[Relay tests] AlterOpLayout - Temporary attr update (#4357)

include/tvm/relay/op.h

@@ -258,6 +258,12 @@ class OpRegistry {
   inline OpRegistry& set_attr(const std::string& attr_name,  // NOLINT(*)
                               const ValueType& value, int plevel = 10);
 
+  /*!
+   * \brief Resets an attr of the registry.
+   * \param attr_name The name of the attribute.
+   */
+  inline void reset_attr(const std::string& attr_name);
+
   // set the name of the op to be the same as registry
   inline OpRegistry& set_name() {  // NOLINT(*)
     if (get()->name.length() == 0) {

python/tvm/relay/op/op.py

@@ -64,6 +64,16 @@ class Op(Expr):
         """
         _OpSetAttr(self, attr_name, value, plevel)
 
+    def reset_attr(self, attr_name):
+        """Reset attribute about the operator.
+
+        Parameters
+        ----------
+        attr_name : str
+            The attribute name
+        """
+        _OpResetAttr(self, attr_name)
+
 
 def get(op_name):
     """Get the Op for a given name

python/tvm/relay/testing/__init__.py

@@ -37,6 +37,7 @@ from . import squeezenet
 from . import vgg
 from . import densenet
 from . import yolo_detection
+from . import temp_op_attr
 
 from .config import ctx_list
 from .init import create_workload

python/tvm/relay/testing/temp_op_attr.py

+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+""" Defines a TempOpAttr class that allows temporarily changing an attr of the
+operator to allow unit testing. This is useful for AlterOpLayout and Legalize
+tests."""
+
+from tvm import relay
+
+class TempOpAttr(object):
+    """ Temporarily changes the attr of an op. """
+    def __init__(self, op_name, attr_key, attr_value):
+        """ Saves the required info for RAII pattern usage.
+
+        Parameters
+        ----------
+        op_name : str
+            The op name.
+
+        attr_key : str
+            The attribute name.
+
+        attr_value : object
+            The attribute value.
+
+        Examples
+        --------
+        .. code-block:: python
+
+        # Temporarily update FTVMAlterOpLayout to a user-defined packed function.
+        # After the test is finished, the attr value will be set back to the original value.
+
+        with TempOpAttr("nn.conv2d", "FTVMAlterOpLayout", alter_conv2d):
+            my_mod = relay.transform.AlterOpLayout()(my_mod)
+
+        """
+        self.op = relay.op.get(op_name)
+        self.attr_key = attr_key
+        self.attr_value = attr_value
+
+    def __enter__(self):
+        self.older_attr = self.op.get_attr(self.attr_key)
+        self.op.reset_attr(self.attr_key)
+        self.op.set_attr(self.attr_key, self.attr_value)
+        return self
+
+    def __exit__(self, ptype, value, trace):
+        self.op.reset_attr(self.attr_key)
+        if self.older_attr:
+            self.op.set_attr(self.attr_key, self.older_attr)

src/relay/ir/op.cc

@@ -95,6 +95,20 @@ const bool Op::HasGenericAttr(const std::string& key) {
   return true;
 }
 
+// Resets attr of the OpMap.
+void OpRegistry::reset_attr(const std::string& key) {
+  OpManager* mgr = OpManager::Global();
+  std::lock_guard<std::mutex> lock(mgr->mutex);
+  std::unique_ptr<GenericOpMap>& op_map = mgr->attr[key];
+  if (op_map == nullptr) {
+    return;
+  }
+  uint32_t index = op_->index_;
+  if (op_map->data_.size() > index) {
+    op_map->data_[index] = std::make_pair(TVMRetValue(), 0);
+  }
+}
+
 void OpRegistry::UpdateAttr(const std::string& key,
                             TVMRetValue value,
                             int plevel) {
@@ -113,7 +127,10 @@ void OpRegistry::UpdateAttr(const std::string& key,
   CHECK(p.second != plevel)
       << "Attribute " << key << " of operator " << this->name
       << " is already registered with same plevel=" << plevel;
-  if (p.second < plevel) {
+  CHECK(value.type_code() != kNull)
+      << "Registered packed_func is Null for " << key
+      << " of operator " << this->name;
+  if (p.second < plevel && value.type_code() != kNull) {
     op_map->data_[index] = std::make_pair(value, plevel);
   }
 }
@@ -152,6 +169,15 @@ TVM_REGISTER_API("relay.op._OpSetAttr")
     reg.set_attr(attr_name, value, plevel);
   });
 
+TVM_REGISTER_API("relay.op._OpResetAttr")
+.set_body([](TVMArgs args, TVMRetValue* rv) {
+    Op op = args[0];
+    std::string attr_name = args[1];
+    auto& reg =
+        OpRegistry::Registry()->__REGISTER_OR_GET__(op->name);
+    reg.reset_attr(attr_name);
+  });
+
 TVM_REGISTER_API("relay.op._Register")
 .set_body([](TVMArgs args, TVMRetValue* rv) {
     std::string op_name = args[0];

tests/python/relay/test_ir_op.py

@@ -15,6 +15,7 @@
 # specific language governing permissions and limitations
 # under the License.
 from tvm import relay
+from tvm.relay.testing.temp_op_attr import TempOpAttr
 
 def test_op_attr():
     log_op = relay.op.get("log")
@@ -27,6 +28,50 @@ def test_op_attr():
     assert log_op.get_attr("ftest") is None
     assert relay.op.get("exp").get_attr("ftest")(1) == 2
 
+def test_op_reset_attr():
+    """ Tests reset_attr functionality. """
+    def add1(x):
+        return x + 1
+
+    def add2(x):
+        return x + 2
+
+    # Register fadd1 and fadd2 attributes.
+    relay.op.register("exp", "fadd1", add1)
+    relay.op.register("log", "fadd1", add1)
+    relay.op.register("log", "fadd2", add2)
+
+    # Reset log fadd1 attr.
+    log_op = relay.op.get("log")
+    log_op.reset_attr("fadd1")
+
+    # Check that fadd1 attr is resetted.
+    assert log_op.get_attr("fadd1") is None
+
+    # Check that fadd1 attr of other ops are intact.
+    assert relay.op.get("exp").get_attr("fadd1")(1) == 2
+
+    # Check that other attrs of the log op are intact.
+    assert relay.op.get("log").get_attr("fadd2")(1) == 3
+
+def test_op_temp_attr():
+    """ Tests reset_attr functionality. """
+    def add1(x):
+        return x + 1
+
+    def add2(x):
+        return x + 2
+
+    # Set original attr value is add1.
+    relay.op.register("sqrt", "ftest", add1)
+
+    with TempOpAttr("sqrt", "ftest", add2):
+        # Check that the attr value is updated to add2.
+        assert relay.op.get("sqrt").get_attr("ftest")(1) == 3
+
+    # Check that the attr value is recovered to add1.
+    assert relay.op.get("sqrt").get_attr("ftest")(1) == 2
+
 def test_op_level1():
     x = relay.Var("x")
 
@@ -47,5 +92,7 @@ def test_op_level3():
 
 if __name__ == "__main__":
     test_op_attr()
+    test_op_reset_attr()
+    test_op_temp_attr()
     test_op_level1()
     test_op_level3()

tests/python/relay/test_op_qnn_conv2d.py

@@ -21,6 +21,14 @@ from tvm import relay
 from tvm.relay import transform
 from tvm.relay.testing import run_infer_type
 from tvm.contrib import graph_runtime
+from tvm.relay.testing.temp_op_attr import TempOpAttr
+
+# We use llvm target for testing functionality. `llvm` points to an older Intel
+# generation machine, that legalizes to a simple lowering. Therefore, the
+# legalization is overwritten such that it can be skipped and we use the
+# QNNCanonicalizeOps lowering for the testing.
+def legalize_qnn_conv2d(attrs, inputs, types):
+    return None
 
 def get_ref_func(data,
                  kernel,
@@ -173,6 +181,8 @@ def verify(ref_func, qnn_func, data_shape, data_dtype, kernel_shape,
     np.testing.assert_equal(qnn_output, golden_output)
 
 def test_no_zero_point():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 1, 2, 4)
         data_dtype = 'uint8'
@@ -220,6 +230,8 @@ def test_no_zero_point():
                 kernel_shape, kernel_dtype)
 
 def test_kernel_zero_point():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 4, 2, 4)
         data_dtype = 'uint8'
@@ -268,6 +280,8 @@ def test_kernel_zero_point():
 
 
 def test_input_zero_point():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 4, 2, 4)
         data_dtype = 'uint8'
@@ -315,6 +329,8 @@ def test_input_zero_point():
                 kernel_shape, kernel_dtype)
 
 def test_both_zero_point():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 4, 2, 4)
         data_dtype = 'uint8'
@@ -362,6 +378,8 @@ def test_both_zero_point():
                 kernel_shape, kernel_dtype)
 
 def test_layout():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 2, 4, 4) # NHWC
         data_dtype = 'uint8'
@@ -411,6 +429,8 @@ def test_layout():
 
 
 def test_padding():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (1, 4, 2, 2)
         data_dtype = 'uint8'
@@ -458,6 +478,8 @@ def test_padding():
                 kernel_shape, kernel_dtype)
 
 def test_dilation():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 4, 4, 4)
         data_dtype = 'uint8'
@@ -483,6 +505,8 @@ def test_dilation():
 
 
 def test_const_folding():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         data_shape = (2, 4, 2, 4)
         data_dtype = 'uint8'
         kernel_shape = (3, 4, 2, 2)
@@ -511,6 +535,8 @@ def test_const_folding():
         assert "reshape" not in folded_func.astext()
 
 def test_kernel_size_1x1():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 4, 2, 4)
         data_dtype = 'uint8'
@@ -536,6 +562,8 @@ def test_kernel_size_1x1():
                 kernel_shape, kernel_dtype)
 
 def test_tflite_large_irregular():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (1, 1024, 1, 1)
         data_dtype = 'uint8'
@@ -571,6 +599,8 @@ def test_tflite_large_irregular():
         np.testing.assert_equal(qnn_output, golden_output)
 
 def test_tflite_output_multiplier_greater_than_one():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (2, 1, 2, 4)
         data_dtype = 'uint8'
@@ -617,6 +647,8 @@ def test_tflite_output_multiplier_greater_than_one():
         np.testing.assert_equal(qnn_output, golden_output)
 
 def test_tflite_anistropic_strides():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # uint8 input
         data_shape = (1, 1, 3, 6)
         data_dtype = 'uint8'
@@ -656,6 +688,8 @@ def test_tflite_anistropic_strides():
         np.testing.assert_equal(qnn_output, golden_output)
 
 def test_broadcast_layout():
+    with TempOpAttr("qnn.conv2d", "FTVMQnnLegalize", legalize_qnn_conv2d):
+
         # Test broadcast support for NHWC layout.
         data_shape = (1, 229, 229, 3) # NHWC
         data_dtype = 'uint8'

tests/python/relay/test_op_qnn_dense.py

@@ -19,6 +19,15 @@ import tvm
 import numpy as np
 from tvm import relay
 from tvm.contrib import graph_runtime
+from tvm.relay.testing.temp_op_attr import TempOpAttr
+
+
+# We use llvm target for testing functionality. `llvm` points to an older Intel
+# generation machine, that legalizes to a simple lowering. Therefore, the
+# legalization is overwritten such that it can be skipped and we use the
+# QNNCanonicalizeOps lowering for the testing.
+def legalize_qnn_dense(attrs, inputs, types):
+    return None
 
 
 def make_requantize_params(input_scale, output_scale, output_zero_point, out_dtype):
@@ -209,18 +218,24 @@ def qnn_dense_driver(test_configuration):
 
 
 def test_qnn_dense_without_bias():
+    with TempOpAttr("qnn.dense", "FTVMQnnLegalize", legalize_qnn_dense):
+
         int32_output_without_bias_params = \
             make_int_configuration(use_bias=False)
         qnn_dense_driver(int32_output_without_bias_params)
 
 
 def test_qnn_dense_with_bias():
+    with TempOpAttr("qnn.dense", "FTVMQnnLegalize", legalize_qnn_dense):
+
         int32_output_with_bias_params = \
             make_int_configuration(use_bias=True)
         qnn_dense_driver(int32_output_with_bias_params)
 
 
 def test_qnn_dense_with_requantized_output():
+    with TempOpAttr("qnn.dense", "FTVMQnnLegalize", legalize_qnn_dense):
+
         int8_requantized_output_with_bias_params = \
             make_int_configuration(use_bias=True, requantize_output=True)
         qnn_dense_driver(int8_requantized_output_with_bias_params)

tests/python/relay/test_pass_legalize.py

@@ -20,8 +20,8 @@ import tvm
 
 from tvm import relay
 from tvm.contrib import graph_runtime
-from tvm.relay.op import register_legalize
 from tvm.relay import transform, analysis
+from tvm.relay.testing.temp_op_attr import TempOpAttr
 
 
 def run_opt_pass(expr, passes):
@@ -46,7 +46,6 @@ def test_legalize():
         y = relay.Function([x, weight], y)
         return y
 
-    @register_legalize("nn.conv2d", level=100)
     def legalize_conv2d(attrs, inputs, types):
         data, weight = inputs
         weight = relay.multiply(weight, relay.const(2.0, "float32"))
@@ -63,6 +62,7 @@ def test_legalize():
         y = relay.Function([x, weight], y)
         return y
 
+    with TempOpAttr("nn.conv2d", "FTVMLegalize", legalize_conv2d):
         a = before()
         a = run_opt_pass(a, transform.Legalize())
         b = run_opt_pass(expected(), transform.InferType())
@@ -79,16 +79,15 @@ def test_legalize_none():
 
     called = [False]
 
-    @register_legalize("nn.global_max_pool2d", level=101)
     def legalize_conv2d(attrs, inputs, types):
         called[0] = True
         return None
 
+    with TempOpAttr("nn.global_max_pool2d", "FTVMLegalize", legalize_conv2d):
         a = before()
         a = run_opt_pass(a, transform.Legalize())
+        b = run_opt_pass(before(), transform.InferType())
 
-    b = before()
-    b = run_opt_pass(b, transform.InferType())
     assert analysis.alpha_equal(a, b), "Actual = \n" + str(a)
     assert(called[0])
 
@@ -105,14 +104,12 @@ def test_legalize_multiple_ops():
         y = relay.Function([x, weight], y)
         return y
 
-    @register_legalize("nn.conv2d", level=102)
     def legalize_conv2d(attrs, inputs, types):
         data, weight = inputs
         weight = relay.multiply(weight, relay.const(2.0, "float32"))
         return relay.nn.conv2d(data, weight, **attrs)
 
-    @register_legalize("nn.relu", level=103)
-    def legalize_conv2d(attrs, inputs, types):
+    def legalize_relu(attrs, inputs, types):
         data = inputs[0]
         add = relay.add(tvm.relay.const(0, "float32"), data)
         return relay.nn.relu(add)
@@ -130,6 +127,8 @@ def test_legalize_multiple_ops():
         y = relay.Function([x, weight], y)
         return y
 
+    with TempOpAttr("nn.conv2d", "FTVMLegalize", legalize_conv2d):
+        with TempOpAttr("nn.relu", "FTVMLegalize", legalize_relu):
             a = before()
             a = run_opt_pass(a, transform.Legalize())
             b = run_opt_pass(expected(), transform.InferType())
@@ -147,7 +146,6 @@ def test_legalize_multi_input():
         func = relay.Function([x, y, z], func)
         return func
 
-    @register_legalize("concatenate", level=104)
     def legalize_concatenate(attrs, inputs, types):
         # Check that the correct multi-input case is handled.
         assert len(inputs) == 1
@@ -165,6 +163,8 @@ def test_legalize_multi_input():
         func = relay.Function([x, y, z], func)
         return func
 
+
+    with TempOpAttr("concatenate", "FTVMLegalize", legalize_concatenate):
         a = before()
         a = run_opt_pass(a, transform.Legalize())
         b = run_opt_pass(expected(), transform.InferType())

tests/python/relay/test_pass_qnn_legalize.py

@@ -20,8 +20,8 @@ import tvm
 
 from tvm import relay
 from tvm.contrib import graph_runtime
-from tvm.relay.qnn.op import register_qnn_legalize
 from tvm.relay import transform, analysis
+from tvm.relay.testing.temp_op_attr import TempOpAttr
 
 def alpha_equal(x, y):
     """
@@ -54,7 +54,6 @@ def test_qnn_legalize():
         y = relay.Function([x], y)
         return y
 
-    @register_qnn_legalize("qnn.requantize", level=100)
     def legalize_qnn_requantize(attrs, inputs, types):
         data = inputs[0]
         data = relay.add(relay.const(0, 'int8'), data)
@@ -80,6 +79,8 @@ def test_qnn_legalize():
 
     a = before()
 
+    with TempOpAttr("qnn.requantize", "FTVMQnnLegalize", legalize_qnn_requantize):
+
         # Check that Relay Legalize does not change the graph.
         a = run_opt_pass(a, relay.transform.Legalize())
         b = run_opt_pass(before(), transform.InferType())