support init_frz_model for hybrid descriptor (#1112)

* support init_frz_model for hybrid descriptor

Refactors some methods to implement it.
Also fixes some typos.

* rename `graph_def` to `model_file`

Co-authored-by: Denghui Lu <[email protected]>

Co-authored-by: Denghui Lu <[email protected]>

Author: njzjz

deepmd/descriptor/descriptor.py

@@ -280,26 +280,65 @@ def get_feed_dict(self,
         feed_dict : dict[str, tf.Tensor]
             The output feed_dict of current descriptor
         """
-        # TODO: currently only SeA has this method, but I think the method can be
-        # moved here as it doesn't contain anything related to a specific descriptor
-        raise NotImplementedError
+        feed_dict = {
+            't_coord:0'  :coord_,
+            't_type:0'   :atype_,
+            't_natoms:0' :natoms,
+            't_box:0'    :box,
+            't_mesh:0'   :mesh
+        }
+        return feed_dict
 
     def init_variables(self,
-                       embedding_net_variables: dict
-                       ) -> None:
+                       model_file: str,
+                       suffix : str = "",
+    ) -> None:
         """
         Init the embedding net variables with the given dict
 
         Parameters
         ----------
-        embedding_net_variables
-                The input dict which stores the embedding net variables
+        model_file : str
+            The input model file
+        suffix : str, optional
+            The suffix of the scope
         
         Notes
         -----
         This method is called by others when the descriptor supported initialization from the given variables.
         """
-        # TODO: currently only SeA has this method, but I think the method can be
-        # moved here as it doesn't contain anything related to a specific descriptor
         raise NotImplementedError(
             "Descriptor %s doesn't support initialization from the given variables!" % type(self).__name__)
+
+    def get_tensor_names(self, suffix : str = "") -> Tuple[str]:
+        """Get names of tensors.
+        
+        Parameters
+        ----------
+        suffix : str
+            The suffix of the scope
+
+        Returns
+        -------
+        Tuple[str]
+            Names of tensors
+        """
+        raise NotImplementedError("Descriptor %s doesn't support this property!" % type(self).__name__)
+
+    def pass_tensors_from_frz_model(self,
+                                    *tensors : tf.Tensor,
+    ) -> None:
+        """
+        Pass the descrpt_reshape tensor as well as descrpt_deriv tensor from the frz graph_def
+
+        Parameters
+        ----------
+        *tensors : tf.Tensor
+            passed tensors
+        
+        Notes
+        -----
+        The number of parameters in the method must be equal to the numbers of returns in
+        :meth:`get_tensor_names`.
+        """
+        raise NotImplementedError("Descriptor %s doesn't support this method!" % type(self).__name__)

deepmd/descriptor/hybrid.py

@@ -253,3 +253,55 @@ def enable_compression(self,
         """
         for idx, ii in enumerate(self.descrpt_list):
             ii.enable_compression(min_nbor_dist, model_file, table_extrapolate, table_stride_1, table_stride_2, check_frequency, suffix=f"{suffix}_{idx}")
+
+    def init_variables(self,
+                       model_file : str,
+                       suffix : str = "",
+    ) -> None:
+        """
+        Init the embedding net variables with the given dict
+
+        Parameters
+        ----------
+        model_file : str
+            The input frozen model file
+        suffix : str, optional
+            The suffix of the scope
+        """
+        for idx, ii in enumerate(self.descrpt_list):
+            ii.init_variables(model_file, suffix=f"{suffix}_{idx}")
+
+    def get_tensor_names(self, suffix : str = "") -> Tuple[str]:
+        """Get names of tensors.
+        
+        Parameters
+        ----------
+        suffix : str
+            The suffix of the scope
+
+        Returns
+        -------
+        Tuple[str]
+            Names of tensors
+        """
+        tensor_names = []
+        for idx, ii in enumerate(self.descrpt_list):
+            tensor_names.extend(ii.get_tensor_names(suffix=f"{suffix}_{idx}"))
+        return tuple(tensor_names)
+
+    def pass_tensors_from_frz_model(self,
+                                    *tensors : tf.Tensor,
+    ) -> None:
+        """
+        Pass the descrpt_reshape tensor as well as descrpt_deriv tensor from the frz graph_def
+
+        Parameters
+        ----------
+        *tensors : tf.Tensor
+            passed tensors
+        """
+        jj = 0
+        for ii in self.descrpt_list:
+            n_tensors = len(ii.get_tensor_names())
+            ii.pass_tensors_from_frz_model(*tensors[jj:jj+n_tensors])
+            jj += n_tensors

deepmd/descriptor/se_a.py

@@ -13,7 +13,7 @@
 from deepmd.utils.tabulate import DPTabulate
 from deepmd.utils.type_embed import embed_atom_type
 from deepmd.utils.sess import run_sess
-from deepmd.utils.graph import load_graph_def, get_tensor_by_name_from_graph
+from deepmd.utils.graph import load_graph_def, get_tensor_by_name_from_graph, get_embedding_net_variables
 from .descriptor import Descriptor
 
 class DescrptSeA (Descriptor):
@@ -433,10 +433,10 @@ def build (self,
         tf.summary.histogram('nlist', self.nlist)
 
         self.descrpt_reshape = tf.reshape(self.descrpt, [-1, self.ndescrpt])
-        self.descrpt_reshape = tf.identity(self.descrpt_reshape, name = 'o_rmat')
-        self.descrpt_deriv = tf.identity(self.descrpt_deriv, name = 'o_rmat_deriv')
-        self.rij = tf.identity(self.rij, name = 'o_rij')
-        self.nlist = tf.identity(self.nlist, name = 'o_nlist')
+        self.descrpt_reshape = tf.identity(self.descrpt_reshape, name = 'o_rmat' + suffix)
+        self.descrpt_deriv = tf.identity(self.descrpt_deriv, name = 'o_rmat_deriv' + suffix)
+        self.rij = tf.identity(self.rij, name = 'o_rij' + suffix)
+        self.nlist = tf.identity(self.nlist, name = 'o_nlist' + suffix)
 
         self.dout, self.qmat = self._pass_filter(self.descrpt_reshape, 
                                                  atype,
@@ -456,6 +456,21 @@ def get_rot_mat(self) -> tf.Tensor:
         """
         return self.qmat
 
+    def get_tensor_names(self, suffix : str = "") -> Tuple[str]:
+        """Get names of tensors.
+        
+        Parameters
+        ----------
+        suffix : str
+            The suffix of the scope
+
+        Returns
+        -------
+        Tuple[str]
+            Names of tensors
+        """
+        return (f'o_rmat{suffix}:0', f'o_rmat_deriv{suffix}:0', f'o_rij{suffix}:0', f'o_nlist{suffix}:0')
+
     def pass_tensors_from_frz_model(self,
                                     descrpt_reshape : tf.Tensor,
                                     descrpt_deriv   : tf.Tensor,
@@ -481,60 +496,21 @@ def pass_tensors_from_frz_model(self,
         self.descrpt_deriv = descrpt_deriv
         self.descrpt_reshape = descrpt_reshape
 
-    def get_feed_dict(self,
-                      coord_, 
-                      atype_, 
-                      natoms, 
-                      box, 
-                      mesh):
-        """
-        generate the deed_dict for current descriptor
-
-        Parameters
-        ----------
-        coord_
-                The coordinate of atoms
-        atype_
-                The type of atoms
-        natoms
-                The number of atoms. This tensor has the length of Ntypes + 2
-                natoms[0]: number of local atoms
-                natoms[1]: total number of atoms held by this processor
-                natoms[i]: 2 <= i < Ntypes+2, number of type i atoms
-        box
-                The box. Can be generated by deepmd.model.make_stat_input
-        mesh
-                For historical reasons, only the length of the Tensor matters.
-                if size of mesh == 6, pbc is assumed. 
-                if size of mesh == 0, no-pbc is assumed. 
-
-        Returns
-        -------
-        feed_dict
-                The output feed_dict of current descriptor
-        """
-        feed_dict = {
-            't_coord:0'  :coord_,
-            't_type:0'   :atype_,
-            't_natoms:0' :natoms,
-            't_box:0'    :box,
-            't_mesh:0'   :mesh
-        }
-        return feed_dict
-
-
     def init_variables(self,
-                       embedding_net_variables: dict
+                       model_file : str,
+                       suffix : str = "",
     ) -> None:
         """
         Init the embedding net variables with the given dict
 
         Parameters
         ----------
-        embedding_net_variables
-                The input dict which stores the embedding net variables
+        model_file : str
+            The input frozen model file
+        suffix : str, optional
+            The suffix of the scope
         """
-        self.embedding_net_variables = embedding_net_variables
+        self.embedding_net_variables = get_embedding_net_variables(model_file, suffix = suffix)
 
 
     def prod_force_virial(self, 

deepmd/model/ener.py

@@ -173,10 +173,11 @@ def build (self,
                 name = 'descrpt_attr/ntypes',
                 dtype = tf.int32)
             feed_dict = self.descrpt.get_feed_dict(coord_, atype_, natoms, box, mesh)
-            return_elements = ['o_rmat:0', 'o_rmat_deriv:0', 'o_rij:0', 'o_nlist:0', 'o_descriptor:0']
-            descrpt_reshape, descrpt_deriv, rij, nlist, dout \
+            return_elements = [*self.descrpt.get_tensor_names(), 'o_descriptor:0']
+            imported_tensors \
                 = self._import_graph_def_from_frz_model(frz_model, feed_dict, return_elements)
-            self.descrpt.pass_tensors_from_frz_model(descrpt_reshape, descrpt_deriv, rij, nlist)
+            dout = imported_tensors[-1]
+            self.descrpt.pass_tensors_from_frz_model(*imported_tensors[:-1])
 
 
         if self.srtab is not None :

deepmd/model/tensor.py

@@ -3,7 +3,7 @@
 
 from deepmd.env import tf
 from deepmd.common import ClassArg
-from deepmd.env import global_cvt_2_ener_float, MODEL_VERSION
+from deepmd.env import global_cvt_2_ener_float, MODEL_VERSION, GLOBAL_TF_FLOAT_PRECISION
 from deepmd.env import op_module
 from deepmd.utils.graph import load_graph_def
 from .model_stat import make_stat_input, merge_sys_stat
@@ -138,10 +138,11 @@ def build (self,
                 name = 'descrpt_attr/ntypes',
                 dtype = tf.int32)
             feed_dict = self.descrpt.get_feed_dict(coord_, atype_, natoms, box, mesh)
-            return_elements = ['o_rmat:0', 'o_rmat_deriv:0', 'o_rij:0', 'o_nlist:0', 'o_descriptor:0']
-            descrpt_reshape, descrpt_deriv, rij, nlist, dout \
+            return_elements = [*self.descrpt.get_tensor_names(), 'o_descriptor:0']
+            imported_tensors \
                 = self._import_graph_def_from_frz_model(frz_model, feed_dict, return_elements)
-            self.descrpt.pass_tensors_from_frz_model(descrpt_reshape, descrpt_deriv, rij, nlist)
+            dout = imported_tensors[-1]
+            self.descrpt.pass_tensors_from_frz_model(*imported_tensors[:-1])
 
         rot_mat = self.descrpt.get_rot_mat()
         rot_mat = tf.identity(rot_mat, name = 'o_rot_mat'+suffix)

deepmd/train/trainer.py

@@ -691,7 +691,7 @@ def _init_from_frz_model(self):
         # self.frz_model will control the self.model to import the descriptor from the given frozen model instead of building from scratch...
         # initialize fitting net with the given compressed frozen model
         if self.model_type == 'original_model':
-            self.descrpt.init_variables(get_embedding_net_variables(self.run_opt.init_frz_model))
+            self.descrpt.init_variables(self.run_opt.init_frz_model)
             self.fitting.init_variables(get_fitting_net_variables(self.run_opt.init_frz_model))
             tf.constant("original_model", name = 'model_type', dtype = tf.string)
         elif self.model_type == 'compressed_model':

deepmd/utils/graph.py

@@ -108,7 +108,7 @@ def get_tensor_by_type(node,
     elif data_type == np.float32:
         tensor = np.array(node.float_val)
     else:
-        raise RunTimeError('model compression does not support the half precision')
+        raise RuntimeError('model compression does not support the half precision')
     return tensor
 
 
@@ -139,40 +139,44 @@ def get_embedding_net_nodes_from_graph_def(graph_def: tf.GraphDef, suffix: str =
     return embedding_net_nodes
 
 
-def get_embedding_net_nodes(model_file: str) -> Dict:
+def get_embedding_net_nodes(model_file: str, suffix: str = "") -> Dict:
     """
     Get the embedding net nodes with the given frozen model(model_file)
 
     Parameters
     ----------
     model_file
         The input frozen model path
+    suffix : str, optional
+        The suffix of the scope
    
     Returns
     ----------
     Dict
         The embedding net nodes with the given frozen model
     """
     _, graph_def = load_graph_def(model_file)
-    return get_embedding_net_nodes_from_graph_def(graph_def)
+    return get_embedding_net_nodes_from_graph_def(graph_def, suffix=suffix)
 
 
-def get_embedding_net_variables_from_graph_def(graph_def : tf.GraphDef) -> Dict:
+def get_embedding_net_variables_from_graph_def(graph_def : tf.GraphDef, suffix: str = "") -> Dict:
     """
     Get the embedding net variables with the given tf.GraphDef object
 
     Parameters
     ----------
     graph_def
         The input tf.GraphDef object
+    suffix : str, optional
+        The suffix of the scope
     
     Returns
     ----------
     Dict
         The embedding net variables within the given tf.GraphDef object 
     """
     embedding_net_variables = {}
-    embedding_net_nodes = get_embedding_net_nodes_from_graph_def(graph_def)
+    embedding_net_nodes = get_embedding_net_nodes_from_graph_def(graph_def, suffix=suffix)
     for item in embedding_net_nodes:
         node = embedding_net_nodes[item]
         dtype = tf.as_dtype(node.dtype).as_numpy_dtype
@@ -184,22 +188,24 @@ def get_embedding_net_variables_from_graph_def(graph_def : tf.GraphDef) -> Dict:
         embedding_net_variables[item] = np.reshape(tensor_value, tensor_shape)
     return embedding_net_variables
 
-def get_embedding_net_variables(model_file : str) -> Dict:
+def get_embedding_net_variables(model_file : str, suffix: str = "") -> Dict:
     """
     Get the embedding net variables with the given frozen model(model_file)
 
     Parameters
     ----------
     model_file
         The input frozen model path
+    suffix : str, optional
+        The suffix of the scope
     
     Returns
     ----------
     Dict
         The embedding net variables within the given frozen model
     """
     _, graph_def = load_graph_def(model_file)
-    return get_embedding_net_variables_from_graph_def(graph_def)
+    return get_embedding_net_variables_from_graph_def(graph_def, suffix=suffix)
 
 
 def get_fitting_net_nodes_from_graph_def(graph_def: tf.GraphDef) -> Dict: