feat(mge): add imperative pad

GitOrigin-RevId: de79de536f

imperative/python/megengine/functional/nn.py

@@ -1592,6 +1592,57 @@ def sliding_window_transpose(
     return output
 
 
+def pad(
+    src: Tensor,
+    pad_witdth: Tuple[Tuple[int, int], ...],
+    mode: str = "CONSTANT",
+    constant_value: float = 0.0,
+) -> Tensor:
+    """
+    pad
+    """
+    p_offsets = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
+
+    assert mode in [
+        "constant",
+        "CONSTANT",
+        "edge",
+        "EDGE",
+        "replicate",
+        "REPLICATE",
+        "reflect",
+        "REFLECT",
+    ]
+
+    if mode.lower() == "edge":
+        mode = "replicate"
+
+    for i in range(0, len(pad_witdth)):
+        p_offsets[i * 2] = pad_witdth[i][0]
+        p_offsets[i * 2 + 1] = pad_witdth[i][1]
+
+    op = builtin.Padding(
+        front_offset_dim0=p_offsets[0],
+        front_offset_dim1=p_offsets[2],
+        front_offset_dim2=p_offsets[4],
+        front_offset_dim3=p_offsets[6],
+        front_offset_dim4=p_offsets[8],
+        front_offset_dim5=p_offsets[10],
+        front_offset_dim6=p_offsets[12],
+        back_offset_dim0=p_offsets[1],
+        back_offset_dim1=p_offsets[3],
+        back_offset_dim2=p_offsets[5],
+        back_offset_dim3=p_offsets[7],
+        back_offset_dim4=p_offsets[9],
+        back_offset_dim5=p_offsets[11],
+        back_offset_dim6=p_offsets[13],
+        padding_val=constant_value,
+        padding_mode=mode.upper(),
+    )
+    (output,) = apply(op, src)
+    return output
+
+
 interpolate = deprecated_func("1.3", "megengine.functional.vision", "interpolate", True)
 roi_pooling = deprecated_func("1.3", "megengine.functional.vision", "roi_pooling", True)
 roi_align = deprecated_func("1.3", "megengine.functional.vision", "roi_align", True)

imperative/python/megengine/module/__init__.py

@@ -31,6 +31,7 @@ from .identity import Identity
 from .linear import Linear
 from .module import Module
 from .normalization import GroupNorm, InstanceNorm, LayerNorm
+from .padding import Pad
 from .pooling import AvgPool2d, MaxPool2d
 from .quant_dequant import DequantStub, QuantStub
 from .sequential import Sequential

imperative/python/megengine/module/padding.py

@@ -0,0 +1,22 @@
+from typing import Tuple
+
+from ..functional import nn
+from .module import Module
+
+
+class Pad(Module):
+    def __init__(
+        self,
+        pad_witdth: Tuple[Tuple[int, int], ...],
+        mode: str = "CONSTANT",
+        constant_val: float = 0.0,
+    ):
+        super().__init__()
+        self.pad_width = pad_witdth
+        self.mode = mode
+        self.pad_val = constant_val
+
+    def forward(self, src):
+        return nn.pad(
+            src, pad_witdth=self.pad_width, mode=self.mode, constant_value=self.pad_val
+        )

imperative/python/test/unit/functional/test_functional.py

@@ -1062,3 +1062,22 @@ def test_sliding_window_transpose():
         dilation=(dh, dw),
     )
     np.testing.assert_equal(gt_out, out.numpy())
+
+
+def test_pad():
+    src = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]], dtype=np.float32)
+    dst = np.pad(src, ((2, 2), (2, 2)), "constant")
+    res = F.nn.pad(tensor(src), ((2, 2), (2, 2)), "CONSTANT")
+    np.testing.assert_allclose(res, dst, atol=1e-5)
+
+    dst = np.pad(src, ((2, 2), (2, 2)), "constant", constant_values=3)
+    res = F.nn.pad(tensor(src), ((2, 2), (2, 2)), "CONSTANT", constant_value=3)
+    np.testing.assert_allclose(res, dst, atol=1e-5)
+
+    dst = np.pad(src, ((2, 2), (2, 2)), "edge")
+    res = F.nn.pad(tensor(src), ((2, 2), (2, 2)), "EDGE")
+    np.testing.assert_allclose(res, dst, atol=1e-5)
+
+    dst = np.pad(src, ((2, 2), (2, 2)), "reflect")
+    res = F.nn.pad(tensor(src), ((2, 2), (2, 2)), "REFLECT")
+    np.testing.assert_allclose(res, dst, atol=1e-5)

imperative/src/impl/ops/specializations.cpp

@@ -660,4 +660,12 @@ OP_TRAIT_REG(Cumsum, Cumsum).apply_on_var_node(apply_on_var_node).fallback();
 }  // namespace cumsum
 }  // namespace
 
+namespace padding {
+auto apply_on_var_node(const OpDef& def, const VarNodeArray& inputs) {
+    auto&& op = static_cast<const Padding&>(def);
+    mgb_assert(inputs.size() == 1);
+    return opr::Padding::make(inputs[0], op.param());
+}
+OP_TRAIT_REG(Padding, Padding).apply_on_var_node(apply_on_var_node).fallback();
+} // namespace padding
 } // namespace mgb::imperative

src/core/include/megbrain/ir/ops.td

@@ -389,4 +389,6 @@ def Split: MgbHashableOp<"Split", [EmptyParam]> {
   );
 }
 
+def Padding: MgbHashableOp<"Padding", [PaddingParam]>;
+
 #endif // MGB_OPS