2021_11_30

readme.md

@@ -14,5 +14,4 @@ ___________________________________
 **References:**
 1. The original EANN model: Yaolong Zhang, Ce Hu and Bin Jiang *J. Phys. Chem. Lett.* 10, 4962-4967 (2019).
 2. The EANN model for dipole/transition dipole/polarizability: Yaolong Zhang  Sheng Ye, Jinxiao Zhang, Jun Jiang and Bin Jiang *J. Phys. Chem. B*  124, 7284–7290 (2020).
-3. The REANN model: Yaolong Zhang, Junfan xia and Bin Jiang *arXiv:2106.08245*
-
+3. The REANN model: Yaolong Zhang, Junfan xia and Bin Jiang *Phys. Rev. Lett.* 127, 156002 (2021).

reann/dm/PES.py

@@ -13,18 +13,18 @@ def __init__(self,nlinked=1):
         global oc_loop,oc_nblock, oc_nl, oc_dropout_p, oc_table_norm, oc_activate
         global nwave, neigh_atoms, cutoff, nipsin, atomtype
         # global parameters for input_nn
-        nblock = 2                   # nblock>=2  resduial NN block will be employed nblock=1: simple feedforward nn
+        nblock = 1                   # nblock>=2  resduial NN block will be employed nblock=1: simple feedforward nn
         nl=[128,128]                # NN structure
         dropout_p=[0.0,0.0,0.0]       # dropout probability for each hidden layer
         activate = 'Relu_like'
-        table_norm= False
-        oc_loop = 0
-        oc_nl = [32,32]          # neural network architecture   
+        table_norm= True
+        oc_loop = 1
+        oc_nl = [128,128]          # neural network architecture   
         oc_nblock = 1
         oc_dropout_p=[0.0,0.0,0.0,0.0]
         oc_activate = 'Relu_like'
         #========================queue_size sequence for laod data into gpu
-        oc_table_norm = False
+        oc_table_norm = True
         norbit = None
         #======================read input_nn==================================
         with open('para/input_nn','r') as f1:
@@ -42,8 +42,9 @@ def __init__(self,nlinked=1):
         outputneuron=1
         #======================read input_nn=============================================
         nipsin=2
-        cutoff=4.0
-        nwave=6
+        cutoff=4.5
+        nwave=7
+        neigh_atoms=150
         with open('para/input_density','r') as f1:
            while True:
               tmp=f1.readline()

reann/dm/script_PES.py

@@ -3,7 +3,7 @@
 import torch
 def jit_pes():
     init_pes=PES.PES()
-    state_dict = torch.load("EANN.pth",map_location='cpu')
+    state_dict = torch.load("REANN.pth",map_location='cpu')
     new_state_dict = OrderedDict()
     for k, v in state_dict['eannparam'].items():
         if k[0:7]=="module.":
@@ -17,6 +17,6 @@ def jit_pes():
     for params in scripted_pes.parameters():
         params.requires_grad=False
     scripted_pes.to(torch.double)
-    scripted_pes.save("EANN_DM_DOUBLE.pt")
+    scripted_pes.save("REANN_DM_DOUBLE.pt")
     scripted_pes.to(torch.float32)
-    scripted_pes.save("EANN_DM_FLOAT.pt")
+    scripted_pes.save("REANN_DM_FLOAT.pt")

reann/lammps/PES.py

@@ -14,16 +14,16 @@ def __init__(self,nlinked=1):
         # global parameters for input_nn
         nblock = 1                    # nblock>=2  resduial NN block will be employed nblock=1: simple feedforward nn
         nl=[128,128]                # NN structure
-        dropout_p=[0.0,0.0,0.0]       # dropout probability for each hidden layer
+        dropout_p=[0.0,0.0]       # dropout probability for each hidden layer
         activate = "Relu_like"
-        table_norm= False
-        oc_loop = 0
-        oc_nl = [32,32]          # neural network architecture   
+        table_norm= True
+        oc_loop = 1
+        oc_nl = [128,128]          # neural network architecture   
         oc_nblock = 1
-        oc_dropout_p=[0.0,0.0,0.0,0.0]
+        oc_dropout_p=[0.0,0.0]
         oc_activate = "Relu_like"
         #========================queue_size sequence for laod data into gpu
-        oc_table_norm=False
+        oc_table_norm=True
         norbit=None
         #======================read input_nn==================================
         with open('para/input_nn','r') as f1:
@@ -41,8 +41,8 @@ def __init__(self,nlinked=1):
         outputneuron=1
         #======================read input_nn=============================================
         nipsin=2
-        cutoff=4.0
-        nwave=6
+        cutoff=4.5
+        nwave=7
         with open('para/input_density','r') as f1:
            while True:
               tmp=f1.readline()

reann/lammps_REANN/PES.py

@@ -14,16 +14,16 @@ def __init__(self,nlinked=1):
         # global parameters for input_nn
         nblock = 1                    # nblock>=2  resduial NN block will be employed nblock=1: simple feedforward nn
         nl=[128,128]                # NN structure
-        dropout_p=[0.0,0.0,0.0]       # dropout probability for each hidden layer
+        dropout_p=[0.0,0.0]       # dropout probability for each hidden layer
         activate="Relu_like"
-        table_norm= False
+        table_norm= True
         oc_loop = 0
-        oc_nl = [32,32]          # neural network architecture   
+        oc_nl = [128,128]          # neural network architecture   
         oc_nblock = 1
-        oc_dropout_p=[0.0,0.0,0.0,0.0]
+        oc_dropout_p=[0.0,0.0]
         oc_activate="Relu_like"
         #========================queue_size sequence for laod data into gpu
-        oc_table_norm=False
+        oc_table_norm= True
         norbit=None
         #======================read input_nn==================================
         with open('para/input_nn','r') as f1:
@@ -41,8 +41,8 @@ def __init__(self,nlinked=1):
         outputneuron=1
         #======================read input_nn=============================================
         nipsin=2
-        cutoff=4.0
-        nwave=6
+        cutoff=4.5
+        nwave=7
         with open('para/input_density','r') as f1:
            while True:
               tmp=f1.readline()

reann/pes/PES.py

@@ -17,14 +17,14 @@ def __init__(self,nlinked=1):
         nl=[128,128]                # NN structure
         dropout_p=[0.0,0.0,0.0]       # dropout probability for each hidden layer
         activate = 'Relu_like'
-        table_norm= False
-        oc_loop = 0
-        oc_nl = [32,32]          # neural network architecture   
+        table_norm= True
+        oc_loop = 1
+        oc_nl = [128,128]          # neural network architecture   
         oc_nblock = 1
         oc_dropout_p=[0.0,0.0,0.0,0.0]
         oc_activate = 'Relu_like'
         #========================queue_size sequence for laod data into gpu
-        oc_table_norm=False
+        oc_table_norm=True
         norbit= None
         #======================read input_nn==================================
         with open('para/input_nn','r') as f1:
@@ -42,8 +42,9 @@ def __init__(self,nlinked=1):
         outputneuron=1
         #======================read input_nn=============================================
         nipsin=2
-        cutoff=4.0
-        nwave=6
+        cutoff=4.5
+        nwave=7
+        neigh_atoms=150
         with open('para/input_density','r') as f1:
            while True:
               tmp=f1.readline()

reann/pol/PES.py

@@ -17,14 +17,14 @@ def __init__(self,nlinked=1):
         nl=[128,128]                # NN structure
         dropout_p=[0.0,0.0,0.0]       # dropout probability for each hidden layer
         activate = 'Relu_like'
-        table_norm= False
-        oc_loop = 0
-        oc_nl = [32,32]          # neural network architecture   
+        table_norm= True
+        oc_loop = 1
+        oc_nl = [128,128]          # neural network architecture   
         oc_nblock = 1
         oc_dropout_p=[0.0,0.0,0.0,0.0]
         oc_activate = 'Relu_like'
         #========================queue_size sequence for laod data into gpu
-        oc_table_norm=False
+        oc_table_norm=True
         norbit=None
         #======================read input_nn==================================
         with open('para/input_nn','r') as f1:
@@ -42,8 +42,9 @@ def __init__(self,nlinked=1):
         outputneuron=1
         #======================read input_nn=============================================
         nipsin=2
-        cutoff=4.0
-        nwave=6
+        cutoff=4.5
+        nwave=7
+        neigh_atoms=150
         with open('para/input_density','r') as f1:
            while True:
               tmp=f1.readline()

reann/run/train.py

@@ -6,6 +6,7 @@
 from src.density import *
 from src.MODEL import *
 from src.EMA import *
+from torch.nn.parallel import DistributedDataParallel as DDP
 if activate=='Tanh_like':
     from src.activate import Tanh_like as actfun
 else:

reann/src/read.py

@@ -7,7 +7,6 @@
 from src.gpu_sel import *
 # used for DDP
 import torch.distributed as dist
-from torch.nn.parallel import DistributedDataParallel as DDP
 
 # open a file for output information in iterations
 fout=open('nn.err','w')

reann/tdm/PES.py

@@ -17,14 +17,14 @@ def __init__(self,nlinked=1):
         nl=[128,128]                # NN structure
         dropout_p=[0.0,0.0,0.0]       # dropout probability for each hidden layer
         activate = 'Relu_like'
-        table_norm = False
-        oc_loop = 0
+        table_norm = True
+        oc_loop = 1
         oc_nl = [32,32]          # neural network architecture   
         oc_nblock = 1
         oc_dropout_p=[0.0,0.0,0.0,0.0]
         oc_activate = 'Relu_like'
         #========================queue_size sequence for laod data into gpu
-        oc_table_norm=False
+        oc_table_norm=True
         norbit=None
         #======================read input_nn==================================
         with open('para/input_nn','r') as f1:
@@ -42,8 +42,9 @@ def __init__(self,nlinked=1):
         outputneuron=3
         #======================read input_nn=============================================
         nipsin=2
-        cutoff=4.0
-        nwave=6
+        cutoff=4.5
+        nwave=7
+        neigh_atoms=150
         with open('para/input_density','r') as f1:
            while True:
               tmp=f1.readline()