graph mode: add hardswish inplace handling (#40284)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40284

Adds graph mode handling for inplace hardswish, and test coverage for functional hardswish.

Test Plan:
```
python test/test_quantization.py TestQuantizeScriptPTSQOps.test_hardswish
```

Imported from OSS

Differential Revision: D22140628

fbshipit-source-id: 55a514f7dc1130d510f69ee4e611d7cb5e08d02e

test/quantization/test_quantize_jit.py

@@ -2177,11 +2177,23 @@ class TestQuantizeJitOps(QuantizationTestCase):
                            .run(m.graph)
 
     def test_hardswish(self):
-        data = [(torch.rand((1, 2, 5, 5), dtype=torch.float), torch.randint(0, 1, (1,), dtype=torch.long)) for _ in range(2)]
-        hardswish = torch.nn.Hardswish()
-        for tracing in [True, False]:
-            m = self.checkGraphModeOp(hardswish, data, "quantized::hardswish", tracing)
+        class FunctionalHardswish(torch.nn.Module):
+            def __init__(self, inplace):
+                super(FunctionalHardswish, self).__init__()
+                self.inplace = inplace
+
+            def forward(self, input):
+                return torch.nn.functional.hardswish(input, inplace=self.inplace)
+
+        modules = [torch.nn.Hardswish(), FunctionalHardswish(True),
+                   FunctionalHardswish(False)]
+
+        for test_case in itertools.product([True, False], modules):
+            tracing, m = test_case
+            m = self.checkGraphModeOp(
+                m, self.img_data, "quantized::hardswish", tracing)
             FileCheck().check_not("aten::hardswish") \
+                       .check_not("aten::hardswish_") \
                        .run(m.graph)
 
     def test_elu(self):

torch/csrc/jit/passes/quantization/helper.cpp

@@ -34,6 +34,7 @@ std::vector<std::string> _static_quantizable_aten_funcs = {
     "addmm",
     "matmul",
     "hardswish",
+    "hardswish_",
     "elu",
     "elu_",
     "batch_norm",

torch/csrc/jit/passes/quantization/quantization_patterns.h

@@ -848,6 +848,9 @@ graph(%a_quant, %alpha, %scale, %input_scale, %r_scale, %r_zero_point, %r_dtype)
   auto hardswish = getObservedQParamOpFusionInfo(
       "aten::hardswish", "quantized::hardswish", {}, {});
 
+  auto hardswish_ = getObservedQParamOpFusionInfo(
+      "aten::hardswish_", "quantized::hardswish", {}, {});
+
   auto layer_norm = getObservedQParamOpFusionInfo(
       "aten::layer_norm",
       "quantized::layer_norm",
@@ -968,6 +971,7 @@ graph(%a_quant, %alpha, %scale, %input_scale, %r_scale, %r_zero_point, %r_dtype)
       {"quantized::mul", mul, quantized_mul},
       {"quantized::mul", inplace_mul, quantized_mul},
       hardswish,
+      hardswish_,
       layer_norm,
       group_norm,
       instance_norm,

torch/testing/_internal/common_quantization.py

@@ -7,6 +7,7 @@ r"""Importing this file includes common utility methods and base clases for
 checking quantization api and properties of resulting modules.
 """
 
+import copy
 import io
 import functools
 import torch
@@ -282,8 +283,12 @@ class QuantizationTestCase(TestCase):
                 # make sure it runs
                 outputs[d] = models[d](inputs)
             else:
+                # module under test can contain in-place ops, and we depend on
+                # input data staying constant for comparisons
+                data_copy = copy.deepcopy(data)
                 models[d] = quantize_jit(
-                    model, qconfig_dict, test_only_eval_fn, [data], inplace=False, debug=d)
+                    model, qconfig_dict, test_only_eval_fn, [data_copy], inplace=False,
+                    debug=d)
                 # make sure it runs
                 outputs[d] = models[d](*inputs)