Merge branch 'master' into feat-se-t

Signed-off-by: Jinzhe Zeng <jinzhe.zeng@ustc.edu.cn>

Author: njzjz

.pre-commit-config.yaml

@@ -29,7 +29,7 @@ repos:
         exclude: ^source/3rdparty
   - repo: https://github.com/astral-sh/ruff-pre-commit
     # Ruff version.
-    rev: v0.14.14
+    rev: v0.15.0
     hooks:
       - id: ruff
         args: ["--fix"]

deepmd/dpmodel/utils/network.py

@@ -886,10 +886,13 @@ def deserialize(cls, data: dict) -> "EmbeddingNet":
         obj = cls(**data)
         # Reinitialize layers from serialized data, using the same layer type
         # that __init__ created (respects subclass overrides via MRO).
-        layer_type = type(obj.layers[0])
-        obj.layers = type(obj.layers)(
-            [layer_type.deserialize(layer) for layer in layers]
-        )
+        if obj.layers:
+            layer_type = type(obj.layers[0])
+            obj.layers = type(obj.layers)(
+                [layer_type.deserialize(layer) for layer in layers]
+            )
+        else:
+            obj.layers = type(obj.layers)([])
         return obj
 
 

deepmd/dpmodel/utils/serialization.py

@@ -106,9 +106,9 @@ def save_dp_model(filename: str, model_dict: dict) -> None:
         with h5py.File(filename, "w") as f:
             model_dict = traverse_model_dict(
                 model_dict,
-                lambda x: f.create_dataset(
-                    f"variable_{variable_counter():04d}", data=x
-                ).name,
+                lambda x: (
+                    f.create_dataset(f"variable_{variable_counter():04d}", data=x).name
+                ),
             )
             save_dict = {
                 **extra_dict,
@@ -118,15 +118,17 @@ def save_dp_model(filename: str, model_dict: dict) -> None:
     elif filename_extension in {".yaml", ".yml"}:
         model_dict = traverse_model_dict(
             model_dict,
-            lambda x: {
-                "@class": "np.ndarray",
-                "@is_variable": True,
-                "@version": 1,
-                "dtype": x.dtype.name,
-                "value": x.tolist(),
-            }
-            if isinstance(x, np.ndarray)
-            else x,
+            lambda x: (
+                {
+                    "@class": "np.ndarray",
+                    "@is_variable": True,
+                    "@version": 1,
+                    "dtype": x.dtype.name,
+                    "value": x.tolist(),
+                }
+                if isinstance(x, np.ndarray)
+                else x
+            ),
         )
         with open(filename, "w") as f:
             yaml.safe_dump(

deepmd/pd/utils/multi_task.py

@@ -150,8 +150,10 @@ def replace_one_item(
     for shared_key in shared_links:
         shared_links[shared_key]["links"] = sorted(
             shared_links[shared_key]["links"],
-            key=lambda x: x["shared_level"]
-            - ("spin" in model_config["model_dict"][x["model_key"]]) * 100,
+            key=lambda x: (
+                x["shared_level"]
+                - ("spin" in model_config["model_dict"][x["model_key"]]) * 100
+            ),
         )
         # little trick to make spin models in the front to be the base models,
         # because its type embeddings are more general.

deepmd/pt/utils/env.py

@@ -34,7 +34,7 @@
     # only linux
     ncpus = len(os.sched_getaffinity(0))
 except AttributeError:
-    ncpus = os.cpu_count()
+    ncpus = os.cpu_count() or 1
 NUM_WORKERS = int(os.environ.get("NUM_WORKERS", min(4, ncpus)))
 if multiprocessing.get_start_method() != "fork":
     # spawn or forkserver does not support NUM_WORKERS > 0 for DataLoader

deepmd/pt/utils/multi_task.py

@@ -155,8 +155,10 @@ def replace_one_item(
     for shared_key in shared_links:
         shared_links[shared_key]["links"] = sorted(
             shared_links[shared_key]["links"],
-            key=lambda x: x["shared_level"]
-            - ("spin" in model_config["model_dict"][x["model_key"]]) * 100,
+            key=lambda x: (
+                x["shared_level"]
+                - ("spin" in model_config["model_dict"][x["model_key"]]) * 100
+            ),
         )
         # little trick to make spin models in the front to be the base models,
         # because its type embeddings are more general.

deepmd/pt_expt/utils/env.py

@@ -34,7 +34,7 @@
     # only linux
     ncpus = len(os.sched_getaffinity(0))
 except AttributeError:
-    ncpus = os.cpu_count()
+    ncpus = os.cpu_count() or 1
 NUM_WORKERS = int(os.environ.get("NUM_WORKERS", min(4, ncpus)))
 if multiprocessing.get_start_method() != "fork":
     # spawn or forkserver does not support NUM_WORKERS > 0 for DataLoader

source/tests/common/dpmodel/test_network.py

@@ -250,6 +250,44 @@ def test_trainable_parameter_variants(self) -> None:
         self.assertTrue(net_mixed.layers[0].trainable)
         self.assertFalse(net_mixed.layers[1].trainable)
 
+    def test_empty_layers_round_trip(self) -> None:
+        """Test EmbeddingNet with empty neuron list (edge case for deserialize).
+
+        This tests the fix for IndexError when neuron=[] results in empty layers.
+        The deserialize method should handle this case without trying to access
+        layers[0] when the list is empty.
+        """
+        in_dim = 4
+        neuron = []  # Empty neuron list
+
+        # Create network with empty layers
+        net = EmbeddingNet(
+            in_dim=in_dim,
+            neuron=neuron,
+            activation_function="tanh",
+            resnet_dt=True,
+            precision="float64",
+        )
+
+        # Verify it has no layers
+        self.assertEqual(len(net.layers), 0)
+
+        # Serialize and deserialize
+        serialized = net.serialize()
+        net_restored = EmbeddingNet.deserialize(serialized)
+
+        # Verify restored network also has no layers
+        self.assertEqual(len(net_restored.layers), 0)
+        self.assertEqual(net_restored.in_dim, in_dim)
+        self.assertEqual(net_restored.neuron, neuron)
+
+        # Verify forward pass works (should return input unchanged)
+        rng = np.random.default_rng()
+        x = rng.standard_normal((5, in_dim))
+        out = net_restored.call(x)
+        # With no layers, output should equal input
+        np.testing.assert_allclose(out, x)
+
 
 class TestFittingNet(unittest.TestCase):
     def test_fitting_net(self) -> None:

source/tests/pt_expt/utils/test_network.py

@@ -165,7 +165,7 @@ def test_cross_backend_consistency(self) -> None:
         # Test forward pass
         rng = np.random.default_rng()
         x_np = rng.standard_normal((5, self.in_dim))
-        x_torch = torch.from_numpy(x_np)
+        x_torch = torch.from_numpy(x_np).to(env.DEVICE)
 
         out_dp = dp_net.call(x_np)
         out_pt = pt_net(x_torch).detach().cpu().numpy()