def RunningAllreduceWithGPUs(self, gpu_ids, allreduce_function):
"""A base function to test different scenarios."""
workspace.ResetWorkspace()
net = core.Net("mujitest")
for id in gpu_ids:
net.ConstantFill([],
"testblob_gpu_" + str(id),
shape=[1, 2, 3, 4],
value=float(id + 1),
device_option=muji.OnGPU(id))
allreduce_function(net, ["testblob_gpu_" + str(i) for i in gpu_ids],
"_reduced", gpu_ids)
workspace.RunNetOnce(net)
target_value = sum(gpu_ids) + len(gpu_ids)
all_blobs = workspace.Blobs()
all_blobs.sort()
for blob in all_blobs:
print blob, workspace.FetchBlob(blob)
for id in gpu_ids:
blob = workspace.FetchBlob("testblob_gpu_" + str(i) + "_reduced")
np.testing.assert_array_equal(blob,
target_value,
err_msg="gpu id %d of %s" %
(id, str(gpu_ids)))
def _testMiniAlexNet(self, order):
# First, we get all the random initialization of parameters.
model = self._MiniAlexNetNoDropout(order)
workspace.ResetWorkspace()
workspace.RunNetOnce(model.param_init_net)
inputs = dict([(str(name), workspace.FetchBlob(str(name)))
for name in model.params])
if order == "NCHW":
inputs["data"] = np.random.rand(4, 3, 227, 227).astype(np.float32)
else:
inputs["data"] = np.random.rand(4, 227, 227, 3).astype(np.float32)
inputs["label"] = np.array([1, 2, 3, 4]).astype(np.int32)
cpu_device = caffe2_pb2.DeviceOption()
cpu_device.device_type = caffe2_pb2.CPU
gpu_device = caffe2_pb2.DeviceOption()
gpu_device.device_type = caffe2_pb2.CUDA
checker = device_checker.DeviceChecker(1e-5, [cpu_device, gpu_device])
ret = checker.CheckNet(
model.net.Proto(),
inputs,
# The indices sometimes may be sensitive to small numerical differences
# in the input, so we ignore checking them.
ignore=['_pool1_idx', '_pool2_idx', '_pool5_idx'])
self.assertEqual(ret, True)
def testMNISTNetworks(self):
# First, we get all the random initialization of parameters.
init_net, train_net = self._MNISTNetworks()
workspace.ResetWorkspace()
workspace.RunNetOnce(init_net)
inputs = dict([(str(name), workspace.FetchBlob(str(name)))
for name in workspace.Blobs()])
cpu_device = caffe2_pb2.DeviceOption()
cpu_device.device_type = caffe2_pb2.CPU
gpu_device = caffe2_pb2.DeviceOption()
gpu_device.device_type = caffe2_pb2.CUDA
checker = device_checker.DeviceChecker(1e-2, [cpu_device, gpu_device])
ret = checker.CheckNet(train_net.Proto(), inputs)
self.assertEqual(ret, True)
def CheckSimple(self, op, inputs, outputs_to_check):
"""Checks the operator in a very simple fashion by stacking a sum of squares
on the top.
Inputs:
op: the operator to be checked.
inputs: the input data in numpy arrays.
input_to_check: an index specifying which input blob we should
check.
outputs_with_grads: indices specifying which output blobs will we
need to check gradients with. For these outputs, we will collect a
squared sum and also feed in their gradients.
grad_operator: the gradient operator. If not given, we will get the
gradient operator from the gradient registry.
Outputs:
boolean: True if it passes, False if it does not pass.
"""
# Entering the checker workspace
old_ws_name = workspace.CurrentWorkspace()
results = []
workspace.SwitchWorkspace("_device_check_", True)
for i, device_option in enumerate(self._device_options):
for i, arr in enumerate(inputs):
workspace.FeedBlob(op.inputs[i], arr, device_option)
op.device_option.CopyFrom(device_option)
workspace.RunOperatorOnce(op)
results.append([
workspace.FetchBlob(op.outputs[idx])
for idx in outputs_to_check
])
# Everything is done, reset the workspace.
workspace.ResetWorkspace()
# After running on all devices, check correctness
success = True
for i in range(1, len(self._device_options)):
for j in range(len(outputs_to_check)):
x = results[i][j]
y = results[0][j]
if np.any(np.abs(x - y) > self._threshold):
print 'Failure in checking device option', i, 'and output ',
print op.outputs[j], '. The outputs are:'
print x.flatten()
print y.flatten()
success = False
continue
workspace.SwitchWorkspace(old_ws_name)
return success
def CheckSimple(self, op, inputs, outputs_to_check):
"""Checks the operator with different device implementations.
Inputs:
op: the operator to be checked.
inputs: the input data in numpy arrays.
outputs_to_check: the outputs to check between devices.
Outputs:
boolean: True if it passes, False if it does not pass.
"""
# Entering the checker workspace
old_ws_name = workspace.CurrentWorkspace()
results = []
workspace.SwitchWorkspace("_device_check_", True)
for i, device_option in enumerate(self._device_options):
for i, arr in enumerate(inputs):
workspace.FeedBlob(op.input[i], arr, device_option)
op.device_option.CopyFrom(device_option)
workspace.RunOperatorOnce(op)
results.append([
workspace.FetchBlob(op.output[idx]) for idx in outputs_to_check
])
# Everything is done, reset the workspace.
workspace.ResetWorkspace()
# After running on all devices, check correctness
success = True
for i in range(1, len(self._device_options)):
for j in range(len(outputs_to_check)):
x = results[i][j]
y = results[0][j]
if np.any(np.abs(x - y) > self._threshold):
print 'Failure in checking device option', i, 'and output ',
print op.output[j], '. The outputs are:'
print x.flatten()
print y.flatten()
success = False
#else:
# print ('Passed device pair (0, %d), %s %s' %
# (i, outputs_to_check[j], y.shape))
workspace.SwitchWorkspace(old_ws_name)
return success
def setUp(self):
workspace.SwitchWorkspace("default")
workspace.ResetWorkspace()
def setUp(self):
self.net = core.Net("test-net")
self.net.ConstantFill([], "testblob", shape=[1, 2, 3, 4], value=1.0)
workspace.ResetWorkspace()
def testResetWorkspace(self):
self.assertEqual(workspace.RunNetOnce(self.net.Proto().SerializeToString()), True)
self.assertEqual(workspace.HasBlob("testblob"), True)
self.assertEqual(workspace.ResetWorkspace(), True)
self.assertEqual(workspace.HasBlob("testblob"), False)
def testRootFolder(self):
self.assertEqual(workspace.ResetWorkspace(), True)
self.assertEqual(workspace.RootFolder(), ".")
self.assertEqual(workspace.ResetWorkspace("/home/test"), True)
self.assertEqual(workspace.RootFolder(), "/home/test")
def CheckSimple(self,
op,
inputs,
input_to_check,
outputs_with_grads,
grad_ops=None):
"""Checks the operator in a very simple fashion by stacking a sum of squares
on the top.
Inputs:
op: the operator to be checked.
inputs: the input data in numpy arrays.
input_to_check: an index specifying which input blob we should
check.
outputs_with_grads: indices specifying which output blobs will we
need to check gradients with. For these outputs, we will collect a
squared sum and also feed in their gradients.
grad_operator: the gradient operator. If not given, we will get the
gradient operator from the gradient registry.
Outputs:
boolean: True if it passes, False if it does not pass.
"""
# Entering the checker workspace
old_ws_name = workspace.CurrentWorkspace()
if self._workspace_name != old_ws_name:
workspace.SwitchWorkspace(self._workspace_name, True)
op.device_option.CopyFrom(self._device_option)
if grad_ops is None:
grad_ops = core.GradientRegistry.GetGradient(op)
dims_to_check = inputs[input_to_check].size
# First, feed in the input.
for i, arr in enumerate(inputs):
workspace.FeedBlob(op.inputs[i], arr, self._device_option)
# Get the loss and gradient for the original.
input_name = op.inputs[input_to_check]
loss, grad = self.GetLossAndGrad(op, grad_ops, inputs[input_to_check],
input_name, outputs_with_grads)
grad_estimate = np.zeros_like(inputs[input_to_check])
for current_dim in range(dims_to_check):
# Positive gradient
inputs[input_to_check].flat[current_dim] += self._stepsize
pos_loss, _ = self.GetLossAndGrad(op, grad_ops,
inputs[input_to_check],
input_name, outputs_with_grads)
# Negative gradient
inputs[input_to_check].flat[current_dim] -= self._stepsize * 2
neg_loss, _ = self.GetLossAndGrad(op, grad_ops,
inputs[input_to_check],
input_name, outputs_with_grads)
# Recover the value
inputs[input_to_check].flat[current_dim] += self._stepsize
grad_estimate.flat[current_dim] = (pos_loss -
neg_loss) / self._stepsize / 2
# Now, check correctness
scale = np.maximum(np.maximum(np.abs(grad), np.abs(grad_estimate)), 1)
fail_mat = (np.abs(grad - grad_estimate) > scale * self._threshold)
if np.any(fail_mat):
idx = np.flatnonzero(fail_mat)
#print 'Failed. [idx, grad, grad_estimate] are:'
#print np.vstack([idx, grad.flat[idx], grad_estimate.flat[idx]]).T
ret = False
else:
ret = True
# After finishing, cleaning up things.
if self._workspace_name != old_ws_name:
# We reset the workspace to make sure everything intermediate is cleaned
# up. Note that there is no need to delete a workspace - when empty it
# takes a very limited amount of memory.
workspace.ResetWorkspace()
workspace.SwitchWorkspace(old_ws_name)
return ret, grad, grad_estimate
def testRootFolder(self):
self.assertEqual(workspace.ResetWorkspace(), True)
self.assertEqual(workspace.RootFolder(), ".")
self.assertEqual(workspace.ResetWorkspace("/tmp/caffe-workspace-test"),
True)
self.assertEqual(workspace.RootFolder(), "/tmp/caffe-workspace-test")