[bug fix] update quantizable add ops detection on IPEX backend (#1456)

Signed-off-by: Xin He <[email protected]>

Author: xin3he

neural_compressor/adaptor/torch_utils/util.py

@@ -465,15 +465,13 @@ def get_quantizable_ops_from_cfgs(ops_name, op_infos_from_cfgs, input_tensor_ids
             op_info = op_infos_from_cfgs[name]
             output_tensors = op_info["output_tensor_infos"]
             input_tensors = op_info["input_tensor_infos"]
-            for input_tensor in input_tensors:
-                if "inf_dtype" not in input_tensor.keys():
-                    continue
-                if input_tensor["inf_dtype"] == torch.float32:
-                    pre_op_name = input_tensor_ids_op_name[input_tensor["id"]]
-                    if pre_op_name[1] in ["q_op_infos"]:
-                        print(pre_op_name, "is not the fuse ops first op.")
-                        start = False
-                        continue
+            start = any(
+                [
+                    input_tensor["inf_dtype"] != "torch.float32"
+                    for input_tensor in input_tensors
+                    if "inf_dtype" in input_tensor.keys()
+                ]
+            )
             if not start:
                 continue
             # add quantizable ops, include op and fuse ops.

test/ipex/test_adaptor_ipex.py

@@ -316,62 +316,6 @@ def forward(self, a):
         )
         self.assertTrue(isinstance(q_model._model, torch.jit.ScriptModule))
 
-    def test_tune_add(self):
-        class M(torch.nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.conv = torch.nn.Conv2d(3, 1, 1)
-                self.linear = torch.nn.Linear(224 * 224, 5)
-
-            def forward(self, a):
-                x = self.conv(a)
-                x = x.view(1, -1)
-                x += x
-                x = self.linear(x)
-                return x
-
-        model = M()
-        from neural_compressor import PostTrainingQuantConfig, quantization
-
-        acc_lst = [1, 0.8, 1.1, 1.2]
-
-        def fake_eval(model):
-            res = acc_lst.pop(0)
-            return res
-
-        conf = PostTrainingQuantConfig(backend="ipex", quant_level=0)
-        calib_dataloader = Dataloader()
-        q_model = quantization.fit(model, conf, calib_dataloader=calib_dataloader, eval_func=fake_eval)
-        self.assertTrue(isinstance(q_model._model, torch.jit.ScriptModule))
-
-    def test_tune_add_with_recipe(self):
-        class M(torch.nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.linear = torch.nn.Linear(224 * 224 * 3, 5)
-
-            def forward(self, x):
-                x += x
-                x = x.view(1, -1)
-                x = self.linear(x)
-                return x
-
-        model = M()
-        from neural_compressor import PostTrainingQuantConfig, quantization
-
-        acc_lst = [1, 0.8, 1.1, 1.2]
-
-        def fake_eval(model):
-            res = acc_lst.pop(0)
-            return res
-
-        conf = PostTrainingQuantConfig(
-            backend="ipex", quant_level=0, recipes={"smooth_quant": True, "smooth_quant_args": {"alpha": 0.5}}
-        )
-        calib_dataloader = Dataloader()
-        q_model = quantization.fit(model, conf, calib_dataloader=calib_dataloader, eval_func=fake_eval)
-        self.assertTrue(isinstance(q_model._model, torch.jit.ScriptModule))
-
     def test_tune_minmax_obs(self):
         class M(torch.nn.Module):
             def __init__(self):
@@ -527,67 +471,6 @@ def forward(self, a):
         )
         self.assertTrue(isinstance(q_model._model, torch.jit.ScriptModule))
 
-    def test_tune_add(self):
-        class M(torch.nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.conv = torch.nn.Conv2d(3, 1, 1)
-                self.linear = torch.nn.Linear(224 * 224, 5)
-
-            def forward(self, a):
-                x = self.conv(a)
-                x = x.view(1, -1)
-                x += x
-                x = self.linear(x)
-                return x
-
-        model = M().to("xpu")
-        from neural_compressor import PostTrainingQuantConfig, quantization
-
-        acc_lst = [1, 0.8, 1.1, 1.2]
-
-        def fake_eval(model):
-            res = acc_lst.pop(0)
-            return res
-
-        conf = PostTrainingQuantConfig(backend="ipex", device="xpu", quant_level=0)
-        calib_dataloader = Dataloader(device="xpu")
-        q_model = quantization.fit(model, conf, calib_dataloader=calib_dataloader, eval_func=fake_eval)
-        self.assertTrue(isinstance(q_model._model, torch.jit.ScriptModule))
-
-    def test_tune_add_with_recipe(self):
-        class M(torch.nn.Module):
-            def __init__(self):
-                super().__init__()
-                self.conv = torch.nn.Conv2d(3, 1, 1)
-                self.linear = torch.nn.Linear(224 * 224, 5)
-
-            def forward(self, a):
-                x = self.conv(a)
-                x = x.view(1, -1)
-                x += x
-                x = self.linear(x)
-                return x
-
-        model = M().to("xpu")
-        from neural_compressor import PostTrainingQuantConfig, quantization
-
-        acc_lst = [1, 0.8, 1.1, 1.2]
-
-        def fake_eval(model):
-            res = acc_lst.pop(0)
-            return res
-
-        conf = PostTrainingQuantConfig(
-            backend="ipex",
-            device="xpu",
-            quant_level=0,
-            recipes={"smooth_quant": True, "smooth_quant_args": {"alpha": 0.5}},
-        )
-        calib_dataloader = Dataloader(device="xpu")
-        q_model = quantization.fit(model, conf, calib_dataloader=calib_dataloader, eval_func=fake_eval)
-        self.assertTrue(isinstance(q_model._model, torch.jit.ScriptModule))
-
 
 class TestMixedPrecision(unittest.TestCase):
     @classmethod