nvnmd

deepmd/nvnmd/data/data.py

@@ -192,6 +192,7 @@
     },
     "ctrl": {
         # NSTDM
+        "MAX_NNEI": 128,
         "NSTDM": 64,
         "NSTDM_M1": 32,
         "NSTDM_M2": 2,

deepmd/nvnmd/entrypoints/mapt.py

@@ -87,9 +87,22 @@ def __init__(self, config_file: str, weight_file: str, map_file: str):
         jdata["weight_file"] = weight_file
         jdata["enable"] = True
 
+        # 0 : xyz_scatter = xyz_scatter * two_embd + xyz_scatter;
+        # Gs + 1, Gt + 0
+        # 1 : xyz_scatter = xyz_scatter * two_embd + two_embd   ;
+        # Gs + 0, Gt + 1
+        self.Gs_Gt_mode = 1
+
         nvnmd_cfg.init_from_jdata(jdata)
 
     def build_map(self):
+        if self.Gs_Gt_mode == 0:
+            self.shift_Gs = 1
+            self.shift_Gt = 0
+        if self.Gs_Gt_mode == 1:
+            self.shift_Gs = 0
+            self.shift_Gt = 1
+        #
         M = nvnmd_cfg.dscp["M1"]
         if nvnmd_cfg.version == 0:
             ndim = nvnmd_cfg.dscp["ntype"]
@@ -482,7 +495,7 @@ def build_s2g_grad(self):
             shift = 0
         if nvnmd_cfg.version == 1:
             ndim = 1
-            shift = 0
+            shift = self.shift_Gs
         #
         dic_ph = {}
         dic_ph["s"] = tf.placeholder(tf.float64, [None, 1], "t_s")
@@ -575,10 +588,8 @@ def build_t2g(self):
             [-1, two_side_type_embedding.shape[-1]],
         )
         # see se_atten.py in dp
-        # old version : xyz_scatter = xyz_scatter * two_embd + xyz_scatter; Gs + 1, Gt + 0
-        # new version : xyz_scatter = xyz_scatter * two_embd + two_embd   ; Gs + 0, Gt + 1
         wbs = [get_filter_type_weight(nvnmd_cfg.weight, ll) for ll in range(1, 5)]
-        dic_ph["gt"] = self.build_embedding_net(two_side_type_embedding, wbs) + 1
+        dic_ph["gt"] = self.build_embedding_net(two_side_type_embedding, wbs) + self.shift_Gt
         return dic_ph
 
     def run_t2g(self):

deepmd/nvnmd/entrypoints/wrap.py

@@ -205,10 +205,10 @@ def wrap_head(self, nhs, nws):
             bs = e.dec2bin(SEL, NBIT_MODEL_HEAD)[0] + bs
         # atom_ener
         # fix the bug: the different energy between qnn and lammps
-        if "t_bias_atom_e" not in weight.keys():
-            log.error("NVNMD: There is not t_bias_atom_e in weight!")
-            exit(1)
-        atom_ener = weight['t_bias_atom_e']
+        if "t_bias_atom_e" in weight.keys():
+            atom_ener = weight['t_bias_atom_e']
+        else:
+            atom_ener = [0] * 32
         nlayer_fit = fitn["nlayer_fit"]
         if VERSION == 0:
             for tt in range(ntype):

deepmd/nvnmd/utils/config.py

@@ -113,6 +113,8 @@ def init_from_config(self, jdata):
         r"""Initialize member element one by one."""
         if "ctrl" in jdata.keys():
             if "VERSION" in jdata["ctrl"].keys():
+                if "MAX_NNEI" not in jdata["ctrl"].keys():
+                    jdata["ctrl"]["MAX_NNEI"] = 128
                 self.init_config_by_version(jdata["ctrl"]["VERSION"], jdata["ctrl"]["MAX_NNEI"])
         #
         self.config = FioDic().update(jdata, self.config)
@@ -336,6 +338,7 @@ def get_nvnmd_jdata(self):
         r"""Generate `nvnmd` in input script."""
         jdata = self.jdata_deepmd_input["nvnmd"]
         jdata["net_size"] = self.net_size
+        jdata["max_nnei"] = self.max_nnei
         jdata["config_file"] = self.config_file
         jdata["weight_file"] = self.weight_file
         jdata["map_file"] = self.map_file

doc/nvnmd/nvnmd.md

@@ -60,6 +60,7 @@ The "nvnmd" section is defined as
 ```json
 {
     "version": 0,
+    "max_nnei":128,
     "net_size":128,
     "sel":[60, 60],
     "rcut":6.0,
@@ -73,6 +74,7 @@ where items are defined as:
 | Item      | Mean                        | Optional Value                                |
 | --------- | --------------------------- | --------------------------------------------- |
 | version | the version of network structure | 0 or 1 |
+| max_nnei  | the maximum number of neighbors that do not distinguish element types | 128  or 256 |
 | net_size  | the size of nueral network  | 128                                     |
 | sel       | the number of neighbors     | version 0: integer list of lengths 1 to 4 are acceptable; version 1: integer |
 | rcut      | the cutoff radial           | (0, 8.0]                                      |
@@ -187,6 +189,15 @@ You can also restart the CNN training from the path prefix of checkpoint files (
 dp train-nvnmd train_cnn.json -r nvnmd_cnn/model.ckpt -s s1
 ```
 
+You can also initialize the CNN model and train it by
+
+``` bash
+mv nvnmd_cnn nvnmd_cnn_bck
+cp train_cnn.json train_cnn2.json
+# please edit train_cnn2.json
+dp train-nvnmd train_cnn2.json -s s1 -i nvnmd_cnn_bck/model.ckpt
+```
+
 
 # Testing
 

examples/nvnmd/train/train_cnn.json

@@ -1,6 +1,7 @@
 {
   "nvnmd": {
     "version": 0,
+    "max_nnei": 128,
     "net_size": 128,
     "sel": [
       60,

examples/nvnmd/train/train_qnn.json

@@ -1,6 +1,7 @@
 {
   "nvnmd": {
     "version": 0,
+    "max_nnei": 128,
     "net_size": 128,
     "sel": [
       60,

source/tests/test_nvnmd_entrypoints.py

@@ -521,6 +521,7 @@ def test_mapt_cnn_v1(self):
         map_file = str(tests_path / "nvnmd" / "out" / "map_v1_cnn.npy")
         # mapt
         mapObj = MapTable(config_file, weight_file, map_file)
+        mapObj.Gs_Gt_mode = 0
         mapt = mapObj.build_map()
         #
         N = 32
@@ -859,8 +860,9 @@ def test_wrap_qnn_v1(self):
         # test
         data = FioBin().load(jdata["nvnmd_model"])
         idx = [0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384]
+        idx = [i + 128*4 for i in idx]
         pred = [data[i] for i in idx]
-        red_dout = [1, 0, 0, 128, 0, 0, 0, 8, 249, 0, 0, 0, 91, 252, 183, 254]
+        red_dout = [249, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 254, 95, 24, 176]
         np.testing.assert_equal(pred, red_dout)
         # close
         nvnmd_cfg.enable = False