#1036: Added pytorch and numba integration test compatible with CPU only VM #485
Conversation
| ''')) | ||
|
|
||
| row = self.query(f"SELECT pytorchbasic.test_pytorch(1000);")[0] | ||
| self.assertIn('Mean Squared Error', row[0]) |
There was a problem hiding this comment.
I would assert if the loss is not NaN, this could hint to something is broken, if it worked before
| / | ||
| ''')) | ||
|
|
||
| row = self.query("SELECT numbabasic.test_numba(10000);")[0] |
There was a problem hiding this comment.
Is it necessary to verify the returned value?
There was a problem hiding this comment.
We should simply out of good practice
There was a problem hiding this comment.
I think not. The return value is some random number. If there is any problem, an exception is thrown and this breaks the test.
There was a problem hiding this comment.
In that case your are assert should be , did we get a row back
There was a problem hiding this comment.
Added an assert for the resulting row
| / | ||
| ''')) | ||
|
|
||
| row = self.query(f"SELECT pytorchbasic.test_pytorch(1000);")[0] | ||
| self.assertIn('Mean Squared Error', row[0]) | ||
|
|
||
| self.assertFalse(math.isnan(row[0])) |
There was a problem hiding this comment.
I think, NaN is transformed to NULL in the DB
There was a problem hiding this comment.
Appearantly the build failed when fetching micromamba. Running the test locally , the assert is correct.
Return the sum of the discrimants and check number
related exasol/script-languages-release#1036