import unittest
import torch
import importlib
import io
import sys


def _dynamic(test, cond, ok, fail):
    if cond:
        test._testMethodName = ok
        test.assertTrue(True)
    else:
        test._testMethodName = fail
        test.fail(fail)


class TestUserCode(unittest.TestCase):

    def setUp(self):
        # Reference tensor for comparison
        self.tensor = torch.tensor([[3, 7, 1],
                                    [2, 5, 9],
                                    [8, 6, 4]], dtype=torch.float32)

    # --- Variable existence ---
    def test_row_sum_exists(self):
        import user_code
        cond = hasattr(user_code, "row_sum")
        _dynamic(
            self,
            cond,
            "`row_sum` variable is defined.",
            "`row_sum` is missing. Compute it using `tensor.sum(dim=1, keepdim=True)`."
        )

    def test_column_mean_exists(self):
        import user_code
        cond = hasattr(user_code, "column_mean")
        _dynamic(
            self,
            cond,
            "`column_mean` variable is defined.",
            "`column_mean` is missing. Compute it using `tensor.mean(dim=0)`."
        )

    def test_row_max_exists(self):
        import user_code
        cond = hasattr(user_code, "row_max")
        _dynamic(
            self,
            cond,
            "`row_max` variable is defined.",
            "`row_max` is missing. Compute it using `tensor.max(dim=1)[0]`."
        )

    def test_normalized_exists(self):
        import user_code
        cond = hasattr(user_code, "normalized_tensor")
        _dynamic(
            self,
            cond,
            "`normalized_tensor` variable is defined.",
            "`normalized_tensor` is missing. Normalize using `tensor / tensor.sum()`."
        )

    # --- Correct calculations ---
    def test_row_sum_correct(self):
        import user_code
        expected = self.tensor.sum(dim=1, keepdim=True)
        cond = torch.allclose(user_code.row_sum, expected)
        _dynamic(
            self,
            cond,
            "`row_sum` is computed correctly.",
            "`row_sum` is incorrect. Use `tensor.sum(dim=1, keepdim=True)`."
        )

    def test_column_mean_correct(self):
        import user_code
        expected = self.tensor.mean(dim=0)
        cond = torch.allclose(user_code.column_mean, expected)
        _dynamic(
            self,
            cond,
            "`column_mean` is computed correctly.",
            "`column_mean` is incorrect. Use `tensor.mean(dim=0)`."
        )

    def test_row_max_correct(self):
        import user_code
        expected = self.tensor.max(dim=1)[0]
        cond = torch.allclose(user_code.row_max, expected)
        _dynamic(
            self,
            cond,
            "`row_max` is computed correctly.",
            "`row_max` is incorrect. Use `tensor.max(dim=1)[0]`."
        )

    def test_normalized_tensor_correct(self):
        import user_code
        expected = self.tensor / self.tensor.sum()
        cond = torch.allclose(user_code.normalized_tensor, expected)
        _dynamic(
            self,
            cond,
            "`normalized_tensor` is computed correctly.",
            "`normalized_tensor` is incorrect. Normalize using `tensor / tensor.sum()`."
        )

    # --- Ensure something was printed ---
    def test_print_output(self):
        import user_code

        buffer = io.StringIO()
        sys.stdout = buffer
        importlib.reload(user_code)
        sys.stdout = sys.__stdout__

        cond = len(buffer.getvalue().strip()) > 0

        _dynamic(
            self,
            cond,
            "Output is successfully printed.",
            "No output detected. Make sure to print all required values."
        )


if __name__ == "__main__":
    unittest.main()


test_code.py

Siirry korkean tason kirjastoista PyTorchin joustavuuteen. Hallitse syväoppimisen perusrakenteet työskentelemällä Tensorien, PyTorchin ydintietorakenteen, kanssa. Tutustu automaattiseen differentiaatioon gradienttien avulla, toteuta takaisinkytkentä ja rakenna kokonaisia neuroverkkoja alusta alkaen. Opettele käsittelemään todellisia tietoaineistoja ja hallitse koko koulutusprosessi – painojen alustuksesta mallin arviointiin.

Haaste: Matemaattisten Operaatioden Suorittaminen

Ratkaisu