[HOTFIX] Fix default dtype handling in abs function

Ensure that the abs function works correctly when dtype is not provided by defaulting to the tensor's existing dtype. This prevents potential errors caused by missing or None-type dtype inputs.

Author: Leguark

gempy_engine/core/backend_tensor.py

@@ -175,7 +175,7 @@ def _transpose(tensor, axes=None):
         cls.tfnp.concatenate = _concatenate
         cls.tfnp.transpose = _transpose
         cls.tfnp.geomspace = lambda start, stop, step: torch.logspace(start, stop, step, base=10)
-        cls.tfnp.abs = lambda tensor, dtype: tensor.abs().type(dtype)
+        cls.tfnp.abs = lambda tensor, dtype = None: tensor.abs().type(dtype) if dtype is not None else tensor.abs()
 
     @classmethod
     def _wrap_pykeops_functions(cls):