Remove unnecessary allclose parameters for consistency.

Author: nilq

flax/nnx/nn/normalization.py

@@ -891,7 +891,7 @@ class WeightNorm(nnx.Module):
 
     >>> w = model.normed_linear.layer_instance.kernel.value
     >>> col_norms = np.linalg.norm(np.array(w), axis=0)
-    >>> np.testing.assert_allclose(col_norms, np.ones(4), rtol=1e-5, atol=1e-5)
+    >>> np.testing.assert_allclose(col_norms, np.ones(4))
 
   Args:
     layer_instance: The layer instance to wrap.