feat(pt): support eval_typeebd for DeepEval (deepmodeling#4110)

Closes deepmodeling#3608.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Introduced a new method for evaluating type embeddings, enhancing
model capabilities in handling type information.
  
- **Bug Fixes**
	- Improved error handling for unsupported type embedding scenarios.

- **Tests**
- Added a new test to validate the functionality and output of the type
embedding evaluation method, ensuring expected performance.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Signed-off-by: Jinzhe Zeng <[email protected]>

deepmd/pt/infer/deep_eval.py

@@ -46,6 +46,9 @@
 from deepmd.pt.model.model import (
     get_model,
 )
+from deepmd.pt.model.network.network import (
+    TypeEmbedNetConsistent,
+)
 from deepmd.pt.train.wrapper import (
     ModelWrapper,
 )
@@ -61,6 +64,7 @@
     RESERVED_PRECISON_DICT,
 )
 from deepmd.pt.utils.utils import (
+    to_numpy_array,
     to_torch_tensor,
 )
 
@@ -556,6 +560,36 @@ def _get_output_shape(self, odef, nframes, natoms):
         else:
             raise RuntimeError("unknown category")
 
+    def eval_typeebd(self) -> np.ndarray:
+        """Evaluate output of type embedding network by using this model.
+
+        Returns
+        -------
+        np.ndarray
+            The output of type embedding network. The shape is [ntypes, o_size] or [ntypes + 1, o_size],
+            where ntypes is the number of types, and o_size is the number of nodes
+            in the output layer. If there are multiple type embedding networks,
+            these outputs will be concatenated along the second axis.
+
+        Raises
+        ------
+        KeyError
+            If the model does not enable type embedding.
+
+        See Also
+        --------
+        deepmd.pt.model.network.network.TypeEmbedNetConsistent :
+            The type embedding network.
+        """
+        out = []
+        for mm in self.dp.model["Default"].modules():
+            if mm.original_name == TypeEmbedNetConsistent.__name__:
+                out.append(mm(DEVICE))
+        if not out:
+            raise KeyError("The model has no type embedding networks.")
+        typeebd = torch.cat(out, dim=1)
+        return to_numpy_array(typeebd)
+
 
 # For tests only
 def eval_model(

source/tests/pt/model/test_deeppot.py

@@ -110,6 +110,14 @@ def test_uni(self):
         self.assertIsInstance(dp, DeepPot)
         # its methods has been tested in test_dp_test
 
+    def test_eval_typeebd(self):
+        dp = DeepPot(str(self.model))
+        eval_typeebd = dp.eval_typeebd()
+        self.assertEqual(
+            eval_typeebd.shape, (len(self.config["model"]["type_map"]) + 1, 8)
+        )
+        np.testing.assert_allclose(eval_typeebd[-1], np.zeros_like(eval_typeebd[-1]))
+
 
 class TestDeepPotFrozen(TestDeepPot):
     def setUp(self):