override forward_factored instead

i6_models/parts/factored_hybrid/triphone.py

@@ -42,7 +42,11 @@ def __init__(self, cfg: FactoredTriphoneBlockV1Config):
             activation=cfg.activation,
         )
 
-    def forward(
+    # update type definitions
+    def forward(self, *args, **kwargs) -> Tuple[Tensor, Tensor, Tensor, Tensor, Tensor]:
+        return super().forward(*args, **kwargs)
+
+    def forward_factored(
         self,
         features: Tensor,  # B, T, F
         contexts_left: Tensor,  # B, T
@@ -58,7 +62,7 @@ def forward(
         :return: tuple of logits for p(c|l,x), p(l|x), p(r|c,l,x) and the embedded left context and center state values.
         """
 
-        logits_center, logits_left, contexts_left_embedded = super().forward(features, contexts_left)
+        logits_center, logits_left, contexts_left_embedded = super().forward_factored(features, contexts_left)
 
         # This logic is very similar to FactoredDiphoneBlockV2.forward, but not the same.
         # This class computes `p(r|c,l,h(x))` while FactoredDiphoneBlockV2 computes `p(r|c,h(x))`.

tests/test_fh.py

@@ -129,9 +129,7 @@ def test_tri_output_shape_and_norm():
             contexts_center = torch.randint(0, tri_block.num_center, (b, t))
             encoder_output = torch.rand((b, t, n_in))
             output_center, output_left, output_right, _, _ = tri_block(
-                features=encoder_output,
-                contexts_left=contexts_left,
-                contexts_center=contexts_center,
+                features=encoder_output, contexts_left=contexts_left, contexts_center=contexts_center
             )
             assert output_left.shape == (b, t, n_ctx)
             assert output_center.shape == (b, t, cdim)