fix: list -> torch.nn.ModuleList

Signed-off-by: zjgemi <liuxin_zijian@163.com>

Author: zjgemi

deepmd/pt/model/atomic_model/density_atomic_model.py

@@ -51,12 +51,12 @@ def __init__(self, descriptor, fitting, type_map, **kwargs):
         self.axis_neuron = self.descriptor.axis_neuron
         neurons = []
         dims = [1 + self.descriptor.repinit_args.tebd_dim] + neurons + [self.descriptor.get_dim_out()]
-        self.grid_embedding_layers = [MLPLayer(
+        self.grid_embedding_layers = torch.nn.ModuleList([MLPLayer(
             dims[i],
             dims[i+1],
             precision=env.DEFAULT_PRECISION,
             activation_function="tanh",
-        ) for i in range(len(neurons)+1)]
+        ) for i in range(len(neurons)+1)])
 
         wanted_shape = (1, self.nnei, 4)
         mean = torch.zeros(