnist_lenet_paramserv_minimum_version.dml @@ -151,14 +151,12 @@ gradients = function(matrix[double] features, [outc1, Houtc1, Woutc1] = conv2d::forward(features, W1, b1, C, Hin, Win, Hf, Wf, stride, stride, pad, pad) outr1 = relu::forward(outc1) - [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, 2, 2, 2, 2, 0, 0) ## layer 2: conv2 -> relu2 -> pool2 [outc2, Houtc2, Woutc2] = conv2d::forward(outp1, W2, b2, F1, Houtp1, Woutp1, Hf, Wf, stride, stride, pad, pad) outr2 = relu::forward(outc2) - [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, 2, 2, 2, 2, 0, 0) ## layer 3: affine3 -> relu3 -> dropout outa3 = affine::forward(outp2, W3, b3) outr3 = relu::forward(outa3) @@ -178,14 +176,12 @@ gradients = function(matrix[double] features, douta3 = relu::backward(doutr3, outa3) [doutp2, dW3, db3] = affine::backward(douta3, outp2, W3, b3) ## layer 2: conv2 -> relu2 -> pool2 - doutr2 = max_pool2d::backward(doutp2, Houtp2, Woutp2, outr2, F2, Houtc2, Woutc2, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + doutr2 = max_pool2d::backward(doutp2, Houtp2, Woutp2, outr2, F2, Houtc2, Woutc2, 2, 2, 2, 2, 0, 0) doutc2 = relu::backward(doutr2, outc2) [doutp1, dW2, db2] = conv2d::backward(doutc2, Houtc2, Woutc2, outp1, W2, b2, F1, Houtp1, Woutp1, Hf, Wf, stride, stride, pad, pad) ## layer 1: conv1 -> relu1 -> pool1 - doutr1 = max_pool2d::backward(doutp1, Houtp1, Woutp1, outr1, F1, Houtc1, Woutc1, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + doutr1 = max_pool2d::backward(doutp1, Houtp1, Woutp1, outr1, F1, Houtc1, Woutc1, 2, 2, 2, 2, 0, 0) doutc1 = relu::backward(doutr1, outc1) [dX_batch, dW1, db1] = conv2d::backward(doutc1, Houtc1, Woutc1, features, W1, b1, C, Hin, Win, Hf, Wf, stride, stride, pad, pad) @@ -307,14 +303,12 @@ predict = function(matrix[double] X, int C, int Hin, int Win, int batch_size, [outc1, Houtc1, Woutc1] = conv2d::forward(X_batch, W1, b1, C, Hin, Win, Hf, Wf, stride, stride, pad, pad) outr1 = relu::forward(outc1) - [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp1, Houtp1, Woutp1] = max_pool2d::forward(outr1, F1, Houtc1, Woutc1, 2, 2, 2, 2, 0, 0) ## layer 2: conv2 -> relu2 -> pool2 [outc2, Houtc2, Woutc2] = conv2d::forward(outp1, W2, b2, F1, Houtp1, Woutp1, Hf, Wf, stride, stride, pad, pad) outr2 = relu::forward(outc2) - [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, Hf=2, Wf=2, - strideh=2, stridew=2, pad=0, pad=0) + [outp2, Houtp2, Woutp2] = max_pool2d::forward(outr2, F2, Houtc2, Woutc2, 2, 2, 2, 2, 0, 0) ## layer 3: affine3 -> relu3 outa3 = affine::forward(outp2, W3, b3) outr3 = relu::forward(outa3)
