wehub-resource-sync
219 lines
6.3 KiB
Python
219 lines
6.3 KiB
Python
# Copyright (c) 2025 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.
|
|
|
|
from __future__ import annotations
|
|
|
|
import unittest
|
|
|
|
import numpy as np
|
|
from test_case_base import (
|
|
TestCaseBase,
|
|
test_instruction_translator_cache_context,
|
|
)
|
|
|
|
import paddle
|
|
from paddle.jit.sot.psdb import check_no_breakgraph
|
|
|
|
|
|
@check_no_breakgraph
|
|
def unary_api(func, x: np.ndarray, *args, **kwargs):
|
|
return func(x, *args, **kwargs)
|
|
|
|
|
|
@check_no_breakgraph
|
|
def binary_api(
|
|
func: np.ufunc,
|
|
x: np.ndarray | list,
|
|
y: np.ndarray | list,
|
|
*args,
|
|
**kwargs,
|
|
):
|
|
return func(x, y, *args, **kwargs)
|
|
|
|
|
|
@check_no_breakgraph
|
|
def binary_operator(op: str, x: np.ndarray, y: np.ndarray):
|
|
if op == "+":
|
|
return x + y
|
|
elif op == "-":
|
|
return x - y
|
|
elif op == "*":
|
|
return x * y
|
|
elif op == "/":
|
|
return x / y
|
|
|
|
|
|
@check_no_breakgraph
|
|
def get_item(x: np.ndarray, index: int | tuple | np.ndarray):
|
|
return x[index]
|
|
|
|
|
|
@check_no_breakgraph
|
|
def set_item(
|
|
x: np.ndarray,
|
|
index: int | list | np.ndarray,
|
|
value: int | list | np.ndarray,
|
|
):
|
|
x[index] = value
|
|
return x
|
|
|
|
|
|
@check_no_breakgraph
|
|
def grad_fn(pd_x, np_y, np_y2):
|
|
pd_y = paddle.to_tensor(np_y)
|
|
pd_z = paddle.add(pd_x, pd_y)
|
|
pd_y2 = paddle.to_tensor(np_y2).astype("float32")
|
|
pd_z2 = paddle.matmul(pd_z, pd_y2)
|
|
return pd_z2
|
|
|
|
|
|
@check_no_breakgraph
|
|
def demo(pd_x):
|
|
np_y = np.array([[1, 2, 3], [4, 5, 6]])
|
|
np_y2 = np.transpose(np_y)
|
|
pd_y2 = paddle.to_tensor(np_y2)
|
|
pd_z = paddle.matmul(pd_x, pd_y2)
|
|
np_z = pd_z.numpy()
|
|
np_z2 = np.mean(np_z)
|
|
np_z3 = np.add(pd_x.numpy(), np_z2)
|
|
np_z4 = np_z * np_z3[:, :2]
|
|
pd_z2 = paddle.subtract(pd_z, paddle.to_tensor(np_z4))
|
|
return pd_z2
|
|
|
|
|
|
def absolute(x):
|
|
return np.absolute(x)
|
|
|
|
|
|
class TestNumPyArray(TestCaseBase):
|
|
def test_guard(self):
|
|
with test_instruction_translator_cache_context() as ctx:
|
|
self.assertEqual(ctx.translate_count, 0)
|
|
self.assert_results(
|
|
binary_api, np.add, np.array([1, 2]), np.array([3, 4])
|
|
)
|
|
self.assertEqual(ctx.translate_count, 1)
|
|
self.assert_results(
|
|
binary_api,
|
|
np.add,
|
|
np.array([1, 2], dtype=np.int32),
|
|
np.array([3, 4], dtype=np.int32),
|
|
)
|
|
self.assertEqual(ctx.translate_count, 2)
|
|
self.assert_results(
|
|
binary_api, np.add, np.array([1]), np.array([3])
|
|
)
|
|
self.assertEqual(ctx.translate_count, 3)
|
|
self.assert_results(
|
|
binary_api, np.add, np.array([4, 3]), np.array([2, 1])
|
|
)
|
|
self.assertEqual(ctx.translate_count, 3)
|
|
|
|
def test_gradient(self):
|
|
pd_x = paddle.randn([2, 3], dtype="float32")
|
|
pd_x.stop_gradient = False
|
|
self.assert_results_with_grad(
|
|
pd_x,
|
|
grad_fn,
|
|
pd_x,
|
|
np.array([[1, 2, 3], [4, 5, 6]], dtype=np.float32),
|
|
np.ones(shape=(3, 2)),
|
|
)
|
|
|
|
def test_add_01(self):
|
|
x = np.array([1, 2])
|
|
y = np.array([3, 4])
|
|
self.assert_results(binary_api, np.add, x, y)
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_add_02(self):
|
|
x = [1, 2]
|
|
y = [3, 4]
|
|
self.assert_results(binary_api, np.add, x, y)
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_sub(self):
|
|
x = np.array([1, 2])
|
|
y = np.array([3, 4])
|
|
self.assert_results(binary_api, np.subtract, x, y)
|
|
self.assert_results(binary_api, np.subtract, [1, 2], [3, 4])
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_mul(self):
|
|
x = np.array([1, 2])
|
|
y = np.array([3, 4])
|
|
self.assert_results(binary_api, np.multiply, x, y)
|
|
self.assert_results(binary_api, np.multiply, [1, 2], [3, 4])
|
|
self.assert_results_with_grad(binary_api, np.multiply, [1, 2], [3, 4])
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_div(self):
|
|
x = np.array([1, 2])
|
|
y = np.array([3, 4])
|
|
self.assert_results(binary_api, np.divide, x, y)
|
|
self.assert_results(binary_api, np.divide, [1, 2], [3, 4])
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_operator(self):
|
|
x = np.array([1, 2])
|
|
y = np.array([3, 4])
|
|
self.assert_results(binary_operator, '+', x, y)
|
|
self.assert_results(binary_operator, '-', x, y)
|
|
self.assert_results(binary_operator, '*', x, y)
|
|
self.assert_results(binary_operator, '/', x, y)
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_argmax(self):
|
|
x = np.array([[1, 2], [3, 4]])
|
|
self.assert_results(unary_api, np.argmax, x)
|
|
self.assert_results(unary_api, np.argmax, x, axis=0)
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_sum(self):
|
|
x = np.array([[1, 2], [3, 4]])
|
|
self.assert_results(unary_api, np.sum, x)
|
|
self.assert_results(unary_api, np.sum, x, axis=0)
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_getitem(self):
|
|
x = np.array([[1, 2], [3, 4]])
|
|
self.assert_results(get_item, x, 0)
|
|
i = tuple(0, 0)
|
|
self.assert_results(get_item, x, i)
|
|
i2 = np.array([0, 1])
|
|
self.assert_results(get_item, x, i2)
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_setitem(self):
|
|
x = np.array([[1, 2], [3, 4]])
|
|
y = np.array([5, 6])
|
|
self.assert_results(set_item, x, 0, y)
|
|
i = np.array([0])
|
|
self.assert_results(set_item, x, i, y)
|
|
self.assert_results(set_item, x, [0], [5, 6])
|
|
|
|
@unittest.skip("Not supported yet")
|
|
def test_demo(self):
|
|
pd_x = paddle.randn([2, 3], dtype="float32")
|
|
pd_x.stop_gradient = False
|
|
self.assert_results_with_grad(pd_x, demo, pd_x)
|
|
|
|
def test_absolute(self):
|
|
x = np.array([[1, -2], [-3, 4]])
|
|
self.assert_results(absolute, x)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|