MAINT: adjustments to test_logistic.py::test_dtype_match (scikit-learn#13645)

Author: oleksandr-pavlyk

sklearn/linear_model/tests/test_logistic.py

@@ -1311,32 +1311,42 @@ def test_dtype_match(solver, multi_class):
  X_64 = np.array(X).astype(np.float64)
  y_64 = np.array(Y1).astype(np.float64)
  X_sparse_32 = sp.csr_matrix(X, dtype=np.float32)
+ solver_tol = 5e-4
 
+ lr_templ = LogisticRegression(
+ solver=solver, multi_class=multi_class,
+ random_state=42, tol=solver_tol, fit_intercept=True)
  # Check type consistency
- lr_32 = LogisticRegression(solver=solver, multi_class=multi_class,
- random_state=42)
+ lr_32 = clone(lr_templ)
  lr_32.fit(X_32, y_32)
  assert_equal(lr_32.coef_.dtype, X_32.dtype)
 
  # check consistency with sparsity
- lr_32_sparse = LogisticRegression(solver=solver,
- multi_class=multi_class,
- random_state=42)
+ lr_32_sparse = clone(lr_templ)
  lr_32_sparse.fit(X_sparse_32, y_32)
  assert_equal(lr_32_sparse.coef_.dtype, X_sparse_32.dtype)
 
  # Check accuracy consistency
- lr_64 = LogisticRegression(solver=solver, multi_class=multi_class,
- random_state=42)
+ lr_64 = clone(lr_templ)
  lr_64.fit(X_64, y_64)
  assert_equal(lr_64.coef_.dtype, X_64.dtype)
 
- rtol = 1e-6
+ # solver_tol bounds the norm of the loss gradient
+ # dw ~= inv(H)*grad ==> |dw| ~= |inv(H)| * solver_tol, where H - hessian
+ #
+ # See https://github.com/scikit-learn/scikit-learn/pull/13645
+ #
+ # with Z = np.hstack((np.ones((3,1)), np.array(X)))
+ # In [8]: np.linalg.norm(np.diag([0,2,2]) + np.linalg.inv((Z.T @ Z)/4))
+ # Out[8]: 1.7193336918135917
+
+ # factor of 2 to get the ball diameter
+ atol = 2 * 1.72 * solver_tol
  if os.name == 'nt' and _IS_32BIT:
  # FIXME
- rtol = 1e-2
+ atol = 1e-2
 
- assert_allclose(lr_32.coef_, lr_64.coef_.astype(np.float32), rtol=rtol)
+ assert_allclose(lr_32.coef_, lr_64.coef_.astype(np.float32), atol=atol)
 
 
 def test_warm_start_converge_LR():