Merge remote-tracking branch 'upstream/master' into feat-pt-expt-pt2-uts

Author: Han Wang

deepmd/dpmodel/loss/__init__.py

@@ -1 +1,24 @@
 # SPDX-License-Identifier: LGPL-3.0-or-later
+from deepmd.dpmodel.loss.dos import (
+    DOSLoss,
+)
+from deepmd.dpmodel.loss.ener import (
+    EnergyLoss,
+)
+from deepmd.dpmodel.loss.ener_spin import (
+    EnergySpinLoss,
+)
+from deepmd.dpmodel.loss.property import (
+    PropertyLoss,
+)
+from deepmd.dpmodel.loss.tensor import (
+    TensorLoss,
+)
+
+__all__ = [
+    "DOSLoss",
+    "EnergyLoss",
+    "EnergySpinLoss",
+    "PropertyLoss",
+    "TensorLoss",
+]

deepmd/dpmodel/loss/dos.py

@@ -0,0 +1,268 @@
+# SPDX-License-Identifier: LGPL-3.0-or-later
+from typing import (
+    Any,
+)
+
+import array_api_compat
+
+from deepmd.dpmodel.array_api import (
+    Array,
+)
+from deepmd.dpmodel.loss.loss import (
+    Loss,
+)
+from deepmd.utils.data import (
+    DataRequirementItem,
+)
+from deepmd.utils.version import (
+    check_version_compatibility,
+)
+
+
+class DOSLoss(Loss):
+    r"""Loss on DOS (density of states) for both local and global predictions.
+
+    Parameters
+    ----------
+    starter_learning_rate : float
+        The learning rate at the start of the training.
+    numb_dos : int
+        The number of DOS components.
+    start_pref_dos : float
+        The prefactor of global DOS loss at the start of the training.
+    limit_pref_dos : float
+        The prefactor of global DOS loss at the end of the training.
+    start_pref_cdf : float
+        The prefactor of global CDF loss at the start of the training.
+    limit_pref_cdf : float
+        The prefactor of global CDF loss at the end of the training.
+    start_pref_ados : float
+        The prefactor of atomic DOS loss at the start of the training.
+    limit_pref_ados : float
+        The prefactor of atomic DOS loss at the end of the training.
+    start_pref_acdf : float
+        The prefactor of atomic CDF loss at the start of the training.
+    limit_pref_acdf : float
+        The prefactor of atomic CDF loss at the end of the training.
+    **kwargs
+        Other keyword arguments.
+    """
+
+    def __init__(
+        self,
+        starter_learning_rate: float,
+        numb_dos: int,
+        start_pref_dos: float = 1.00,
+        limit_pref_dos: float = 1.00,
+        start_pref_cdf: float = 1000,
+        limit_pref_cdf: float = 1.00,
+        start_pref_ados: float = 0.0,
+        limit_pref_ados: float = 0.0,
+        start_pref_acdf: float = 0.0,
+        limit_pref_acdf: float = 0.0,
+        **kwargs: Any,
+    ) -> None:
+        self.starter_learning_rate = starter_learning_rate
+        self.numb_dos = numb_dos
+        self.start_pref_dos = start_pref_dos
+        self.limit_pref_dos = limit_pref_dos
+        self.start_pref_cdf = start_pref_cdf
+        self.limit_pref_cdf = limit_pref_cdf
+        self.start_pref_ados = start_pref_ados
+        self.limit_pref_ados = limit_pref_ados
+        self.start_pref_acdf = start_pref_acdf
+        self.limit_pref_acdf = limit_pref_acdf
+
+        assert (
+            self.start_pref_dos >= 0.0
+            and self.limit_pref_dos >= 0.0
+            and self.start_pref_cdf >= 0.0
+            and self.limit_pref_cdf >= 0.0
+            and self.start_pref_ados >= 0.0
+            and self.limit_pref_ados >= 0.0
+            and self.start_pref_acdf >= 0.0
+            and self.limit_pref_acdf >= 0.0
+        ), "Can not assign negative weight to `pref` and `pref_atomic`"
+
+        self.has_dos = start_pref_dos != 0.0 or limit_pref_dos != 0.0
+        self.has_cdf = start_pref_cdf != 0.0 or limit_pref_cdf != 0.0
+        self.has_ados = start_pref_ados != 0.0 or limit_pref_ados != 0.0
+        self.has_acdf = start_pref_acdf != 0.0 or limit_pref_acdf != 0.0
+
+        assert self.has_dos or self.has_cdf or self.has_ados or self.has_acdf, (
+            "Can not assign zero weight to all pref terms"
+        )
+
+    def call(
+        self,
+        learning_rate: float,
+        natoms: int,
+        model_dict: dict[str, Array],
+        label_dict: dict[str, Array],
+        mae: bool = False,
+    ) -> tuple[Array, dict[str, Array]]:
+        """Calculate loss from model results and labeled results."""
+        # Get array namespace from any available tensor
+        first_key = next(iter(model_dict))
+        xp = array_api_compat.array_namespace(model_dict[first_key])
+
+        coef = learning_rate / self.starter_learning_rate
+        pref_dos = (
+            self.limit_pref_dos + (self.start_pref_dos - self.limit_pref_dos) * coef
+        )
+        pref_cdf = (
+            self.limit_pref_cdf + (self.start_pref_cdf - self.limit_pref_cdf) * coef
+        )
+        pref_ados = (
+            self.limit_pref_ados + (self.start_pref_ados - self.limit_pref_ados) * coef
+        )
+        pref_acdf = (
+            self.limit_pref_acdf + (self.start_pref_acdf - self.limit_pref_acdf) * coef
+        )
+
+        loss = 0
+        more_loss = {}
+
+        if self.has_ados and "atom_dos" in model_dict and "atom_dos" in label_dict:
+            find_local = label_dict.get("find_atom_dos", 0.0)
+            pref_ados = pref_ados * find_local
+            local_pred = xp.reshape(model_dict["atom_dos"], (-1, natoms, self.numb_dos))
+            local_label = xp.reshape(
+                label_dict["atom_dos"], (-1, natoms, self.numb_dos)
+            )
+            diff = xp.reshape(local_pred - local_label, (-1, self.numb_dos))
+            if "mask" in model_dict:
+                mask = xp.reshape(model_dict["mask"], (-1,))
+                mask_float = xp.astype(mask, diff.dtype)
+                diff = diff * mask_float[:, None]
+                n_valid = xp.sum(mask_float)
+                l2_local_loss_dos = xp.sum(xp.square(diff)) / (n_valid * self.numb_dos)
+            else:
+                l2_local_loss_dos = xp.mean(xp.square(diff))
+            loss += pref_ados * l2_local_loss_dos
+            more_loss["rmse_local_dos"] = self.display_if_exist(
+                xp.sqrt(l2_local_loss_dos), find_local
+            )
+
+        if self.has_acdf and "atom_dos" in model_dict and "atom_dos" in label_dict:
+            find_local = label_dict.get("find_atom_dos", 0.0)
+            pref_acdf = pref_acdf * find_local
+            local_pred_cdf = xp.cumulative_sum(
+                xp.reshape(model_dict["atom_dos"], (-1, natoms, self.numb_dos)),
+                axis=-1,
+            )
+            local_label_cdf = xp.cumulative_sum(
+                xp.reshape(label_dict["atom_dos"], (-1, natoms, self.numb_dos)),
+                axis=-1,
+            )
+            diff = xp.reshape(local_pred_cdf - local_label_cdf, (-1, self.numb_dos))
+            if "mask" in model_dict:
+                mask = xp.reshape(model_dict["mask"], (-1,))
+                mask_float = xp.astype(mask, diff.dtype)
+                diff = diff * mask_float[:, None]
+                n_valid = xp.sum(mask_float)
+                l2_local_loss_cdf = xp.sum(xp.square(diff)) / (n_valid * self.numb_dos)
+            else:
+                l2_local_loss_cdf = xp.mean(xp.square(diff))
+            loss += pref_acdf * l2_local_loss_cdf
+            more_loss["rmse_local_cdf"] = self.display_if_exist(
+                xp.sqrt(l2_local_loss_cdf), find_local
+            )
+
+        if self.has_dos and "dos" in model_dict and "dos" in label_dict:
+            find_global = label_dict.get("find_dos", 0.0)
+            pref_dos = pref_dos * find_global
+            global_pred = xp.reshape(model_dict["dos"], (-1, self.numb_dos))
+            global_label = xp.reshape(label_dict["dos"], (-1, self.numb_dos))
+            diff = global_pred - global_label
+            if "mask" in model_dict:
+                atom_num = xp.sum(model_dict["mask"], axis=-1, keepdims=True)
+                l2_global_loss_dos = xp.mean(
+                    xp.sum(xp.square(diff) * atom_num, axis=0) / xp.sum(atom_num)
+                )
+                atom_num = xp.mean(xp.astype(atom_num, diff.dtype))
+            else:
+                atom_num = natoms
+                l2_global_loss_dos = xp.mean(xp.square(diff))
+            loss += pref_dos * l2_global_loss_dos
+            more_loss["rmse_global_dos"] = self.display_if_exist(
+                xp.sqrt(l2_global_loss_dos) / atom_num, find_global
+            )
+
+        if self.has_cdf and "dos" in model_dict and "dos" in label_dict:
+            find_global = label_dict.get("find_dos", 0.0)
+            pref_cdf = pref_cdf * find_global
+            global_pred_cdf = xp.cumulative_sum(
+                xp.reshape(model_dict["dos"], (-1, self.numb_dos)), axis=-1
+            )
+            global_label_cdf = xp.cumulative_sum(
+                xp.reshape(label_dict["dos"], (-1, self.numb_dos)), axis=-1
+            )
+            diff = global_pred_cdf - global_label_cdf
+            if "mask" in model_dict:
+                atom_num = xp.sum(model_dict["mask"], axis=-1, keepdims=True)
+                l2_global_loss_cdf = xp.mean(
+                    xp.sum(xp.square(diff) * atom_num, axis=0) / xp.sum(atom_num)
+                )
+                atom_num = xp.mean(xp.astype(atom_num, diff.dtype))
+            else:
+                atom_num = natoms
+                l2_global_loss_cdf = xp.mean(xp.square(diff))
+            loss += pref_cdf * l2_global_loss_cdf
+            more_loss["rmse_global_cdf"] = self.display_if_exist(
+                xp.sqrt(l2_global_loss_cdf) / atom_num, find_global
+            )
+
+        more_loss["rmse"] = xp.sqrt(loss)
+        return loss, more_loss
+
+    @property
+    def label_requirement(self) -> list[DataRequirementItem]:
+        """Return data label requirements needed for this loss calculation."""
+        label_requirement = []
+        if self.has_ados or self.has_acdf:
+            label_requirement.append(
+                DataRequirementItem(
+                    "atom_dos",
+                    ndof=self.numb_dos,
+                    atomic=True,
+                    must=False,
+                    high_prec=False,
+                )
+            )
+        if self.has_dos or self.has_cdf:
+            label_requirement.append(
+                DataRequirementItem(
+                    "dos",
+                    ndof=self.numb_dos,
+                    atomic=False,
+                    must=False,
+                    high_prec=False,
+                )
+            )
+        return label_requirement
+
+    def serialize(self) -> dict:
+        """Serialize the loss module."""
+        return {
+            "@class": "DOSLoss",
+            "@version": 1,
+            "starter_learning_rate": self.starter_learning_rate,
+            "numb_dos": self.numb_dos,
+            "start_pref_dos": self.start_pref_dos,
+            "limit_pref_dos": self.limit_pref_dos,
+            "start_pref_cdf": self.start_pref_cdf,
+            "limit_pref_cdf": self.limit_pref_cdf,
+            "start_pref_ados": self.start_pref_ados,
+            "limit_pref_ados": self.limit_pref_ados,
+            "start_pref_acdf": self.start_pref_acdf,
+            "limit_pref_acdf": self.limit_pref_acdf,
+        }
+
+    @classmethod
+    def deserialize(cls, data: dict) -> "DOSLoss":
+        """Deserialize the loss module."""
+        data = data.copy()
+        check_version_compatibility(data.pop("@version"), 1, 1)
+        data.pop("@class")
+        return cls(**data)

deepmd/dpmodel/loss/ener.py

@@ -166,6 +166,7 @@ def call(
         natoms: int,
         model_dict: dict[str, Array],
         label_dict: dict[str, Array],
+        mae: bool = False,
     ) -> tuple[Array, dict[str, Array]]:
         """Calculate loss from model results and labeled results."""
         energy = model_dict["energy"]
@@ -266,6 +267,11 @@ def call(
                 raise NotImplementedError(
                     f"Loss type {self.loss_func} is not implemented for energy loss."
                 )
+            if mae:
+                mae_e = xp.mean(xp.abs(energy - energy_hat)) * atom_norm_ener
+                more_loss["mae_e"] = self.display_if_exist(mae_e, find_energy)
+                mae_e_all = xp.mean(xp.abs(energy - energy_hat))
+                more_loss["mae_e_all"] = self.display_if_exist(mae_e_all, find_energy)
         if self.has_f:
             if self.loss_func == "mse":
                 l2_force_loss = xp.mean(xp.square(diff_f))
@@ -304,6 +310,9 @@ def call(
                 raise NotImplementedError(
                     f"Loss type {self.loss_func} is not implemented for force loss."
                 )
+            if mae:
+                mae_f = xp.mean(xp.abs(diff_f))
+                more_loss["mae_f"] = self.display_if_exist(mae_f, find_force)
         if self.has_v:
             virial_reshape = xp.reshape(virial, (-1,))
             virial_hat_reshape = xp.reshape(virial_hat, (-1,))
@@ -333,6 +342,9 @@ def call(
                 raise NotImplementedError(
                     f"Loss type {self.loss_func} is not implemented for virial loss."
                 )
+            if mae:
+                mae_v = xp.mean(xp.abs(virial_hat_reshape - virial_reshape)) * atom_norm
+                more_loss["mae_v"] = self.display_if_exist(mae_v, find_virial)
         if self.has_ae:
             atom_ener_reshape = xp.reshape(atom_ener, (-1,))
             atom_ener_hat_reshape = xp.reshape(atom_ener_hat, (-1,))