update sponge

This commit is contained in:
zhangxinfeng3 2021-03-03 15:03:01 +08:00
parent 885177769a
commit e1c277cdc5
1 changed files with 294 additions and 30 deletions

View File

@ -18,6 +18,7 @@
from ..primitive import PrimitiveWithInfer, prim_attr_register
from ..._checkparam import Validator as validator
from ...common import dtype as mstype
from ..._checkparam import Rel
class BondForce(PrimitiveWithInfer):
@ -50,12 +51,30 @@ class BondForce(PrimitiveWithInfer):
``GPU``
Examples:
"""
@prim_attr_register
def __init__(self, bond_numbers):
self.bond_numbers = bond_numbers
self.init_prim_io_names(inputs=['uint_crd_f', 'scaler_f', 'atom_a', 'atom_b', 'bond_k', 'bond_r0'],
outputs=['frc_f'])
self.add_prim_attr('bond_numbers', self.bond_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, bond_k_shape, bond_r0_shape):
cls_name = self.name
# N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
bond_k_shape[0], M, Rel.EQ, "bond_k_shape", cls_name)
validator.check_int(
bond_r0_shape[0], M, Rel.EQ, "bond_r0_shape", cls_name)
return uint_crd_f_shape
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, bond_k_type, bond_r0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
validator.check_tensor_dtype_valid('scaler_f_type', scaler_f_type, [mstype.float32], self.name)
@ -81,6 +100,11 @@ class BondEnergy(PrimitiveWithInfer):
Inputs:
Same as operator BondForce().
.. math::
dr = (x_1-x_2, y_1-y_2, z_1-z_2)
E = k*(|dr| - r_0)^2
Outputs:
- **bond_ene** (Tensor, float32) - [M, 1], the harmonic potential energy
for each bond.
@ -89,12 +113,31 @@ class BondEnergy(PrimitiveWithInfer):
``GPU``
Examples:
"""
@prim_attr_register
def __init__(self, bond_numbers):
self.bond_numbers = bond_numbers
self.init_prim_io_names(inputs=['uint_crd_f', 'scaler_f', 'atom_a', 'atom_b', 'bond_k', 'bond_r0'],
outputs=['bond_ene'])
self.add_prim_attr('bond_numbers', self.bond_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, bond_k_shape, bond_r0_shape):
cls_name = self.name
# N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
bond_k_shape[0], M, Rel.EQ, "bond_k_shape", cls_name)
validator.check_int(
bond_r0_shape[0], M, Rel.EQ, "bond_r0_shape", cls_name)
return bond_k_shape
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, bond_k_type, bond_r0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
validator.check_tensor_dtype_valid('scaler_f_type', scaler_f_type, [mstype.float32], self.name)
@ -125,12 +168,30 @@ class BondAtomEnergy(PrimitiveWithInfer):
``GPU``
Examples:
"""
@prim_attr_register
def __init__(self, bond_numbers):
self.bond_numbers = bond_numbers
self.init_prim_io_names(inputs=['uint_crd_f', 'scaler_f', 'atom_a', 'atom_b', 'bond_k', 'bond_r0'],
outputs=['atom_ene'])
self.add_prim_attr('bond_numbers', self.bond_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, bond_k_shape, bond_r0_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
bond_k_shape[0], M, Rel.EQ, "bond_k_shape", cls_name)
validator.check_int(
bond_r0_shape[0], M, Rel.EQ, "bond_r0_shape", cls_name)
return [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, bond_k_type, bond_r0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
validator.check_tensor_dtype_valid('scaler_f_type', scaler_f_type, [mstype.float32], self.name)
@ -167,13 +228,28 @@ class BondForceWithAtomEnergy(PrimitiveWithInfer):
self.init_prim_io_names(inputs=['uint_crd_f', 'scaler_f', 'atom_a', 'atom_b', 'bond_k', 'bond_r0'],
outputs=['frc_f', 'atom_e'])
self.add_prim_attr('bond_numbers', self.bond_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, bond_k_shape, bond_r0_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
bond_k_shape[0], M, Rel.EQ, "bond_k_shape", cls_name)
validator.check_int(
bond_r0_shape[0], M, Rel.EQ, "bond_r0_shape", cls_name)
return uint_crd_f_shape, [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, bond_k_type, bond_r0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
validator.check_tensor_dtype_valid('scaler_f_type', scaler_f_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('atom_a_type', atom_a_type, [mstype.int32], self.name)
validator.check_tensor_dtype_valid('atom_b_type', atom_b_type, [mstype.int32], self.name)
validator.check_tensor_dtype_valid('bond_k_type', bond_k_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('bond_r0_type', bond_r0_type, [mstype.float32], self.name)
return bond_r0_type, bond_r0_type
@ -213,17 +289,33 @@ class BondForceWithAtomVirial(PrimitiveWithInfer):
self.init_prim_io_names(inputs=['uint_crd_f', 'scaler_f', 'atom_a', 'atom_b', 'bond_k', 'bond_r0'],
outputs=['frc_f', 'atom_v'])
self.add_prim_attr('bond_numbers', self.bond_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, bond_k_shape, bond_r0_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
bond_k_shape[0], M, Rel.EQ, "bond_k_shape", cls_name)
validator.check_int(
bond_r0_shape[0], M, Rel.EQ, "bond_r0_shape", cls_name)
return uint_crd_f_shape, [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, bond_k_type, bond_r0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
validator.check_tensor_dtype_valid('scaler_f_type', scaler_f_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('atom_a_type', atom_a_type, [mstype.int32], self.name)
validator.check_tensor_dtype_valid('atom_b_type', atom_b_type, [mstype.int32], self.name)
validator.check_tensor_dtype_valid('bond_k_type', bond_k_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('bond_r0_type', bond_r0_type, [mstype.float32], self.name)
return bond_r0_type, bond_r0_type
class DihedralForce(PrimitiveWithInfer):
"""
DihedralForce:
@ -259,18 +351,18 @@ class DihedralForce(PrimitiveWithInfer):
Inputs:
- **uint_crd_f** (Tensor, uint32) - [N, 3], the unsigned int coordinates
value of each atom.
- **scalar_f** (Tensor, float32) - [3, 1], the 3-D scale factor between
- **scalar_f** (Tensor, float32) - [3, ], the 3-D scale factor between
the real space float coordinates and the unsigned int coordinates.
- **atom_a** (Tensor, int32) - [M, 1], the 1st atom index of each dihedral.
- **atom_b** (Tensor, int32) - [M, 1], the 2nd atom index of each dihedral.
- **atom_c** (Tensor, int32) - [M, 1], the 3rd atom index of each dihedral.
- **atom_d** (Tensor, int32) - [M, 1], the 4th atom index of each dihedral.
- **atom_a** (Tensor, int32) - [M, ], the 1st atom index of each dihedral.
- **atom_b** (Tensor, int32) - [M, ], the 2nd atom index of each dihedral.
- **atom_c** (Tensor, int32) - [M, ], the 3rd atom index of each dihedral.
- **atom_d** (Tensor, int32) - [M, ], the 4th atom index of each dihedral.
4 atoms are connected in the form a-b-c-d.
- **ipn** (Tensor, int32) - [M, 1], the period of dihedral angle of each dihedral.
- **pk** (Tensor, float32) - [M, 1], the force constant of each dihedral.
- **gamc** (Tensor, float32) - [M, 1], k*cos(phi_0) of each dihedral.
- **gams** (Tensor, float32) - [M, 1], k*sin(phi_0) of each dihedral.
- **pn** (Tensor, float32) - [M, 1], the floating point form of ipn.
- **ipn** (Tensor, int32) - [M, ], the period of dihedral angle of each dihedral.
- **pk** (Tensor, float32) - [M, ], the force constant of each dihedral.
- **gamc** (Tensor, float32) - [M, ], k*cos(phi_0) of each dihedral.
- **gams** (Tensor, float32) - [M, ], k*sin(phi_0) of each dihedral.
- **pn** (Tensor, float32) - [M, ], the floating point form of ipn.
Outputs:
- **frc_f** (Tensor, float32) - [N, 3], the force felt by each atom.
@ -289,6 +381,29 @@ class DihedralForce(PrimitiveWithInfer):
outputs=['frc_f'])
self.add_prim_attr('dihedral_numbers', self.dihedral_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, atom_d_shape,
ipn_shape, pk_shape, gamc_shape, gams_shape, pn_shape):
cls_name = self.name
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
atom_d_shape[0], M, Rel.EQ, "atom_d_shape", cls_name)
validator.check_int(ipn_shape[0], M, Rel.EQ, "ipn_shape", cls_name)
validator.check_int(pk_shape[0], M, Rel.EQ, "pk_shape", cls_name)
validator.check_int(gamc_shape[0], M, Rel.EQ, "gamc_shape", cls_name)
validator.check_int(gams_shape[0], M, Rel.EQ, "gams_shape", cls_name)
validator.check_int(pn_shape[0], M, Rel.EQ, "pn_shape", cls_name)
return uint_crd_f_shape
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, atom_d_type,
ipn_type, pk_type, gamc_type, gams_type, pn_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -302,7 +417,6 @@ class DihedralForce(PrimitiveWithInfer):
validator.check_tensor_dtype_valid('gamc_type', gamc_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('gams_type', gams_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('pn_type', pn_type, [mstype.float32], self.name)
return pn_type
@ -321,7 +435,7 @@ class DihedralEnergy(PrimitiveWithInfer):
Same as operator DihedralForce().
Outputs:
- **ene** (Tensor, float32) - [M, 1], the potential energy for each
- **ene** (Tensor, float32) - [M, ], the potential energy for each
dihedral term.
Supported Platforms:
@ -338,6 +452,29 @@ class DihedralEnergy(PrimitiveWithInfer):
outputs=['ene'])
self.add_prim_attr('dihedral_numbers', self.dihedral_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, atom_d_shape,
ipn_shape, pk_shape, gamc_shape, gams_shape, pn_shape):
cls_name = self.name
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
atom_d_shape[0], M, Rel.EQ, "atom_d_shape", cls_name)
validator.check_int(ipn_shape[0], M, Rel.EQ, "ipn_shape", cls_name)
validator.check_int(pk_shape[0], M, Rel.EQ, "pk_shape", cls_name)
validator.check_int(gamc_shape[0], M, Rel.EQ, "gamc_shape", cls_name)
validator.check_int(gams_shape[0], M, Rel.EQ, "gams_shape", cls_name)
validator.check_int(pn_shape[0], M, Rel.EQ, "pn_shape", cls_name)
return [M,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, atom_d_type,
ipn_type, pk_type, gamc_type, gams_type, pn_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -351,7 +488,6 @@ class DihedralEnergy(PrimitiveWithInfer):
validator.check_tensor_dtype_valid('gamc_type', gamc_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('gams_type', gams_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('pn_type', pn_type, [mstype.float32], self.name)
return pn_type
@ -368,7 +504,7 @@ class DihedralAtomEnergy(PrimitiveWithInfer):
Same as operator DihedralEnergy().
Outputs:
- **ene** (Tensor, float32) - [N, 1], the accumulated potential
- **ene** (Tensor, float32) - [N, ], the accumulated potential
energy for each atom.
Supported Platforms:
@ -385,6 +521,30 @@ class DihedralAtomEnergy(PrimitiveWithInfer):
outputs=['ene'])
self.add_prim_attr('dihedral_numbers', self.dihedral_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, atom_d_shape,
ipn_shape, pk_shape, gamc_shape, gams_shape, pn_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
atom_d_shape[0], M, Rel.EQ, "atom_d_shape", cls_name)
validator.check_int(ipn_shape[0], M, Rel.EQ, "ipn_shape", cls_name)
validator.check_int(pk_shape[0], M, Rel.EQ, "pk_shape", cls_name)
validator.check_int(gamc_shape[0], M, Rel.EQ, "gamc_shape", cls_name)
validator.check_int(gams_shape[0], M, Rel.EQ, "gams_shape", cls_name)
validator.check_int(pn_shape[0], M, Rel.EQ, "pn_shape", cls_name)
return [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, atom_d_type,
ipn_type, pk_type, gamc_type, gams_type, pn_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -398,7 +558,6 @@ class DihedralAtomEnergy(PrimitiveWithInfer):
validator.check_tensor_dtype_valid('gamc_type', gamc_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('gams_type', gams_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('pn_type', pn_type, [mstype.float32], self.name)
return pn_type
@ -415,7 +574,7 @@ class DihedralForceWithAtomEnergy(PrimitiveWithInfer):
Outputs:
- **frc_f** (Tensor, float32) - [N, 3], same as operator DihedralForce().
- **ene** (Tensor, float32) - [N, 1], same as operator DihedralAtomEnergy().
- **ene** (Tensor, float32) - [N, ], same as operator DihedralAtomEnergy().
Supported Platforms:
``GPU``
@ -431,6 +590,30 @@ class DihedralForceWithAtomEnergy(PrimitiveWithInfer):
outputs=['frc_f', 'ene'])
self.add_prim_attr('dihedral_numbers', self.dihedral_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, atom_d_shape,
ipn_shape, pk_shape, gamc_shape, gams_shape, pn_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
atom_d_shape[0], M, Rel.EQ, "atom_d_shape", cls_name)
validator.check_int(ipn_shape[0], M, Rel.EQ, "ipn_shape", cls_name)
validator.check_int(pk_shape[0], M, Rel.EQ, "pk_shape", cls_name)
validator.check_int(gamc_shape[0], M, Rel.EQ, "gamc_shape", cls_name)
validator.check_int(gams_shape[0], M, Rel.EQ, "gams_shape", cls_name)
validator.check_int(pn_shape[0], M, Rel.EQ, "pn_shape", cls_name)
return uint_crd_f_shape, [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, atom_d_type,
ipn_type, pk_type, gamc_type, gams_type, pn_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -444,7 +627,6 @@ class DihedralForceWithAtomEnergy(PrimitiveWithInfer):
validator.check_tensor_dtype_valid('gamc_type', gamc_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('gams_type', gams_type, [mstype.float32], self.name)
validator.check_tensor_dtype_valid('pn_type', pn_type, [mstype.float32], self.name)
return pn_type, pn_type
@ -470,14 +652,14 @@ class AngleForce(PrimitiveWithInfer):
Inputs:
- **uint_crd_f** (Tensor, uint32) - [N, 3], the unsigned int coordinate
value of each atom.
- **scaler_f** (Tensor, float32) - [3, 1], the 3-D scale factor between
- **scaler_f** (Tensor, float32) - [3, ], the 3-D scale factor between
the real space float coordinates and the unsigned int coordinates.
- **atom_a** (Tensor, int32) - [M, 1], the 1st atom index of each angle.
- **atom_b** (Tensor, int32) - [M, 1], the 2nd and the central atom index
- **atom_a** (Tensor, int32) - [M, ], the 1st atom index of each angle.
- **atom_b** (Tensor, int32) - [M, ], the 2nd and the central atom index
of each angle.
- **atom_c** (Tensor, int32) - [M, 1], the 3rd atom index of each angle.
- **angle_k** (Tensor, float32) - [M, 1], the force constant for each angle.
- **angle_theta0** (Tensor, float32) - [M, 1], the equilibrium position value
- **atom_c** (Tensor, int32) - [M, ], the 3rd atom index of each angle.
- **angle_k** (Tensor, float32) - [M, ], the force constant for each angle.
- **angle_theta0** (Tensor, float32) - [M, ], the equilibrium position value
for each angle.
Outputs:
@ -497,6 +679,26 @@ class AngleForce(PrimitiveWithInfer):
outputs=['frc_f'])
self.add_prim_attr('angle_numbers', self.angle_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, angle_k_shape,
angle_theta0_shape):
cls_name = self.name
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
angle_k_shape[0], M, Rel.EQ, "angle_k_shape", cls_name)
validator.check_int(
angle_theta0_shape[0], M, Rel.EQ, "angle_theta0_shape", cls_name)
return uint_crd_f_shape
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, angle_k_type,
angle_theta0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -526,7 +728,7 @@ class AngleEnergy(PrimitiveWithInfer):
Same as operator AngleForce().
Outputs:
- **ene** (Tensor, float32) - [M, 1], the potential energy for
- **ene** (Tensor, float32) - [M, ], the potential energy for
each angle term.
Supported Platforms:
@ -543,6 +745,26 @@ class AngleEnergy(PrimitiveWithInfer):
outputs=['ene'])
self.add_prim_attr('angle_numbers', self.angle_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, angle_k_shape,
angle_theta0_shape):
cls_name = self.name
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
angle_k_shape[0], M, Rel.EQ, "angle_k_shape", cls_name)
validator.check_int(
angle_theta0_shape[0], M, Rel.EQ, "angle_theta0_shape", cls_name)
return [M,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, angle_k_type,
angle_theta0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -568,7 +790,7 @@ class AngleAtomEnergy(PrimitiveWithInfer):
Same as operator AngleForce().
Outputs:
- **ene** (Tensor, float32) - [N, 1], the accumulated potential energy
- **ene** (Tensor, float32) - [N, ], the accumulated potential energy
for each atom.
Supported Platforms:
@ -585,6 +807,27 @@ class AngleAtomEnergy(PrimitiveWithInfer):
outputs=['ene'])
self.add_prim_attr('angle_numbers', self.angle_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, angle_k_shape,
angle_theta0_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
angle_k_shape[0], M, Rel.EQ, "angle_k_shape", cls_name)
validator.check_int(
angle_theta0_shape[0], M, Rel.EQ, "angle_theta0_shape", cls_name)
return [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, angle_k_type,
angle_theta0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)
@ -610,7 +853,7 @@ class AngleForceWithAtomEnergy(PrimitiveWithInfer):
Outputs:
- **frc_f** (Tensor, float32) - [N, 3], same as operator AngleForce().
- **ene** (Tensor, float) - [N, 1], same as operator AngleAtomEnergy().
- **ene** (Tensor, float) - [N, ], same as operator AngleAtomEnergy().
Supported Platforms:
``GPU``
@ -626,6 +869,27 @@ class AngleForceWithAtomEnergy(PrimitiveWithInfer):
outputs=['frc_f', 'ene'])
self.add_prim_attr('angle_numbers', self.angle_numbers)
def infer_shape(self, uint_crd_f_shape, scaler_f_shape, atom_a_shape, atom_b_shape, atom_c_shape, angle_k_shape,
angle_theta0_shape):
cls_name = self.name
N = uint_crd_f_shape[0]
M = atom_a_shape[0]
validator.check_int(
uint_crd_f_shape[1], 3, Rel.EQ, "uint_crd_f_shape", cls_name)
validator.check_int(
scaler_f_shape[0], 3, Rel.EQ, "scaler_f_shape", cls_name)
validator.check_int(
atom_a_shape[0], M, Rel.EQ, "atom_a_shape", cls_name)
validator.check_int(
atom_b_shape[0], M, Rel.EQ, "atom_b_shape", cls_name)
validator.check_int(
atom_c_shape[0], M, Rel.EQ, "atom_c_shape", cls_name)
validator.check_int(
angle_k_shape[0], M, Rel.EQ, "angle_k_shape", cls_name)
validator.check_int(
angle_theta0_shape[0], M, Rel.EQ, "angle_theta0_shape", cls_name)
return uint_crd_f_shape, [N,]
def infer_dtype(self, uint_crd_f_dtype, scaler_f_type, atom_a_type, atom_b_type, atom_c_type, angle_k_type,
angle_theta0_type):
validator.check_tensor_dtype_valid('uint_crd_f_dtype', uint_crd_f_dtype, [mstype.uint32], self.name)