Add support for special test cases.

theano/gpuarray/tests/check_dnn.py

@@ -40,6 +40,79 @@ def ifilter(function, sequence):
     return (element for element in sequence if function(element))
 
 
+class DnnCase:
+    """
+    Help class to generate special test cases quickly.
+
+    """
+
+    def __init__(self,
+                 type, inputs_shape, filters_shape,
+                 algo=None, dtype=None, precision=None,
+                 subsample=None, dilation=None, border_mode='valid',
+                 conv_mode='conv', alpha=1, beta=0,
+                 should_fail=False):
+        assert type in ('fwd', 'bwd-filter', 'bwd-data')
+        assert len(inputs_shape) == len(filters_shape) > 2
+        ndim = len(inputs_shape) - 2
+        if dtype is None:
+            dtype = theano.config.floatX
+        if precision is None:
+            precision = theano.config.floatX
+        if subsample is None:
+            subsample = (1,) * ndim
+        if dilation is None:
+            dilation = (1,) * ndim
+        assert dtype in ('float16', 'float32', 'float64')
+        assert precision in ('float16', 'float32', 'float64')
+        assert len(subsample) == len(dilation) == ndim
+        assert border_mode in ('valid', 'full', 'half') or len(border_mode) == ndim
+        assert conv_mode in ('conv', 'cross')
+        assert alpha != 0
+
+        self.type = type
+        self.ndim = ndim
+        self.algo = algo
+        self.inputs_shape = inputs_shape
+        self.filters_shape = filters_shape
+        self.dtype = dtype
+        self.precision = precision
+        self.subsample = subsample
+        self.dilation = dilation
+        self.border_mode = border_mode
+        self.conv_mode = conv_mode
+        self.alpha = alpha
+        self.beta = beta
+        self.should_fail = bool(should_fail)
+
+    def is_fwd(self):
+        return self.type == 'fwd'
+
+    def is_bwd_filter(self):
+        return self.type == 'bwd-filter'
+
+    def is_bwd_data(self):
+        return self.type == 'bwd-data'
+
+    def get_case(self):
+        return (self.algo, self.dtype, self.precision,
+                (self.inputs_shape, self.filters_shape,
+                 self.subsample, self.dilation, self.border_mode,
+                 self.conv_mode, self.alpha, self.beta))
+
+    @staticmethod
+    def fwd(*args, **kwargs):
+        return DnnCase('fwd', *args, **kwargs)
+
+    @staticmethod
+    def bwd_filter(*args, **kwargs):
+        return DnnCase('bwd-filter', *args, **kwargs)
+
+    @staticmethod
+    def bwd_data(*args, **kwargs):
+        return DnnCase('bwd-data', *args, **kwargs)
+
+
 class DnnCaseGenerator:
     """
     Main class used to generate test cases.
@@ -261,6 +334,8 @@ class BaseTestDnnConv(object):
     cpu_gradinput_class = None
     cpu_gradweight_class = None
 
+    special_cases = []  # List of DnnCases.
+
     # Utility methods.
 
     def get_cases(self):
@@ -455,7 +530,16 @@ class BaseTestDnnConv(object):
             algos = (algo for algo in self.bwd_filter_algorithms
                      if cudnn.bwd_filter_algo_supports_dtype_config(algo, dtype, precision, self.ndim))
             count_contexts += sum(1 for algo in algos) + len(SUPPORTED_DNN_CONV_ALGO_RUNTIME)
-        return len_cases * count_contexts
+        return len(self.special_cases) + len_cases * count_contexts
+
+    def should_fail(self, callable, *args):
+        try:
+            print('(should fail)', file=sys.stderr, end=' ')
+            callable(*args)
+        except Exception:
+            pass
+        else:
+            raise AssertionError('Should fail', callable.__name__, *args)
 
     # Iterable test methods.
 
@@ -466,6 +550,12 @@ class BaseTestDnnConv(object):
             for algo in chain(algos, SUPPORTED_DNN_CONV_ALGO_RUNTIME):
                 for parameters in self.get_cases():
                     yield (self.run_conv_fwd, algo, dtype, precision, parameters)
+        for dnn_case in self.special_cases:
+            if dnn_case.is_fwd():
+                if dnn_case.should_fail:
+                    yield (self.should_fail, self.run_conv_fwd,) + dnn_case.get_case()
+                else:
+                    yield (self.run_conv_fwd,) + dnn_case.get_case()
 
     def test_gradinput(self):
         for dtype, precision in cudnn.get_bwd_data_dtype_configs():
@@ -474,6 +564,12 @@ class BaseTestDnnConv(object):
             for algo in chain(algos, SUPPORTED_DNN_CONV_ALGO_RUNTIME):
                 for parameters in self.get_cases():
                     yield (self.run_conv_gradinput, algo, dtype, precision, parameters)
+        for dnn_case in self.special_cases:
+            if dnn_case.is_bwd_data():
+                if dnn_case.should_fail:
+                    yield (self.should_fail, self.run_conv_gradinput,) + dnn_case.get_case()
+                else:
+                    yield (self.run_conv_gradinput,) + dnn_case.get_case()
 
     def test_gradweight(self):
         for dtype, precision in cudnn.get_bwd_filter_dtype_configs():
@@ -482,6 +578,12 @@ class BaseTestDnnConv(object):
             for algo in chain(algos, SUPPORTED_DNN_CONV_ALGO_RUNTIME):
                 for parameters in self.get_cases():
                     yield (self.run_conv_gradweight, algo, dtype, precision, parameters)
+        for dnn_case in self.special_cases:
+            if dnn_case.is_bwd_filter():
+                if dnn_case.should_fail:
+                    yield (self.should_fail, self.run_conv_gradweight,) + dnn_case.get_case()
+                else:
+                    yield (self.run_conv_gradweight,) + dnn_case.get_case()
 
 
 class TestDnnConv2D(BaseTestDnnConv):
@@ -495,6 +597,10 @@ class TestDnnConv2D(BaseTestDnnConv):
     cpu_gradinput_class = theano.tensor.nnet.corr.CorrMM_gradInputs
     cpu_gradweight_class = theano.tensor.nnet.corr.CorrMM_gradWeights
 
+    special_cases = [DnnCase.bwd_filter(algo='deterministic', dtype='float32', precision='float32',
+                                        inputs_shape=(1, 1, 541211, 10), filters_shape=(50, 1, 3, 10),
+                                        border_mode=(1, 0), should_fail=True)]
+
 
 class TestDnnConv3D(BaseTestDnnConv):
     ndim = 3