[SigureMo]

PR types

Others

PR changes

Others

Describe

一些非 self.assertTrue 开头的「漏网之鱼」，如 self.op_test.assertTrue

做 #46686 时发现报错信息非常不友好，因此修一下

emmm，本地单测运行出了点问题，本地不方便调试，之后再说

[paddle-bot]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[paddle-bot]

很抱歉，经过我们的反复讨论，你的PR暂未达到合入标准，请阅读飞桨原生算子开发规范，你可以重新提交新的PR，我们先将此PR关闭，感谢你的贡献。
Sorry to inform you that through our discussion, your PR fails to meet the merging standard (Reference: Paddle Custom Operator Design Doc). You can also submit an new one. Thank you.

[paddle-bot]

很抱歉，经过我们的反复讨论，你的PR暂未达到合入标准，请阅读飞桨原生算子开发规范，你可以重新提交新的PR，我们先将此PR关闭，感谢你的贡献。
Sorry to inform you that through our discussion, your PR fails to meet the merging standard (Reference: Paddle Custom Operator Design Doc). You can also submit an new one. Thank you.

[SigureMo]

因为本地跑不了单测，利用 CI 解决的一部分问题，其他问题后续逐步解决

完整 log：build.log

不过目测有的问题可能确实解决不了……

[SigureMo]

2022-10-01 17:28:28 ======================================================================
2022-10-01 17:28:28 FAIL: test_check_output (test_isclose_op.TestIscloseOpFloat64)
2022-10-01 17:28:28 ----------------------------------------------------------------------
2022-10-01 17:28:28 Traceback (most recent call last):
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/paddle/fluid/tests/unittests/test_isclose_op.py", line 228, in test_check_output
2022-10-01 17:28:28     self.check_output(check_eager=True)
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/op_test.py", line 1799, in check_output
2022-10-01 17:28:28     check_eager=check_eager)
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/op_test.py", line 1664, in check_output_with_place
2022-10-01 17:28:28     static_checker.check()
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/op_test.py", line 1494, in check
2022-10-01 17:28:28     self.compare_outputs_with_expects()
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/op_test.py", line 1484, in compare_outputs_with_expects
2022-10-01 17:28:28     self.compare_single_output_with_expect(out_name, expect)
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/op_test.py", line 1465, in compare_single_output_with_expect
2022-10-01 17:28:28     self._compare_numpy(name, actual_np, expect_np)
2022-10-01 17:28:28   File "/workspace/Paddle/build/python/op_test.py", line 1448, in _compare_numpy
2022-10-01 17:28:28     " in " + self.checker_name)
2022-10-01 17:28:28   File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/numpy/testing/_private/utils.py", line 1531, in assert_allclose
2022-10-01 17:28:28     verbose=verbose, header=header, equal_nan=equal_nan)
2022-10-01 17:28:28   File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/numpy/testing/_private/utils.py", line 763, in assert_array_compare
2022-10-01 17:28:28     raise AssertionError(msg)
2022-10-01 17:28:28 AssertionError: 
2022-10-01 17:28:28 Not equal to tolerance rtol=1e-05, atol=0
2022-10-01 17:28:28 Output (Out) has diff at Place(cpu) in static checker
2022-10-01 17:28:28 (shapes (1,), (1, 1) mismatch)
2022-10-01 17:28:28  x: array([ True])
2022-10-01 17:28:28  y: array([[ True]])
2022-10-01 17:28:28 

所有的 expect 都比 actual 多一个维度，因此删掉外层 np.array([...])

[SigureMo]

2022-10-01 17:31:33 ======================================================================
2022-10-01 17:31:33 FAIL: test_check_output (test_reduce_op.TestAll8DOp)
2022-10-01 17:31:33 ----------------------------------------------------------------------
2022-10-01 17:31:33 Traceback (most recent call last):
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/paddle/fluid/tests/unittests/test_reduce_op.py", line 365, in test_check_output
2022-10-01 17:31:33     self.check_output(check_eager=True)
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/op_test.py", line 1799, in check_output
2022-10-01 17:31:33     check_eager=check_eager)
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/op_test.py", line 1664, in check_output_with_place
2022-10-01 17:31:33     static_checker.check()
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/op_test.py", line 1494, in check
2022-10-01 17:31:33     self.compare_outputs_with_expects()
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/op_test.py", line 1484, in compare_outputs_with_expects
2022-10-01 17:31:33     self.compare_single_output_with_expect(out_name, expect)
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/op_test.py", line 1465, in compare_single_output_with_expect
2022-10-01 17:31:33     self._compare_numpy(name, actual_np, expect_np)
2022-10-01 17:31:33   File "/workspace/Paddle/build/python/op_test.py", line 1448, in _compare_numpy
2022-10-01 17:31:33     " in " + self.checker_name)
2022-10-01 17:31:33   File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/numpy/testing/_private/utils.py", line 1531, in assert_allclose
2022-10-01 17:31:33     verbose=verbose, header=header, equal_nan=equal_nan)
2022-10-01 17:31:33   File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/numpy/testing/_private/utils.py", line 763, in assert_array_compare
2022-10-01 17:31:33     raise AssertionError(msg)
2022-10-01 17:31:33 AssertionError: 
2022-10-01 17:31:33 Not equal to tolerance rtol=1e-05, atol=1e-05
2022-10-01 17:31:33 Output (Out) has diff at Place(cpu) in static checker
2022-10-01 17:31:33 (shapes (1,), (2, 5, 3, 4, 2) mismatch)
2022-10-01 17:31:33  x: array([False])
2022-10-01 17:31:33  y: array([[[[[False, False],
2022-10-01 17:31:33           [False, False],
2022-10-01 17:31:33           [False, False],...

reduce_all 会使维度为 [1]

Paddle/paddle/phi/infermeta/unary.cc

Lines 2701 to 2722 in 0929bc8

	std::vector<int64_t> out_dim_vector;
	if (keep_dim) {
	for (int64_t i = 0; i < x.dims().size(); ++i) {
	if (reduce_all || dims_set.find(i) != dims_set.end()) {
	out_dim_vector.push_back(1);
	} else {
	out_dim_vector.push_back(x.dims().at(i));
	}
	}
	} else {
	for (int64_t i = 0; i < x.dims().size(); ++i) {
	if (reduce_all || dims_set.find(i) != dims_set.end()) {
	continue;
	} else {
	out_dim_vector.push_back(x.dims().at(i));
	}
	}
	if (out_dim_vector.size() == 0) {
	out_dim_vector.push_back(1);
	}
	}

因此这里需要移除 reduce_all 参数（或者设为默认值 False）才能与 NumPy 计算的 expect 一致

[paddle-bot]

很抱歉，经过我们的反复讨论，你的PR暂未达到合入标准，请阅读飞桨原生算子开发规范，你可以重新提交新的PR，我们先将此PR关闭，感谢你的贡献。
Sorry to inform you that through our discussion, your PR fails to meet the merging standard (Reference: Paddle Custom Operator Design Doc). You can also submit an new one. Thank you.