Fix: Allow hashing of metrics with lists in their state (PL^5939)

* Fix: Allow hashing of metrics with lists in their state

* Add test case and modify semantics of Metric __hash__ in order to be compatible with structural equality checks

* Fix pep8 style issue

Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
(cherry picked from commit a1c26d36e4260543f8642a4fb581f04c5b09c04c)

pytorch_lightning/metrics/metric.py

@@ -333,7 +333,13 @@ class Metric(nn.Module, ABC):
         hash_vals = [self.__class__.__name__]
 
         for key in self._defaults.keys():
-            hash_vals.append(getattr(self, key))
+            val = getattr(self, key)
+            # Special case: allow list values, so long
+            # as their elements are hashable
+            if hasattr(val, '__iter__') and not isinstance(val, torch.Tensor):
+                hash_vals.extend(val)
+            else:
+                hash_vals.append(val)
 
         return hash(tuple(hash_vals))
 

tests/metrics/test_metric.py

@@ -154,6 +154,32 @@ def test_compute():
     assert a.compute() == 5
 
 
+def test_hash():
+
+    class A(Dummy):
+        pass
+
+    class B(DummyList):
+        pass
+
+    a1 = A()
+    a2 = A()
+    assert hash(a1) != hash(a2)
+
+    b1 = B()
+    b2 = B()
+    assert hash(b1) == hash(b2)
+    assert isinstance(b1.x, list) and len(b1.x) == 0
+    b1.x.append(torch.tensor(5))
+    assert isinstance(hash(b1), int)  # <- check that nothing crashes
+    assert isinstance(b1.x, list) and len(b1.x) == 1
+    b2.x.append(torch.tensor(5))
+    # Sanity:
+    assert isinstance(b2.x, list) and len(b2.x) == 1
+    # Now that they have tensor contents, they should have different hashes:
+    assert hash(b1) != hash(b2)
+
+
 def test_forward():
 
     class A(Dummy):