[tests]Improved math tests and modified some file formats under test/python #535

Old-cpu · 2025-08-24T12:04:00Z

Improved math operation tests and modified the following file formats

test_sigmoid.py
test_relu.py
test_max_pool2d.py
test_iota.py

linuxlonelyeagle · 2025-08-28T01:53:43Z

tests/Python/test_erf.py

+# CHECK: return %{{.*}}
+# CHECK: }
+# CHECK: }
+


Avoid unnecessary checks whenever possible.Pay attention to the aesthetics of the formatting; the following approach should look much nicer.
Here you should not use {{.*}}; you should capture the variable name.

# CHECK-LABEL: func @forward # CHECK: %{{.*}} = math.erf # CHECK: return %{{.*}} # CHECK: }

linuxlonelyeagle

Leave a few simple comments.

linuxlonelyeagle · 2025-08-28T01:56:18Z

tests/Python/test_iota.py

 print(graph._imported_module)
+
 # CHECK: module {
 # CHECK-LABEL: func.func @forward


The same goes for the test below; I think you can modify them together.

Thank you for your suggestion , I'll continue to improve this PR

linuxlonelyeagle · 2025-08-30T16:01:23Z

@Old-cpu use rebase, don't use merge

Old-cpu · 2025-08-30T16:13:04Z

@Old-cpu use rebase, don't use merge
Got it, I'll edit it tomorrow

linuxlonelyeagle · 2025-09-01T05:44:49Z

Let me give an example.

# CHECK-LABEL: func.func @forward
# CHECK: %{{[a-zA-Z0-9_]+}} = math.cos 
# CHECK: return %{{[a-zA-Z0-9_]+}}
# CHECK: }

You should let it become.

# CHECK-LABEL: func.func @forward
#       CHECK: %[[cos:.*]] = math.cos 
#       CHECK: return %[[cos]]
#       CHECK: }

Old-cpu · 2025-09-01T05:54:59Z

I understand what you mean, I will modify it immediately

linuxlonelyeagle

@zhanghb97

linuxlonelyeagle · 2025-09-03T06:34:12Z

tests/Python/test_cos.py

+# CHECK-LABEL: func.func @forward
+#       CHECK: %[[cos:.*]] = math.cos 
+#       CHECK: return %[[cos]]
+


I was wondering if you left two blank lines at the end—just one would suffice.

I was wondering if you left two blank lines at the end—just one would suffice.

Yes, I left two blank lines locally, but if I only leave one blank line, there will be no blank line when I submit it.

zhanghb97 · 2025-09-17T23:30:25Z

Please check the following error from the CI:

FAIL: BUDDY :: Python/test_max_pool2d.py (79 of 97)
******************** TEST 'BUDDY :: Python/test_max_pool2d.py' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
RUN: at line 1: /home/buddy/miniconda3/envs/buddy/bin/python3 /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py 2>&1 | /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/llvm/build/bin/FileCheck /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py
+ /home/buddy/miniconda3/envs/buddy/bin/python3 /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py
+ /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/llvm/build/bin/FileCheck /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py
/home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py:40:10: error: CHECK: expected string not found in input
# CHECK: return %[[transpose_2]]s
         ^
<stdin>:7:21: note: scanning from here
 %4 = tosa.transpose %2, %3 : (tensor<1x214x160x3xf32>, tensor<4xi32>) -> tensor<1x3x214x160xf32>
                    ^
<stdin>:7:21: note: with "transpose_2" equal to "4"
 %4 = tosa.transpose %2, %3 : (tensor<1x214x160x3xf32>, tensor<4xi32>) -> tensor<1x3x214x160xf32>
                    ^
<stdin>:8:2: note: possible intended match here
 return %4 : tensor<1x3x214x160xf32>
 ^

Input file: <stdin>
Check file: /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py

-dump-input=help explains the following input dump.

Input was:
<<<<<<
            1: module { 
            2:  func.func @forward(%arg0: tensor<1x3x640x480xf32>) -> tensor<1x3x214x160xf32> { 
            3:  %0 = "tosa.const"() <{value = dense<[0, 2, 3, 1]> : tensor<4xi32>}> : () -> tensor<4xi32> 
            4:  %1 = tosa.transpose %arg0, %0 : (tensor<1x3x640x480xf32>, tensor<4xi32>) -> tensor<1x640x480x3xf32> 
            5:  %2 = tosa.max_pool2d %1 {kernel = array<i64: 5, 5>, pad = array<i64: 2, 2, 2, 2>, stride = array<i64: 3, 3>} : (tensor<1x640x480x3xf32>) -> tensor<1x214x160x3xf32> 
            6:  %3 = "tosa.const"() <{value = dense<[0, 3, 1, 2]> : tensor<4xi32>}> : () -> tensor<4xi32> 
            7:  %4 = tosa.transpose %2, %3 : (tensor<1x214x160x3xf32>, tensor<4xi32>) -> tensor<1x3x214x160xf32> 
check:40'0                         X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ error: no match found
check:40'1                                                                                                        with "transpose_2" equal to "4"
            8:  return %4 : tensor<1x3x214x160xf32> 
check:40'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
check:40'2      ?                                    possible intended match
            9:  } 
check:40'0     ~~~
           10: } 
check:40'0     ~~
           11:  
check:40'0     ~
>>>>>>

--

********************
Testing:  0.. 10.. 20.. 30.. 40.. 50.. 60.. 70.. 80.. 90.. 
********************
Failed Tests (1):
  BUDDY :: Python/test_max_pool2d.py


Testing Time: 66.88s

Total Discovered Tests: 97
  Passed: 96 (98.97%)
  Failed:  1 (1.03%)
FAILED: tests/CMakeFiles/check-tests /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/build/tests/CMakeFiles/check-tests 
cd /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/build/tests && /home/buddy/miniconda3/envs/buddy/bin/python3 /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/llvm/build/./bin/llvm-lit -sv /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/build/tests
ninja: build stopped: subcommand failed.

Old-cpu · 2025-09-23T03:58:57Z

Please check the following error from the CI:请检查来自 CI 的以下错误：

FAIL: BUDDY :: Python/test_max_pool2d.py (79 of 97)
******************** TEST 'BUDDY :: Python/test_max_pool2d.py' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
RUN: at line 1: /home/buddy/miniconda3/envs/buddy/bin/python3 /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py 2>&1 | /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/llvm/build/bin/FileCheck /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py
+ /home/buddy/miniconda3/envs/buddy/bin/python3 /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py
+ /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/llvm/build/bin/FileCheck /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py
/home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py:40:10: error: CHECK: expected string not found in input
# CHECK: return %[[transpose_2]]s
         ^
<stdin>:7:21: note: scanning from here
 %4 = tosa.transpose %2, %3 : (tensor<1x214x160x3xf32>, tensor<4xi32>) -> tensor<1x3x214x160xf32>
                    ^
<stdin>:7:21: note: with "transpose_2" equal to "4"
 %4 = tosa.transpose %2, %3 : (tensor<1x214x160x3xf32>, tensor<4xi32>) -> tensor<1x3x214x160xf32>
                    ^
<stdin>:8:2: note: possible intended match here
 return %4 : tensor<1x3x214x160xf32>
 ^

Input file: <stdin>
Check file: /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/tests/Python/test_max_pool2d.py

-dump-input=help explains the following input dump.

Input was:
<<<<<<
            1: module { 
            2:  func.func @forward(%arg0: tensor<1x3x640x480xf32>) -> tensor<1x3x214x160xf32> { 
            3:  %0 = "tosa.const"() <{value = dense<[0, 2, 3, 1]> : tensor<4xi32>}> : () -> tensor<4xi32> 
            4:  %1 = tosa.transpose %arg0, %0 : (tensor<1x3x640x480xf32>, tensor<4xi32>) -> tensor<1x640x480x3xf32> 
            5:  %2 = tosa.max_pool2d %1 {kernel = array<i64: 5, 5>, pad = array<i64: 2, 2, 2, 2>, stride = array<i64: 3, 3>} : (tensor<1x640x480x3xf32>) -> tensor<1x214x160x3xf32> 
            6:  %3 = "tosa.const"() <{value = dense<[0, 3, 1, 2]> : tensor<4xi32>}> : () -> tensor<4xi32> 
            7:  %4 = tosa.transpose %2, %3 : (tensor<1x214x160x3xf32>, tensor<4xi32>) -> tensor<1x3x214x160xf32> 
check:40'0                         X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ error: no match found
check:40'1                                                                                                        with "transpose_2" equal to "4"
            8:  return %4 : tensor<1x3x214x160xf32> 
check:40'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
check:40'2      ?                                    possible intended match
            9:  } 
check:40'0     ~~~
           10: } 
check:40'0     ~~
           11:  
check:40'0     ~
>>>>>>

--

********************
Testing:  0.. 10.. 20.. 30.. 40.. 50.. 60.. 70.. 80.. 90.. 
********************
Failed Tests (1):
  BUDDY :: Python/test_max_pool2d.py


Testing Time: 66.88s

Total Discovered Tests: 97
  Passed: 96 (98.97%)
  Failed:  1 (1.03%)
FAILED: tests/CMakeFiles/check-tests /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/build/tests/CMakeFiles/check-tests 
cd /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/build/tests && /home/buddy/miniconda3/envs/buddy/bin/python3 /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/llvm/build/./bin/llvm-lit -sv /home/buddy/actions-runner/_work/buddy-mlir/buddy-mlir/build/tests
ninja: build stopped: subcommand failed.

OK, it has been modified. Please review it when you have time.

linuxlonelyeagle reviewed Aug 28, 2025

View reviewed changes

linuxlonelyeagle suggested changes Aug 28, 2025

View reviewed changes

Old-cpu force-pushed the main branch from b3b2116 to 6fea99d Compare September 1, 2025 04:40

Old-cpu requested a review from linuxlonelyeagle September 1, 2025 04:40

[tests]Improve math operator testing

4e746cc

Old-cpu force-pushed the main branch from 6fea99d to 4e746cc Compare September 1, 2025 05:09

[tests] Fix formatting issues in the file

6284215

linuxlonelyeagle approved these changes Sep 3, 2025

View reviewed changes

linuxlonelyeagle mentioned this pull request Sep 17, 2025

[tests/Python] Add ResNet18 test #548

Closed

zhanghb97 added ci Related to CI/CD final review labels Sep 17, 2025

Fixed bug in test_max_pool2d.py

f623e6e

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[tests]Improved math tests and modified some file formats under test/python #535

[tests]Improved math tests and modified some file formats under test/python #535

Old-cpu commented Aug 24, 2025

Uh oh!

linuxlonelyeagle Aug 28, 2025 •

edited

Loading

Uh oh!

linuxlonelyeagle left a comment

Uh oh!

linuxlonelyeagle Aug 28, 2025

Uh oh!

Old-cpu Aug 28, 2025

Uh oh!

linuxlonelyeagle commented Aug 30, 2025

Uh oh!

Old-cpu commented Aug 30, 2025

Uh oh!

linuxlonelyeagle commented Sep 1, 2025

Uh oh!

Old-cpu commented Sep 1, 2025

Uh oh!

linuxlonelyeagle left a comment

Uh oh!

linuxlonelyeagle Sep 3, 2025

Uh oh!

Old-cpu Sep 3, 2025

Uh oh!

zhanghb97 commented Sep 17, 2025

Uh oh!

Old-cpu commented Sep 23, 2025

Uh oh!

Uh oh!

[tests]Improved math tests and modified some file formats under test/python #535

Are you sure you want to change the base?

[tests]Improved math tests and modified some file formats under test/python #535

Conversation

Old-cpu commented Aug 24, 2025

Uh oh!

linuxlonelyeagle Aug 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

linuxlonelyeagle left a comment

Choose a reason for hiding this comment

Uh oh!

linuxlonelyeagle Aug 28, 2025

Choose a reason for hiding this comment

Uh oh!

Old-cpu Aug 28, 2025

Choose a reason for hiding this comment

Uh oh!

linuxlonelyeagle commented Aug 30, 2025

Uh oh!

Old-cpu commented Aug 30, 2025

Uh oh!

linuxlonelyeagle commented Sep 1, 2025

Uh oh!

Old-cpu commented Sep 1, 2025

Uh oh!

linuxlonelyeagle left a comment

Choose a reason for hiding this comment

Uh oh!

linuxlonelyeagle Sep 3, 2025

Choose a reason for hiding this comment

Uh oh!

Old-cpu Sep 3, 2025

Choose a reason for hiding this comment

Uh oh!

zhanghb97 commented Sep 17, 2025

Uh oh!

Old-cpu commented Sep 23, 2025

Uh oh!

Uh oh!

linuxlonelyeagle Aug 28, 2025 •

edited

Loading