support export hardsigmoid in torch<=1.8 (open-mmlab#169)

* support export hardsigmoid in torch<=1.8

* fix lint

mmdeploy/pytorch/ops/__init__.py

@@ -3,12 +3,14 @@
  adaptive_avg_pool2d__default,
  adaptive_avg_pool3d__default)
 from .grid_sampler import grid_sampler__default
+from .hardsigmoid import hardsigmoid__default
 from .instance_norm import instance_norm__tensorrt
 from .lstm import generic_rnn__ncnn
 from .squeeze import squeeze__default
 
 __all__ = [
  'adaptive_avg_pool1d__default', 'adaptive_avg_pool2d__default',
  'adaptive_avg_pool3d__default', 'grid_sampler__default',
- 'instance_norm__tensorrt', 'generic_rnn__ncnn', 'squeeze__default'
+ 'hardsigmoid__default', 'instance_norm__tensorrt', 'generic_rnn__ncnn',
+ 'squeeze__default'
 ]

mmdeploy/pytorch/ops/hardsigmoid.py

@@ -0,0 +1,12 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved
+# Modified from:
+# https://github.com/pytorch/pytorch/blob/9ade03959392e5a90b74261012de1d806cab2253/torch/onnx/symbolic_opset9.py
+from mmdeploy.core import SYMBOLIC_REWRITER
+
+
+@SYMBOLIC_REWRITER.register_symbolic(
+ 'hardsigmoid', is_pytorch=True, arg_descriptors=['v'])
+def hardsigmoid__default(ctx, g, self):
+ """Support export hardsigmoid This rewrite enable export hardsigmoid in
+ torch<=1.8.2."""
+ return g.op('HardSigmoid', self, alpha_f=1 / 6)

tests/test_pytorch/test_pytorch_ops.py

@@ -116,3 +116,10 @@ def test_squeeze(self):
  nodes = get_model_onnx_nodes(model, x)
  assert nodes[0].attribute[0].ints == [0]
  assert nodes[0].op_type == 'Squeeze'
+
+
+def test_hardsigmoid():
+ x = torch.rand(1, 2, 3, 4)
+ model = torch.nn.Hardsigmoid().eval()
+ nodes = get_model_onnx_nodes(model, x)
+ assert nodes[0].op_type == 'HardSigmoid'