deepmd.tf.model.frozen

deepmd.tf.model.frozen#

Classes#

FrozenModel

Load model from a frozen model, which cannot be trained.

Module Contents#

class deepmd.tf.model.frozen.FrozenModel(model_file: str, **kwargs)[source]#

Bases: deepmd.tf.model.model.Model

Load model from a frozen model, which cannot be trained.

Parameters:

model_filestr: The path to the frozen model

model_file[source]#

model[source]#

build(coord_: deepmd.tf.env.tf.Tensor, atype_: deepmd.tf.env.tf.Tensor, natoms: deepmd.tf.env.tf.Tensor, box: deepmd.tf.env.tf.Tensor, mesh: deepmd.tf.env.tf.Tensor, input_dict: dict, frz_model: str | None = None, ckpt_meta: str | None = None, suffix: str = '', reuse: bool | enum.Enum | None = None) → dict[source]#

Build the model.

Parameters:

coord_tf.Tensor: The coordinates of atoms
atype_tf.Tensor: The atom types of atoms
natomstf.Tensor: The number of atoms
boxtf.Tensor: The box vectors
meshtf.Tensor: The mesh vectors
input_dictdict: The input dict
frz_modelstr, optional: The path to the frozen model
ckpt_metastr, optional: The path prefix of the checkpoint and meta files
suffixstr, optional: The suffix of the scope
reusebool or tf.AUTO_REUSE, optional: Whether to reuse the variables

Returns:

dict: The output dict

get_fitting() → deepmd.tf.fit.fitting.Fitting | dict[source]#: Get the fitting(s).

get_loss(loss: dict, lr) → deepmd.tf.loss.loss.Loss | dict | None[source]#: Get the loss function(s).

get_rcut()[source]#: Get cutoff radius of the model.

get_ntypes() → int[source]#: Get the number of types.

data_stat(data) → None[source]#: Data staticis.

init_variables(graph: deepmd.tf.env.tf.Graph, graph_def: deepmd.tf.env.tf.GraphDef, model_type: str = 'original_model', suffix: str = '') → None[source]#

Init the embedding net variables with the given frozen model.

Parameters:

graphtf.Graph: The input frozen model graph
graph_deftf.GraphDef: The input frozen model graph_def
model_typestr: the type of the model
suffixstr: suffix to name scope

enable_compression(suffix: str = '') → None[source]#

Enable compression.

Parameters:

suffixstr: suffix to name scope

get_type_map() → list[source]#: Get the type map.

classmethod update_sel(train_data: deepmd.utils.data_system.DeepmdDataSystem, type_map: list[str] | None, local_jdata: dict) → tuple[dict, float | None][source]#

Update the selection and perform neighbor statistics.

Parameters:

train_dataDeepmdDataSystem: data used to do neighbor statistics
type_maplist[str], optional: The name of each type of atoms
local_jdatadict: The local data refer to the current class

Returns:

dict: The updated local data
float: The minimum distance between two atoms

serialize(suffix: str = '') → dict[source]#

Serialize the model.

There is no suffix in a native DP model, but it is important for the TF backend.

Returns:

dict: The serialized data
suffixstr, optional: Name suffix to identify this descriptor

classmethod deserialize(data: dict, suffix: str = '') → NoReturn[source]#

Deserialize the model.

There is no suffix in a native DP model, but it is important for the TF backend.

Parameters:

datadict: The serialized data
suffixstr, optional: Name suffix to identify this model

Returns:

Model: The deserialized Model

property input_requirement: list[deepmd.utils.data.DataRequirementItem][source]#: Return data requirements needed for the model input.