项目文件夹

文件
2026-07-13 12:40:42 +08:00

104 行
2.5 KiB
Python

# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import re
from type_mapping import attr_types_map, input_types_map, output_type_map
# tests for typename
def is_input(s):
return s in input_types_map
def is_attr(s):
return s in attr_types_map
def is_output(s):
return s in output_type_map
def is_vec(s):
return s.endswith("[]")
def is_scalar(s):
return re.match(r"Scalar(\(\w+\))*", s) is not None
def is_intarray(s):
return s == 'IntArray'
def is_datatype(s):
return s == 'DataType'
def is_initializer_list(s):
return s == "{}"
def is_base_op(op):
return "kernel" in op and "infer_meta" in op
# this func describe a op that only has composite implementation,
# without kernel implementation. kernel implementation include
# other op (invoke) or c++ kernel (kernel + infermeta)
def is_only_composite_op(op):
return "composite" in op and "kernel" not in op and "invoke" not in op
# this func describe a op that has composite implementation,
# maybe also has kernel implementation.
def is_composite_op(op):
return "composite" in op
def supports_selected_rows_kernel(op):
return is_base_op(op) and len(op["kernel"]["func"]) == 2
def supports_inplace(op):
return op['inplace'] is not None and len(op['inplace']) == 1
def supports_no_need_buffer(op):
for input in op["inputs"]:
if input["no_need_buffer"]:
return True
return False
def is_tensor_list(s):
return s == 'Tensor[]'
def exist_mutable_attribute(attributes):
for attribute in attributes:
if (
is_scalar(attribute['typename'])
or is_intarray(attribute['typename'])
) and attribute.get('support_tensor', False):
return True
else:
return False
def is_mutable_attribute(attribute):
return (
is_scalar(attribute['typename']) or is_intarray(attribute['typename'])
) and attribute.get('support_tensor', False)