rm _checkparam usage

Author: HaoyangLee

mindocr/optim/adamw.py

@@ -4,8 +4,6 @@
 
 import mindspore as ms
 from mindspore import ops
-from mindspore._checkparam import Rel
-from mindspore._checkparam import Validator as validator
 from mindspore.common.initializer import initializer
 from mindspore.common.parameter import Parameter
 from mindspore.common.tensor import Tensor
@@ -15,12 +13,12 @@
 
 def _check_param_value(beta1, beta2, eps, prim_name):
     """Check the type of inputs."""
-    validator.check_value_type("beta1", beta1, [float], prim_name)
-    validator.check_value_type("beta2", beta2, [float], prim_name)
-    validator.check_value_type("eps", eps, [float], prim_name)
-    validator.check_float_range(beta1, 0.0, 1.0, Rel.INC_NEITHER, "beta1", prim_name)
-    validator.check_float_range(beta2, 0.0, 1.0, Rel.INC_NEITHER, "beta2", prim_name)
-    validator.check_positive_float(eps, "eps", prim_name)
+    assert isinstance(beta1, float), f"For '{prim_name}', the type of 'beta1' must be 'float', but got type '{type(beta1).__name__}'."
+    assert isinstance(beta2, float), f"For '{prim_name}', the type of 'beta2' must be 'float', but got type '{type(beta2).__name__}'."
+    assert isinstance(eps, float), f"For '{prim_name}', the type of 'eps' must be 'float', but got type '{type(eps).__name__}'."
+    assert 0.0 < beta1 < 1.0, f"For '{prim_name}', the range of 'beta1' must be (0.0, 1.0), but got {beta1}."
+    assert 0.0 < beta2 < 1.0, f"For '{prim_name}', the range of 'beta2' must be (0.0, 1.0), but got {beta2}."
+    assert eps > 0, f"For '{prim_name}', the 'eps' must be positive, but got {eps}."
 
 
 _grad_scale = ops.MultitypeFuncGraph("grad_scale")

mindocr/optim/adan.py

@@ -15,8 +15,6 @@
 """adan"""
 from __future__ import absolute_import
 
-from mindspore._checkparam import Rel
-from mindspore._checkparam import Validator as validator
 from mindspore.common import dtype as mstype
 from mindspore.common.api import ms_function
 from mindspore.common.tensor import Tensor
@@ -127,14 +125,15 @@ def _update_run_op(
     return op_cast(next_param, F.dtype(param))
 
 
-def _check_param_value(beta1, beta2, eps, prim_name):
+def _check_param_value(beta1, beta2, eps, use_locking, prim_name):
     """Check the type of inputs."""
-    validator.check_value_type("beta1", beta1, [float], prim_name)
-    validator.check_value_type("beta2", beta2, [float], prim_name)
-    validator.check_value_type("eps", eps, [float], prim_name)
-    validator.check_float_range(beta1, 0.0, 1.0, Rel.INC_NEITHER, "beta1", prim_name)
-    validator.check_float_range(beta2, 0.0, 1.0, Rel.INC_NEITHER, "beta2", prim_name)
-    validator.check_positive_float(eps, "eps", prim_name)
+    assert isinstance(beta1, float), f"For '{prim_name}', the type of 'beta1' must be 'float', but got type '{type(beta1).__name__}'."
+    assert isinstance(beta2, float), f"For '{prim_name}', the type of 'beta2' must be 'float', but got type '{type(beta2).__name__}'."
+    assert isinstance(eps, float), f"For '{prim_name}', the type of 'eps' must be 'float', but got type '{type(eps).__name__}'."
+    assert 0.0 < beta1 < 1.0, f"For '{prim_name}', the range of 'beta1' must be (0.0, 1.0), but got {beta1}."
+    assert 0.0 < beta2 < 1.0, f"For '{prim_name}', the range of 'beta2' must be (0.0, 1.0), but got {beta2}."
+    assert eps > 0, f"For '{prim_name}', the 'eps' must be positive, but got {eps}."
+    assert isinstance(use_locking, bool), f"For '{prim_name}', the type of 'use_locking' must be 'bool', but got type '{type(use_locking).__name__}'."
 
 
 class Adan(Optimizer):
@@ -161,8 +160,7 @@ def __init__(
             learning_rate, params, weight_decay=weight_decay, loss_scale=loss_scale
         )  # Optimized inherit weight decay is bloaked. weight decay is computed in this py.
 
-        _check_param_value(beta1, beta2, eps, self.cls_name)
-        validator.check_value_type("use_locking", use_locking, [bool], self.cls_name)
+        _check_param_value(beta1, beta2, eps, use_locking, self.cls_name)
 
         self.beta1 = Tensor(beta1, mstype.float32)
         self.beta2 = Tensor(beta2, mstype.float32)

mindocr/optim/lion.py

@@ -2,8 +2,6 @@
 
 import mindspore as ms
 from mindspore import ops
-from mindspore._checkparam import Rel
-from mindspore._checkparam import Validator as validator
 from mindspore.common.initializer import initializer
 from mindspore.common.parameter import Parameter
 from mindspore.common.tensor import Tensor
@@ -13,10 +11,10 @@
 
 def _check_param_value(beta1, beta2, prim_name):
     """Check the type of inputs."""
-    validator.check_value_type("beta1", beta1, [float], prim_name)
-    validator.check_value_type("beta2", beta2, [float], prim_name)
-    validator.check_float_range(beta1, 0.0, 1.0, Rel.INC_NEITHER, "beta1", prim_name)
-    validator.check_float_range(beta2, 0.0, 1.0, Rel.INC_NEITHER, "beta2", prim_name)
+    assert isinstance(beta1, float), f"For '{prim_name}', the type of 'beta1' must be 'float', but got type '{type(beta1).__name__}'."
+    assert isinstance(beta2, float), f"For '{prim_name}', the type of 'beta2' must be 'float', but got type '{type(beta2).__name__}'."
+    assert 0.0 < beta1 < 1.0, f"For '{prim_name}', the range of 'beta1' must be (0.0, 1.0), but got {beta1}."
+    assert 0.0 < beta2 < 1.0, f"For '{prim_name}', the range of 'beta2' must be (0.0, 1.0), but got {beta2}."
 
 
 _grad_scale = ops.MultitypeFuncGraph("grad_scale")

mindocr/optim/nadam.py

@@ -3,8 +3,6 @@
 
 import mindspore as ms
 from mindspore import ops
-from mindspore._checkparam import Rel
-from mindspore._checkparam import Validator as validator
 from mindspore.common.api import ms_function
 from mindspore.common.initializer import initializer
 from mindspore.common.parameter import Parameter
@@ -15,12 +13,12 @@
 
 def _check_param_value(beta1, beta2, eps, prim_name):
     """Check the type of inputs."""
-    validator.check_value_type("beta1", beta1, [float], prim_name)
-    validator.check_value_type("beta2", beta2, [float], prim_name)
-    validator.check_value_type("eps", eps, [float], prim_name)
-    validator.check_float_range(beta1, 0.0, 1.0, Rel.INC_NEITHER, "beta1", prim_name)
-    validator.check_float_range(beta2, 0.0, 1.0, Rel.INC_NEITHER, "beta2", prim_name)
-    validator.check_positive_float(eps, "eps", prim_name)
+    assert isinstance(beta1, float), f"For '{prim_name}', the type of 'beta1' must be 'float', but got type '{type(beta1).__name__}'."
+    assert isinstance(beta2, float), f"For '{prim_name}', the type of 'beta2' must be 'float', but got type '{type(beta2).__name__}'."
+    assert isinstance(eps, float), f"For '{prim_name}', the type of 'eps' must be 'float', but got type '{type(eps).__name__}'."
+    assert 0.0 < beta1 < 1.0, f"For '{prim_name}', the range of 'beta1' must be (0.0, 1.0), but got {beta1}."
+    assert 0.0 < beta2 < 1.0, f"For '{prim_name}', the range of 'beta2' must be (0.0, 1.0), but got {beta2}."
+    assert eps > 0, f"For '{prim_name}', the 'eps' must be positive, but got {eps}."
 
 
 _scaler_one = Tensor(1, ms.float32)