[RELAY] CompileEngine update, nn conv2d, fix dense, pool. (#2082)

include/tvm/relay/op_attr_types.h

@@ -72,7 +72,8 @@ using FTVMCompute = runtime::TypedPackedFunc<
  * \return schedule The computation schedule.
  */
 using FTVMSchedule = runtime::TypedPackedFunc<
-  Schedule(const Array<Tensor>& outs,
+  Schedule(const Attrs& attrs,
+           const Array<Tensor>& outs,
            const Target& target)>;
 }  // namespace relay
 }  // namespace tvm

python/tvm/relay/op/_tensor.py

@@ -2,13 +2,8 @@
 """Backend compiler related feature registration"""
 from __future__ import absolute_import
 import topi
-import topi.cuda
-from .op import register_compute, register_schedule, register_pattern, OpPattern
-
-def schedule_injective(outputs, target):
-    """Generic schedule for binary broadcast."""
-    with target:
-        return topi.generic.schedule_injective(outputs)
+from .op import register_compute, register_schedule, register_pattern
+from .op import schedule_injective, OpPattern
 
 schedule_broadcast = schedule_injective
 schedule_elemwise = schedule_injective

python/tvm/relay/op/nn/__init__.py

@@ -2,3 +2,4 @@
 """Neural network related operators."""
 from __future__ import absolute_import as _abs
 from .nn import *
+from . import _nn

python/tvm/relay/op/nn/_nn.py

 #pylint: disable=invalid-name, unused-argument
 """Backend compiler related feature registration"""
-import tvm
 import topi
-from .. import register
+from topi.util import get_const_int, get_const_tuple
+from .. import op as reg
+from ..op import OpPattern, schedule_injective
 
-def dense_compiler(attrs, inputs, output_type):
-    assert len(inputs) == 2
+# dense
+@reg.register_compute("nn.dense")
+def compute_dense(attrs, inputs, out_type, target):
+    """Compute definition of dense"""
     return [topi.nn.dense(inputs[0], inputs[1])]
 
-def dense_schedule(outputs, target):
-    assert len(outputs) == 1
-    return tvm.create_schedule(outputs[0].op)
+@reg.register_schedule("nn.dense")
+def schedule_dense(attrs, outputs, target):
+    """Schedule definition of dense"""
+    with target:
+        return topi.generic.schedule_dense(outputs)
 
-register("nn.dense", "FTVMCompute", dense_compiler)
-register("nn.dense", "FTVMSchedule", dense_schedule)
+reg.register_pattern("nn.dense", reg.OpPattern.OUT_ELEMWISE_FUSABLE)
+
+
+# conv2d
+@reg.register_compute("nn.conv2d")
+def compute_conv2d(attrs, inputs, out_type, target):
+    """Compute definition of conv2d"""
+    padding = get_const_tuple(attrs.padding)
+    strides = get_const_tuple(attrs.strides)
+    dilation = get_const_tuple(attrs.dilation)
+    groups = attrs.groups
+    layout = attrs.data_layout
+    weight_layout = attrs.weight_layout
+    out_dtype = attrs.out_dtype
+    out_dtype = (inputs[0].dtype if (out_dtype == "same" or out_dtype == "")
+                 else out_dtype)
+
+    assert layout in ["NCHW", "NHWC", "NCHW4c"]
+    (dilation_h, dilation_w) = dilation
+    if dilation_h < 1 or dilation_w < 1:
+        raise ValueError("dilation should be positive value")
+
+    if groups == 1:
+        out = topi.nn.conv2d(
+            inputs[0], inputs[1], strides, padding,
+            dilation, layout, out_dtype=out_dtype)
+    elif layout == "NCHW" and \
+         weight_layout == "OIHW" and \
+         get_const_int(inputs[1].shape[0]) == groups and \
+         get_const_int(inputs[1].shape[1]) == 1:
+        out = topi.nn.depthwise_conv2d_nchw(
+            inputs[0], inputs[1], strides, padding, dilation, out_dtype=out_dtype)
+    elif layout == "NHWC" and \
+         kernel_layout == "HWOI" and\
+         get_const_int(inputs[1].shape[2]) == groups and \
+         get_const_int(inputs[1].shape[3]) == 1:
+        out = topi.nn.depthwise_conv2d_nhwc(
+            inputs[0], inputs[1], strides, padding, dilation, out_dtype=out_dtype)
+    else:
+        raise ValueError("not support arbitrary group number for now")
+    return [out]
+
+
+@reg.register_schedule("nn.conv2d")
+def schedule_conv2d(attrs, outs, target):
+    """Schedule definition of conv2d"""
+    groups = attrs.groups
+    layout = attrs.data_layout
+    kernel_layout = attrs.weight_layout
+    with target:
+        if groups == 1 and layout == "NCHW":
+            return topi.generic.schedule_conv2d_nchw(outs)
+        elif groups == 1 and layout == "NCHW4c":
+            return topi.generic.schedule_conv2d_nchw(outs)
+        elif groups == 1 and layout == "NHWC":
+            return topi.generic.schedule_conv2d_nhwc(outs)
+        elif groups != 1:
+            if layout == "NCHW":
+                # TODO(leyuan, merrymercy, Huyuwei): fold depthwise topi into conv2d.
+                return topi.generic.schedule_depthwise_conv2d_nchw(outs)
+            elif layout == "NHWC" and kernel_layout == "HWOI":
+                return topi.generic.schedule_depthwise_conv2d_nhwc(outs)
+    raise ValueError("No compatible schedule")
+
+reg.register_pattern("nn.conv2d", OpPattern.OUT_ELEMWISE_FUSABLE)
+
+
+# conv2d_transpose
+@reg.register_compute("nn.conv2d_transpose")
+def compute_conv2d_transpose(attrs, inputs, out_dtype, target):
+    """Compute definition of conv2d_transpose"""
+    padding = get_const_tuple(attrs.padding)
+    strides = get_const_tuple(attrs.strides)
+    dilation = get_const_tuple(attrs.dilation)
+    groups = attrs.groups
+    layout = attrs.data_layout
+    out_dtype = attrs.out_dtype
+    out_dtype = (inputs[0].dtype if (out_dtype == "same" or out_dtype == "")
+                 else out_dtype)
+    assert layout == "NCHW", "only support nchw for now"
+    assert dilation == (1, 1), "not support dilate now"
+    assert groups == 1, "only support groups == 1 for now"
+    out = topi.nn.conv2d_transpose_nchw(inputs[0], inputs[1], strides, padding, out_dtype)
+    output_padding = get_const_tuple(attrs.output_padding)
+    out = topi.nn.pad(out,
+                      [0, 0, 0, 0], [0, 0, output_padding[0], output_padding[1]])
+    return [out]
+
+@reg.register_schedule("nn.conv2d_transpose")
+def schedule_conv2d_transpose(attrs, outs, target):
+    """Schedule definition of conv2d_transpose"""
+    with target:
+        return topi.generic.schedule_conv2d_transpose_nchw(outs)
+
+reg.register_pattern("nn.conv2d_transpose", OpPattern.OUT_ELEMWISE_FUSABLE)
+
+# bias_add
+@reg.register_compute("nn.bias_add")
+def compute_bias_add(attrs, inputs, out_dtype, target):
+    """Compute definition of conv2d_transpose"""
+    axis = attrs.axis
+    bias = inputs[1]
+    data_ndim = len(inputs[0].shape)
+    if axis < 0:
+        axis = axis + data_ndim
+    num_newaxis = data_ndim - axis - 1
+
+    if num_newaxis:
+        bias = topi.expand_dims(bias, axis=1, num_newaxis=num_newaxis)
+    return [topi.add(inputs[0], bias)]
+
+reg.register_schedule("nn.bias_add", schedule_injective)
+reg.register_pattern("nn.bias_add", OpPattern.BROADCAST)
+
+
+# max_pool2d
+@reg.register_schedule("nn.max_pool2d")
+def schedule_max_pool2d(attrs, outs, target):
+    """Schedule definition of max_pool2d"""
+    layout = attrs.layout
+    with target:
+        return topi.generic.schedule_pool(outs, layout)
+
+reg.register_pattern("nn.max_pool2d", OpPattern.OUT_ELEMWISE_FUSABLE)
+
+
+# avg_pool2d
+@reg.register_schedule("nn.avg_pool2d")
+def schedule_avg_pool2d(attrs, outs, target):
+    """Schedule definition of avg_pool2d"""
+    layout = attrs.layout
+    with target:
+        return topi.generic.schedule_pool(outs, layout)
+
+reg.register_pattern("nn.avg_pool2d", OpPattern.OUT_ELEMWISE_FUSABLE)
+
+
+# global_max_pool2d
+@reg.register_schedule("nn.global_max_pool2d")
+def schedule_global_max_pool2d(_, outs, target):
+    """Schedule definition of global_max_pool2d"""
+    with target:
+        return topi.generic.schedule_global_pool(outs)
+
+reg.register_pattern("nn.global_max_pool2d", OpPattern.OUT_ELEMWISE_FUSABLE)
+
+
+# global_avg_pool2d
+@reg.register_schedule("nn.global_avg_pool2d")
+def schedule_global_avg_pool2d(_, outs, target):
+    """Schedule definition of global_avg_pool2d"""
+    with target:
+        return topi.generic.schedule_global_pool(outs)
+
+reg.register_pattern("nn.global_avg_pool2d", OpPattern.OUT_ELEMWISE_FUSABLE)

python/tvm/relay/op/op.py

+#pylint: disable=unused-argument
 """The base node types for the Relay language."""
+import topi
+
 from ..._ffi.function import _init_api
 
 from ..base import register_relay_node
@@ -156,3 +159,9 @@ def _lower(name, schedule, inputs, outputs):
 @register_func("relay.op.compiler._build")
 def _build(lowered_funcs):
     return build(lowered_funcs, target="llvm")
+
+
+def schedule_injective(attrs, outputs, target):
+    """Generic schedule for binary broadcast."""
+    with target:
+        return topi.generic.schedule_injective(outputs)

src/relay/backend/compile_engine.cc

@@ -89,7 +89,7 @@ class ScheduleGetter :
     CachedFunc cfunc(cache_node);
     CHECK(master_op_.defined());
     Schedule schedule = fschedule[master_op_](
-        cache_node->outputs, target_);
+        master_attrs_, cache_node->outputs, target_);
     return std::make_pair(schedule, cfunc);
   }
 
@@ -145,6 +145,7 @@ class ScheduleGetter :
     }
     if (op_pattern >= master_op_patetrn_) {
       master_op_ = op;
+      master_attrs_ = call_node->attrs;
       master_op_patetrn_ = op_pattern;
     }
     if (outputs.size() != 1) {
@@ -193,6 +194,7 @@ class ScheduleGetter :
  private:
   tvm::Target target_;
   Op master_op_;
+  Attrs master_attrs_;
   int master_op_patetrn_{0};
   std::ostringstream readable_name_stream_;
   std::unordered_map<Expr, Array<Tensor>, NodeHash, NodeEqual> memo_;
@@ -285,6 +287,9 @@ class CompileEngineImpl : public CompileEngineNode {
    * \return Updated name which is unique.
    */
   std::string GetUniqeName(std::string name) {
+    for (size_t i = 0; i < name.length(); ++i) {
+      if (name[i] == '.') name[i] = '_';
+    }
     while (true) {
       auto it = name_map_.find(name);
       if (it == name_map_.end()) {

src/relay/op/nn/nn.cc

@@ -91,16 +91,15 @@ bool DenseRel(const Array<Type>& types,
   Array<tvm::Expr> oshape = data->shape;
   if (param->units.defined()) {
     Array<tvm::Expr> dshape = data->shape;
-
     // validate the weight shape is proper if defined
     // Assign weight type
-    Array<IndexExpr> wshape({dshape[dshape.size() - 1], param->units});
+    Array<IndexExpr> wshape({param->units, dshape[dshape.size() - 1]});
     reporter->Assign(types[1], TensorTypeNode::make(wshape, data->dtype));
     oshape.Set((oshape.size() - 1), param->units);
   } else {
     if (weight == nullptr) return false;
     Array<tvm::Expr> wshape = weight->shape;
-    oshape.Set((oshape.size() - 1), wshape[wshape.size() - 1]);
+    oshape.Set((oshape.size() - 1), wshape[0]);
   }
 
   // assign output type

src/relay/op/nn/pooling.cc

@@ -4,7 +4,9 @@
  * \brief Pooling operators
  */
 #include <tvm/relay/op.h>
+#include <tvm/relay/op_attr_types.h>
 #include <tvm/relay/attrs/nn.h>
+#include <topi/nn/pooling.h>
 #include <vector>
 #include "layout.h"
 
@@ -14,7 +16,7 @@ namespace relay {
 TVM_REGISTER_NODE_TYPE(MaxPool2DAttrs);
 TVM_REGISTER_NODE_TYPE(AvgPool2DAttrs);
 
-template <typename AttrTtype>
+template <typename AttrType>
 bool Pool2DRel(const Array<Type>& types,
                int num_inputs,
                const Attrs& attrs,
@@ -27,7 +29,7 @@ bool Pool2DRel(const Array<Type>& types,
   CHECK_NE(dshape.size(), 0);
   CHECK_GE(dshape.size(), 2U)
       << "Pool2D only support input >= 2-D: input must have height and width";
-  const auto param = attrs.as<AttrTtype>();
+  const auto param = attrs.as<AttrType>();
   CHECK(param != nullptr);
 
   Layout layout(param->layout);
@@ -88,6 +90,46 @@ Expr MakeMaxPool2D(Expr data,
   return CallNode::make(op, {data}, Attrs(attrs), {});
 }
 
+template<typename AttrType, topi::nn::PoolType mode>
+Array<Tensor> Pool2DCompute(const Attrs& attrs,
+                            const Array<Tensor>& inputs,
+                            const Type& out_type,
+                            const Target& target) {
+  const auto* param = attrs.as<AttrType>();
+  CHECK(param != nullptr);
+  auto pool_size = param->pool_size;
+  auto strides = param->strides;
+  auto padding = param->padding;
+  auto ceil_mode = param->ceil_mode;
+  Layout layout(param->layout);
+  CHECK(layout.convertible(Layout("NCHW")))
+      << "max_pool2d currently only supports layouts that are convertible from NCHW";
+  CHECK_EQ(layout.indexof('h'), -1) << "max_pool2d does not support input split on height";
+  CHECK_EQ(layout.indexof('w'), -1) << "max_pool2d does not support input split on width";
+
+  CHECK(inputs[0].ndim() == 4U || inputs[0].ndim() == 5U)
+      << "Pool2D only support 4-D input (e.g., NCHW)"
+      << " or 5-D input (last dimension is a split of channel)";
+
+  if (param->padding.size() == 1) {
+    padding.push_back(padding[0]);
+    padding.push_back(padding[0]);
+    padding.push_back(padding[0]);
+  } else if (param->padding.size() == 2) {
+    padding.push_back(padding[0]);
+    padding.push_back(padding[1]);
+  }
+  if (mode == topi::nn::kAvgPool) {
+    bool count_include_pad = reinterpret_cast<const AvgPool2DAttrs*>(param)->count_include_pad;
+    return Array<Tensor>{
+      topi::nn::pool(inputs[0], pool_size, strides, padding,
+                     mode, ceil_mode, layout.name(), count_include_pad)};
+  } else {
+    return Array<Tensor>{
+      topi::nn::pool(inputs[0], pool_size, strides, padding,
+                     mode, ceil_mode, layout.name())};
+  }
+}
 
 TVM_REGISTER_API("relay.op.nn._make.max_pool2d")
 .set_body([](const TVMArgs& args, TVMRetValue* rv) {
@@ -120,7 +162,8 @@ RELAY_REGISTER_OP("nn.max_pool2d")
 .set_num_inputs(1)
 .add_argument("data", "Tensor", "The input tensor.")
 .set_support_level(2)
-.add_type_rel("MaxPool2D", Pool2DRel<MaxPool2DAttrs>);
+.add_type_rel("MaxPool2D", Pool2DRel<MaxPool2DAttrs>)
+.set_attr<FTVMCompute>("FTVMCompute", Pool2DCompute<MaxPool2DAttrs, topi::nn::kMaxPool>);
 
 
 // AvgPool2D
@@ -175,7 +218,8 @@ Average pooling operation for one dimensional data.
 .set_num_inputs(1)
 .add_argument("data", "Tensor", "The input tensor.")
 .set_support_level(2)
-.add_type_rel("AvgPool2D", Pool2DRel<AvgPool2DAttrs>);
+.add_type_rel("AvgPool2D", Pool2DRel<AvgPool2DAttrs>)
+.set_attr<FTVMCompute>("FTVMCompute", Pool2DCompute<AvgPool2DAttrs, topi::nn::kAvgPool>);
 
 // Global Pool
 TVM_REGISTER_NODE_TYPE(GlobalPool2DAttrs);
@@ -211,6 +255,29 @@ bool GlobalPool2DRel(const Array<Type>& types,
   return true;
 }
 
+
+template<topi::nn::PoolType mode>
+Array<Tensor> GlobalPool2DCompute(const Attrs& attrs,
+                                  const Array<Tensor>& inputs,
+                                  const Type& out_type,
+                                  const Target& target) {
+  const auto* param = attrs.as<GlobalPool2DAttrs>();
+  CHECK(param != nullptr);
+  Layout layout(param->layout);
+  CHECK(layout.convertible(Layout("NCHW")))
+    << "global_avg_pool2d currently only supports layouts that are convertible from NCHW";
+  CHECK_EQ(layout.indexof('h'), -1)
+    << "global_avg_pool2d does not support input split on height";
+  CHECK_EQ(layout.indexof('w'), -1)
+    << "global_avg_pool2d does not support input split on width";
+
+  CHECK(inputs[0].ndim() == 4U || inputs[0].ndim() == 5U)
+    << "Pool2D only support 4-D input (e.g., NCHW)"
+    << " or 5-D input (last dimension is a split of channel)";
+  return Array<Tensor>{
+    topi::nn::global_pool(inputs[0], mode, layout.name()) };
+}
+
 Expr MakeGlobalAvgPool2D(Expr data,
                          std::string layout) {
   auto attrs = make_node<GlobalPool2DAttrs>();
@@ -239,7 +306,8 @@ RELAY_REGISTER_OP("nn.global_avg_pool2d")
 .set_num_inputs(1)
 .add_argument("data", "Tensor", "The input tensor.")
 .set_support_level(2)
-.add_type_rel("GlobalAvgPool2D", GlobalPool2DRel);
+.add_type_rel("GlobalAvgPool2D", GlobalPool2DRel)
+.set_attr<FTVMCompute>("FTVMCompute", GlobalPool2DCompute<topi::nn::kAvgPool>);
 
 // GlobalMaxPool
 Expr MakeGlobalMaxPool2D(Expr data,
@@ -269,7 +337,8 @@ RELAY_REGISTER_OP("nn.global_max_pool2d")
 .set_num_inputs(1)
 .add_argument("data", "Tensor", "The input tensor.")
 .set_support_level(2)
-.add_type_rel("GlobalMaxPool2D", GlobalPool2DRel);
+.add_type_rel("GlobalMaxPool2D", GlobalPool2DRel)
+.set_attr<FTVMCompute>("FTVMCompute", GlobalPool2DCompute<topi::nn::kMaxPool>);
 
 }  // namespace relay
 }  // namespace tvm

tests/python/relay/test_backend_interpreter.py

@@ -55,28 +55,6 @@ def test_mul_param():
     check_eval(func, [x_data, y_data], x_data * y_data)
 
 
-# failing due to numeric issues
-
-# def test_dense():
-#     x = relay.var('x', shape=(10, 10))
-#     w = relay.var('w', shape=(10, 10))
-#     y = relay.nn.dense(x, w)
-#     func = relay.Function([x, w], y)
-#     x_data = np.random.rand(10, 10).astype('float32')
-#     w_data = np.random.rand(10, 10).astype('float32')
-#     check_eval(func, [x_data, w_data], x_data @ w_data, rtol=0.1)
-
-# def test_linear():
-#     x = relay.var('x', shape=(10, 10))
-#     w = relay.var('w', shape=(10, 10))
-#     b = relay.var('b', shape=(10,))
-#     y = relay.add(relay.nn.dense(x, w), b)
-#     func = relay.Function([x, w, b], y)
-#     x_data = np.random.rand(10, 10).astype('float32')
-#     w_data = np.random.rand(10, 10).astype('float32')
-#     b_data = np.random.rand(10).astype('float32')
-#     check_eval(func, [x_data, w_data, b_data], x_data @ w_data + b_data)
-
 def test_equal():
     i = relay.var('i', shape=[], dtype='int32')
     j = relay.var('i', shape=[], dtype='int32')

tests/python/relay/test_op_level1.py

@@ -74,6 +74,7 @@ def test_binary_op():
             y_data = np.random.rand(5, 10, 5).astype(t2.dtype)
             ref_res = ref(x_data, y_data)
             func = relay.Function([x, y], z)
+
             for target, ctx in ctx_list():
                 # use graph by execuor default for testing, as we need
                 # create function explicitly to avoid constant-folding.
@@ -89,12 +90,24 @@ def test_binary_op():
 
 
 def test_bias_add():
-    x = relay.var("x", shape=(10, 2, 3, 4))
+    xshape=(10, 2, 3, 4)
+    bshape=(2,)
+    dtype="float32"
+    x = relay.var("x", shape=xshape)
     bias = relay.var("bias")
     z = relay.nn.bias_add(x, bias)
     zz = relay.ir_pass.infer_type(z)
     assert "axis=" not in zz.astext()
-    assert zz.args[1].checked_type == relay.TensorType((2,))
+    assert zz.args[1].checked_type == relay.TensorType(bshape)
+
+    func = relay.Function([x, bias], z)
+    x_data = np.random.uniform(size=xshape).astype(dtype)
+    y_data = np.random.uniform(size=bshape).astype(dtype)
+    ref_res = x_data + y_data.reshape((2, 1, 1))
+    for target, ctx in ctx_list():
+        intrp = relay.create_executor("graph", ctx=ctx, target=target)
+        op_res = intrp.evaluate(func)(x_data, y_data)
+        np.testing.assert_allclose(op_res.asnumpy(), ref_res, rtol=1e-5)
 
 
 def test_expand_dims_infer_type():
@@ -217,6 +230,50 @@ def test_batch_norm():
     ]))
 
 
+def test_dense():
+    n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), tvm.var("w")
+    x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
+    w = relay.var("w", relay.TensorType((2, w), "float32"))
+    y = relay.nn.dense(x, w, units=2)
+    "units=2" in y.astext()
+    yy = relay.ir_pass.infer_type(y)
+    assert yy.checked_type == relay.TensorType((n, c, h, 2), "float32")
+
+    n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), 2
+    x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
+    wh, ww = tvm.var("wh"), tvm.var("ww")
+    w = relay.var("w", relay.TensorType((ww, wh), "float32"))
+    y = relay.nn.dense(x, w)
+    yy = relay.ir_pass.infer_type(y)
+    assert yy.checked_type == relay.TensorType((n, c, h, ww), "float32")
+
+    n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), 2
+    x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
+    w = relay.var("w", relay.IncompleteType())
+    y = relay.nn.dense(x, w, units=2)
+    yy = relay.ir_pass.infer_type(y)
+    assert yy.checked_type == relay.TensorType((n, c, h, 2), "float32")
+
+    x = relay.var("x", shape=(10, 5))
+    w = relay.var("w", shape=(2, 5))
+    z = relay.nn.dense(x, w)
+
+    # Check result.
+    func = relay.Function([x, w], z)
+    x_data = np.random.rand(10, 5).astype('float32')
+    w_data = np.random.rand(2, 5).astype('float32')
+    ref_res = np.dot(x_data, w_data.T)
+
+    for target, ctx in ctx_list():
+        intrp1 = relay.create_executor("graph", ctx=ctx, target=target)
+        intrp2 = relay.create_executor("debug", ctx=ctx, target=target)
+        op_res1 = intrp1.evaluate(func)(x_data, w_data)
+        tvm.testing.assert_allclose(op_res1.asnumpy(), ref_res, rtol=1e-5)
+        op_res2 = intrp2.evaluate(func)(x_data, w_data)
+        tvm.testing.assert_allclose(op_res2.asnumpy(), ref_res, rtol=1e-5)
+
+
+
 if __name__ == "__main__":
     test_bias_add()
     test_unary_op()
@@ -227,3 +284,4 @@ if __name__ == "__main__":
     test_log_softmax()
     test_dropout()
     test_batch_norm()
+    test_dense()

tests/python/relay/test_op_level2.py

@@ -2,7 +2,9 @@
 """
 import tvm
 from tvm import relay
-
+from tvm.relay.testing import ctx_list
+import numpy as np
+import topi.testing
 
 def test_conv2d_infer_type():
     # symbolic in batch dimension
@@ -62,6 +64,62 @@ def test_conv2d_infer_type():
         (n, h, w, 16), "int32")
 
 
+def test_conv2d_run():
+    def run_test_conv2d(dtype, out_dtype, scale, dshape, kshape,
+                        padding=(1, 1),
+                        fref=None,
+                        groups=1,
+                        dilation=(1, 1),
+                        **attrs):
+        x = relay.var("x", shape=dshape)
+        w = relay.var("w")
+        y = relay.nn.conv2d(x, w,
+                            padding=padding,
+                            dilation=dilation,
+                            groups=groups,
+                            **attrs)
+        func = relay.Function([x, w], y)
+        data = np.random.uniform(-scale, scale, size=dshape).astype(dtype)
+        kernel = np.random.uniform(-scale, scale, size=kshape).astype(dtype)
+        dkernel = topi.testing.dilate_python(kernel, (1, 1) + dilation)
+        if fref is None:
+            ref_res = topi.testing.conv2d_nchw_python(
+                data.astype(out_dtype), dkernel.astype(out_dtype), 1, padding)
+        else:
+            ref_res = fref(data.astype(out_dtype), dkernel.astype(out_dtype))
+
+        for target, ctx in ctx_list():
+            intrp1 = relay.create_executor("graph", ctx=ctx, target=target)
+            op_res1 = intrp1.evaluate(func)(data, kernel)
+            tvm.testing.assert_allclose(op_res1.asnumpy(), ref_res, rtol=1e-5, atol=1e-5)
+
+    # depthwise conv2d
+    dshape = (1, 32, 18, 18)
+    kshape = (32, 1, 3, 3)
+    run_test_conv2d("float32", "float32", 1, dshape, kshape,
+                    padding=(1, 1), channels=32, groups=32, kernel_size=(3 ,3),
+                    fref=lambda x, w: topi.testing.depthwise_conv2d_python_nchw(
+                        x, w, (1, 1), "SAME"))
+
+    # normal conv2d
+    dshape = (1, 3, 224, 224)
+    kshape = (10, 3, 3, 3)
+    run_test_conv2d("float32", "float32", 1, dshape, kshape,
+                    padding=(1, 1), channels=10, kernel_size=(3 ,3))
+    # mixed precision
+    run_test_conv2d("int8", "int32", 1, dshape, kshape,
+                    padding=(1, 1), channels=10, kernel_size=(3 ,3))
+    kshape = (10, 3, 1, 3)
+    # mixed precision.
+    run_test_conv2d("int8", "int32", 1, dshape, kshape,
+                    padding=(0, 1), channels=10, kernel_size=(1 ,3))
+    # dilated conv2d
+    dshape = (1, 3, 18, 18)
+    kshape = (10, 3, 3, 3)
+    run_test_conv2d("float32", "float32", 1, dshape, kshape,
+                    padding=(1, 1), channels=10, kernel_size=(3 ,3), dilation=(3, 3))
+
+
 def test_conv2d_transpose_infer_type():
     # symbolic in batch dimension
     n, c, h, w = tvm.var("n"), 10, 10, 12
@@ -90,6 +148,33 @@ def test_conv2d_transpose_infer_type():
     assert yy.checked_type == relay.TensorType(
         (n, 15, 15, 11), "float32")
 
+
+def test_conv2d_transpose_run():
+    dshape = (1, 3, 18, 18)
+    kshape = (3, 10, 3, 3)
+    oshape = (1, 10, 37, 37)
+    x = relay.var("x", shape=dshape)
+    w = relay.var("w")
+    y = relay.nn.conv2d_transpose(x, w,
+                                  channels=10, kernel_size=(3,3), strides=(2,2),
+                                  padding=(1,1), output_padding=(2, 2))
+    func = relay.Function([x, w], y)
+    dtype = "float32"
+    data = np.random.uniform(size=dshape).astype(dtype)
+    kernel = np.random.uniform(size=kshape).astype(dtype)
+    c_np = topi.testing.conv2d_transpose_nchw_python(
+        data, kernel, 2, 1)
+    d_np = np.zeros(shape=oshape)
+    d_np[:,:,0:c_np.shape[2],0:c_np.shape[3]] = c_np
+    ref_res = d_np
+
+    for target, ctx in ctx_list():
+        intrp1 = relay.create_executor("graph", ctx=ctx, target=target)
+        op_res1 = intrp1.evaluate(func)(data, kernel)
+        tvm.testing.assert_allclose(op_res1.asnumpy(), ref_res, rtol=1e-5, atol=1e-5)
+
+
+
 def test_upsampling_infer_type():
     n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), tvm.var("w")
     x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
@@ -103,15 +188,29 @@ def test_upsampling_infer_type():
     yy = relay.ir_pass.infer_type(y)
     assert yy.checked_type == relay.TensorType((n, c, 200, 400), "float32")
 
-def _test_pool2d_infer_type(opfunc):
+
+def _test_pool2d(opfunc, reffunc):
     n, c, h, w = tvm.var("n"), 10, 224, 224
     x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
     y = opfunc(x, pool_size=(1, 1))
     assert "pool_size=" in y.astext()
     yy = relay.ir_pass.infer_type(y)
     assert yy.checked_type == relay.TensorType((n, 10, 224, 224), "float32")
+    # test execution
+    dtype = "float32"
+    dshape = (1, 3, 28, 28)
+    x = relay.var("x", shape=dshape)
+    y = opfunc(x, pool_size=(2, 2), strides=(2, 2), padding=(0, 0))
+    func = relay.Function([x], y)
+    data = np.random.uniform(size=dshape).astype(dtype)
+    ref_res = reffunc(data.reshape(1,3,14,2,14,2), axis=(3,5))
+    for target, ctx in ctx_list():
+        intrp1 = relay.create_executor("graph", ctx=ctx, target=target)
+        op_res1 = intrp1.evaluate(func)(data)
+        tvm.testing.assert_allclose(op_res1.asnumpy(), ref_res, rtol=1e-5, atol=1e-5)
 
-def _test_global_pool2d_infer_type(opfunc):
+
+def _test_global_pool2d(opfunc, reffunc):
     n, c, h, w = tvm.var("n"), tvm.var("c"), 224, 224
     x = relay.var("x", relay.TensorType((n, h, w, c), "float32"))
     y = opfunc(x, layout="NHWC")
@@ -123,12 +222,61 @@ def _test_global_pool2d_infer_type(opfunc):
     y = opfunc(x)
     yy = relay.ir_pass.infer_type(y)
     assert yy.checked_type == relay.TensorType((n, c, 1, 1), "float32")
+    # test execution
+    dtype = "float32"
+    dshape = (1, 1024, 7, 7)
+    x = relay.var("x", shape=dshape)
+    y = opfunc(x)
+    func = relay.Function([x], y)
+    data = np.random.uniform(size=dshape).astype(dtype)
+    ref_res = reffunc(data, axis=(2,3), keepdims=True)
+    for target, ctx in ctx_list():
+        intrp1 = relay.create_executor("graph", ctx=ctx, target=target)
+        op_res1 = intrp1.evaluate(func)(data)
+        tvm.testing.assert_allclose(op_res1.asnumpy(), ref_res, rtol=1e-5, atol=1e-5)
+
+
+def test_pool2d():
+    _test_pool2d(relay.nn.max_pool2d, np.max)
+    _test_pool2d(relay.nn.avg_pool2d, np.mean)
+    _test_global_pool2d(relay.nn.global_max_pool2d, np.max)
+    _test_global_pool2d(relay.nn.global_avg_pool2d, np.mean)
+
+
+def test_avg_pool2d_no_count_pad():
+    kh, kw = (4, 4)
+    sh, sw = (2, 2)
+    ph, pw = (2, 2)
+    n = 1
+    (ic, ih, iw) = (3, 28, 28)
+    (oc, oh, ow) = (3, 15, 15)
+    dshape = (n, ic, ih, iw)
+    x = relay.var("x", shape=dshape)
+    y = relay.nn.avg_pool2d(x,
+                            pool_size=(kh, kw),
+                            strides=(sw, sw),
+                            padding=(ph, pw),
+                            count_include_pad=False)
+    func = relay.Function([x], y)
+    dtype = "float32"
+    a_np = np.random.uniform(low=0.001, size=(n, ic, ih, iw)).astype(dtype)
+    pad_np = np.zeros(shape=(n, ic, ih+2*ph, iw+2*pw)).astype(dtype)
+    no_zero = (range(n), range(ic), (range(ph, ih+ph)), (range(pw, iw+pw)))
+    pad_np[np.ix_(*no_zero)] = a_np
+    b_np = np.zeros(shape=(n, oc, oh, ow)).astype(dtype)
+    for i in range(oh):
+        for j in range(ow):
+            pad_count = np.sum(pad_np[:, :, i*sh:i*sh+kh, j*sw:j*sw+kw] > 0, axis=(2,3))
+            b_np[:,:,i,j] = np.sum(pad_np[:, :, i*sh:i*sh+kh, j*sw:j*sw+kw],
+                                   axis=(2,3)) / np.maximum(pad_count, 1)
+    ref_res = np.maximum(b_np, 0.0)
+    data = a_np
+
+    for target, ctx in ctx_list():
+        intrp1 = relay.create_executor("graph", ctx=ctx, target=target)
+        op_res1 = intrp1.evaluate(func)(data)
+        tvm.testing.assert_allclose(op_res1.asnumpy(), ref_res, rtol=1e-5, atol=1e-5)
 
-def test_pool2d_infer_type():
-    _test_pool2d_infer_type(relay.nn.max_pool2d)
-    _test_pool2d_infer_type(relay.nn.avg_pool2d)
-    _test_global_pool2d_infer_type(relay.nn.global_avg_pool2d)
-    _test_global_pool2d_infer_type(relay.nn.global_avg_pool2d)
 
 def test_flatten_infer_type():
     d1, d2, d3, d4 = tvm.var("d1"), tvm.var("d2"), tvm.var("d3"), tvm.var("d4")
@@ -163,30 +311,6 @@ def test_pad_infer_type():
     yy = relay.ir_pass.infer_type(y)
     assert yy.checked_type == relay.TensorType((n + 2, 6, 9, w + 8), "float32")
 
-def test_dense_infer_type():
-    n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), tvm.var("w")
-    x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
-    w = relay.var("w", relay.TensorType((w, 2), "float32"))
-    y = relay.nn.dense(x, w, units=2)
-    "units=2" in y.astext()
-    yy = relay.ir_pass.infer_type(y)
-    assert yy.checked_type == relay.TensorType((n, c, h, 2), "float32")
-
-    n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), 2
-    x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
-    wh, ww = tvm.var("wh"), tvm.var("ww")
-    w = relay.var("w", relay.TensorType((wh, ww), "float32"))
-    y = relay.nn.dense(x, w)
-    yy = relay.ir_pass.infer_type(y)
-    assert yy.checked_type == relay.TensorType((n, c, h, ww), "float32")
-
-    n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), 2
-    x = relay.var("x", relay.TensorType((n, c, h, w), "float32"))
-    w = relay.var("w", relay.IncompleteType())
-    y = relay.nn.dense(x, w, units=2)
-    yy = relay.ir_pass.infer_type(y)
-    assert yy.checked_type == relay.TensorType((n, c, h, 2), "float32")
-
 
 def test_lrn():
     n, c , h, w = tvm.var("n"), tvm.var("c"), tvm.var("h"), tvm.var("w")
@@ -206,12 +330,14 @@ def test_l2_normalize():
 
 
 if __name__ == "__main__":
+    test_pool2d()
+    test_avg_pool2d_no_count_pad()
     test_lrn()
     test_l2_normalize()
     test_conv2d_infer_type()
-    test_pool2d_infer_type()
     test_upsampling_infer_type()
     test_flatten_infer_type()
     test_pad_infer_type()
     test_conv2d_transpose_infer_type()
-    test_dense_infer_type()
+    test_conv2d_transpose_run()
+    test_conv2d_run()

topi/python/topi/util.py

 # pylint: disable=invalid-name
 """Common topi utilities"""
 from __future__ import absolute_import as _abs
-import tvm
+from numbers import Integral
 
+import tvm
 from . import tag
 
 def traverse_inline(s, final_op, callback):
@@ -68,13 +69,13 @@ def get_const_int(expr):
     out_value : int
         The output.
     """
-    if isinstance(expr, int):
+    if isinstance(expr, Integral):
         return expr
     if not isinstance(expr, (tvm.expr.IntImm, tvm.expr.UIntImm)):
         expr = tvm.ir_pass.Simplify(expr)
     if not isinstance(expr, (tvm.expr.IntImm, tvm.expr.UIntImm)):
         raise ValueError("Expect value to be constant int")
-    return expr.value
+    return int(expr.value)
 
 
 def equal_const_int(expr, value):
@@ -90,7 +91,7 @@ def equal_const_int(expr, value):
     equal : bool
         Whether they equals.
     """
-    if isinstance(expr, int):
+    if isinstance(expr, Integral):
         return expr == value
     if not isinstance(expr, (tvm.expr.IntImm, tvm.expr.UIntImm)):
         expr = tvm.ir_pass.Simplify(expr)