bug: fix spin nlist in spin_model (#3718)

## Summary by CodeRabbit - **New Features** - Expanded backend support to include PyTorch and DP in the energy spin training documentation. - Updated spin settings for TensorFlow and PyTorch/DP in the energy spin training documentation. - Added new sections for loss functions and data preparation in the energy spin training documentation. - **Bug Fixes** - Adjusted test conditions and initialization parameters in various model tests to align with updated functionalities. - **Tests** - Increased selection values in permutation tests to enhance test coverage and reliability.  --------- Co-authored-by: Han Wang <[email protected]>
deepmodeling · May 8, 2024 · 6d0874e · 6d0874e
1 parent 6f06adb
commit 6d0874e
Show file tree

Hide file tree

Showing 4 changed files with 94 additions and 22 deletions.
diff --git a/deepmd/pt/model/model/spin_model.py b/deepmd/pt/model/model/spin_model.py
@@ -170,15 +170,20 @@ def extend_nlist(extended_atype, nlist):
         nlist_shift = nlist + nall
         nlist[~nlist_mask] = -1
         nlist_shift[~nlist_mask] = -1
-        self_spin = torch.arange(0, nloc, dtype=nlist.dtype, device=nlist.device) + nall
-        self_spin = self_spin.view(1, -1, 1).expand(nframes, -1, -1)
-        # self spin + real neighbor + virtual neighbor
+        self_real = (
+            torch.arange(0, nloc, dtype=nlist.dtype, device=nlist.device)
+            .view(1, -1, 1)
+            .expand(nframes, -1, -1)
+        )
+        self_spin = self_real + nall
+        # real atom's neighbors: self spin + real neighbor + virtual neighbor
+        # nf x nloc x (1 + nnei + nnei)
+        real_nlist = torch.cat([self_spin, nlist, nlist_shift], dim=-1)
+        # spin atom's neighbors: real + real neighbor + virtual neighbor
         # nf x nloc x (1 + nnei + nnei)
-        extended_nlist = torch.cat([self_spin, nlist, nlist_shift], dim=-1)
+        spin_nlist = torch.cat([self_real, nlist, nlist_shift], dim=-1)
         # nf x (nloc + nloc) x (1 + nnei + nnei)
-        extended_nlist = torch.cat(
-            [extended_nlist, -1 * torch.ones_like(extended_nlist)], dim=-2
-        )
+        extended_nlist = torch.cat([real_nlist, spin_nlist], dim=-2)
         # update the index for switch
         first_part_index = (nloc <= extended_nlist) & (extended_nlist < nall)
         second_part_index = (nall <= extended_nlist) & (extended_nlist < (nall + nloc))

diff --git a/doc/model/train-energy-spin.md b/doc/model/train-energy-spin.md
@@ -1,14 +1,35 @@
-# Fit spin energy {{ tensorflow_icon }}
+# Fit spin energy {{ tensorflow_icon }} {{ pytorch_icon }} {{ dpmodel_icon }}
 
 :::{note}
-**Supported backends**: TensorFlow {{ tensorflow_icon }}
+**Supported backends**: TensorFlow {{ tensorflow_icon }}, PyTorch {{ pytorch_icon }}, DP {{ dpmodel_icon }}
 :::
 
-In this section, we will take `$deepmd_source_dir/examples/NiO/se_e2_a/input.json` as an example of the input file.
+To train a model that takes additional spin information as input, you only need to modify the following sections to define the spin-specific settings,
+keeping other sections the same as the normal energy model's input script.
+
+:::{warning}
+Note that when adding spin into the model, there will be some implicit modifications automatically done by the program:
+
+- In the TensorFlow backend, the `se_e2_a` descriptor will treat those atom types with spin as new (virtual) types,
+  and duplicate their corresponding selected numbers of neighbors ({ref}`sel <model/descriptor[se_e2_a]/sel>`) from their real atom types.
+- In the PyTorch backend, if spin settings are added, all the types (with or without spin) will have their virtual types.
+  The `se_e2_a` descriptor will thus double the {ref}`sel <model/descriptor[se_e2_a]/sel>` list,
+  while in other descriptors with mixed types (such as `dpa1` or `dpa2`), the sel number will not be changed for clarity.
+  If you are using descriptors with mixed types, to achieve better performance,
+  you should manually extend your sel number (maybe double) depending on the balance between performance and efficiency.
+  :::
 
 ## Spin
 
-The construction of the fitting net is give by section {ref}`spin <model/spin>`
+The spin settings are given by the {ref}`spin <model/spin>` section, which sets the magnetism for each type of atoms as described in the following sections.
+
+:::{note}
+Note that the construction of spin settings is different between TensorFlow and PyTorch/DP.
+:::
+
+### Spin settings in TensorFlow
+
+The implementation in TensorFlow only supports `se_e2_a` descriptor. See examples in `$deepmd_source_dir/examples/spin/se_e2_a/input_tf.json`, the {ref}`spin <model/spin>` section is defined as the following:
 
 ```json
     "spin" : {
@@ -18,10 +39,38 @@ The construction of the fitting net is give by section {ref}`spin <model/spin>`
     },
 ```
 
-- {ref}`use_spin <model/spin[ener_spin]/use_spin>` determines whether to turn on the magnetism of the atoms.The index of this option matches option `type_map <model/type_map>`.
+- {ref}`use_spin <model/spin[ener_spin]/use_spin>` is a list of boolean values indicating whether to use atomic spin for each atom type.
+  True for spin and False for not. The index of this option matches option `type_map <model/type_map>`.
 - {ref}`virtual_len <model/spin[ener_spin]/virtual_len>` specifies the distance between virtual atom and the belonging real atom.
 - {ref}`spin_norm <model/spin[ener_spin]/spin_norm>` gives the magnitude of the magnetic moment for each magnatic atom.
 
+### Spin settings in PyTorch/DP
+
+In PyTorch/DP, the spin implementation is more flexible and so far supports the following descriptors:
+
+- `se_e2_a`
+- `dpa1`(`se_atten`)
+- `dpa2`
+
+See `se_e2_a` examples in `$deepmd_source_dir/examples/spin/se_e2_a/input_torch.json`, the {ref}`spin <model/spin>` section is defined as the following with a much more clear interface:
+
+```json
+    "spin": {
+      "use_spin": [true, false],
+      "virtual_scale": [0.3140]
+    },
+```
+
+- {ref}`use_spin <model/spin[ener_spin]/use_spin>` is a list of boolean values indicating whether to use atomic spin for each atom type, or a list of type indexes that use atomic spin.
+  The index of this option matches option `type_map <model/type_map>`.
+- {ref}`virtual_len <model/spin[ener_spin]/virtual_scale>` defines the scaling factor to determine the virtual distance
+  between a virtual atom representing spin and its corresponding real atom
+  for each atom type with spin. This factor is defined as the virtual distance
+  divided by the magnitude of atomic spin for each atom type with spin.
+  The virtual coordinate is defined as the real coordinate plus spin \* virtual_scale.
+  List of float values with shape of `ntypes` or `ntypes_spin` or one single float value for all types,
+  only used when {ref}`use_spin <model/spin[ener_spin]/use_spin>` is True for each atom type.
+
 ## Spin Loss
 
 The spin loss function $L$ for training energy is given by
@@ -59,3 +108,7 @@ The {ref}`loss <loss>` section in the `input.json` is
 The options {ref}`start_pref_e <loss[ener_spin]/start_pref_e>`, {ref}`limit_pref_e <loss[ener_spin]/limit_pref_e>`, {ref}`start_pref_fr <loss[ener_spin]/start_pref_fr>`, {ref}`limit_pref_fm <loss[ener_spin]/limit_pref_fm>`, {ref}`start_pref_v <loss[ener_spin]/start_pref_v>` and {ref}`limit_pref_v <loss[ener_spin]/limit_pref_v>` determine the start and limit prefactors of energy, atomic force, magnatic force and virial, respectively.
 
 If one does not want to train with virial, then he/she may set the virial prefactors {ref}`start_pref_v <loss[ener_spin]/start_pref_v>` and {ref}`limit_pref_v <loss[ener_spin]/limit_pref_v>` to 0.
+
+## Data preparation
+
+(Need a documentation for data format for TensorFlow and PyTorch/DP.)
diff --git a/source/tests/pt/model/test_forward_lower.py b/source/tests/pt/model/test_forward_lower.py
@@ -82,7 +82,9 @@ def test(
         ) = extend_input_and_build_neighbor_list(
             coord.unsqueeze(0),
             atype.unsqueeze(0),
-            self.model.get_rcut(),
+            self.model.get_rcut() + 1.0
+            if test_spin
+            else self.model.get_rcut(),  # buffer region for spin nlist
             self.model.get_sel(),
             mixed_types=self.model.mixed_types(),
             box=cell.unsqueeze(0),
@@ -128,15 +130,13 @@ class TestEnergyModelSeA(unittest.TestCase, ForwardLowerTest):
     def setUp(self):
         self.prec = 1e-10
         model_params = copy.deepcopy(model_se_e2_a)
-        self.type_split = False
         self.model = get_model(model_params).to(env.DEVICE)
 
 
 class TestEnergyModelDPA1(unittest.TestCase, ForwardLowerTest):
     def setUp(self):
         self.prec = 1e-10
         model_params = copy.deepcopy(model_dpa1)
-        self.type_split = True
         self.model = get_model(model_params).to(env.DEVICE)
 
 
@@ -151,24 +151,38 @@ def setUp(self):
             "repinit_nsel"
         ]
         model_params = copy.deepcopy(model_dpa2)
-        self.type_split = True
         self.model = get_model(model_params).to(env.DEVICE)
 
 
 class TestEnergyModelZBL(unittest.TestCase, ForwardLowerTest):
     def setUp(self):
         self.prec = 1e-10
         model_params = copy.deepcopy(model_zbl)
-        self.type_split = False
         self.model = get_model(model_params).to(env.DEVICE)
 
 
 class TestEnergyModelSpinSeA(unittest.TestCase, ForwardLowerTest):
     def setUp(self):
-        # still need to figure out why only 1e-5 rtol and atol
-        self.prec = 1e-5
+        self.prec = 1e-10
+        model_params = copy.deepcopy(model_spin)
+        self.test_spin = True
+        self.model = get_model(model_params).to(env.DEVICE)
+
+
+class TestEnergyModelSpinDPA1(unittest.TestCase, ForwardLowerTest):
+    def setUp(self):
+        self.prec = 1e-10
+        model_params = copy.deepcopy(model_spin)
+        model_params["descriptor"] = copy.deepcopy(model_dpa1)["descriptor"]
+        self.test_spin = True
+        self.model = get_model(model_params).to(env.DEVICE)
+
+
+class TestEnergyModelSpinDPA2(unittest.TestCase, ForwardLowerTest):
+    def setUp(self):
+        self.prec = 1e-10
         model_params = copy.deepcopy(model_spin)
-        self.type_split = False
+        model_params["descriptor"] = copy.deepcopy(model_dpa2)["descriptor"]
         self.test_spin = True
         self.model = get_model(model_params).to(env.DEVICE)
 

diff --git a/source/tests/pt/model/test_permutation.py b/source/tests/pt/model/test_permutation.py
@@ -116,10 +116,10 @@
         "type": "dpa2",
         "repinit_rcut": 6.0,
         "repinit_rcut_smth": 2.0,
-        "repinit_nsel": 30,
+        "repinit_nsel": 100,
         "repformer_rcut": 4.0,
         "repformer_rcut_smth": 0.5,
-        "repformer_nsel": 20,
+        "repformer_nsel": 40,
         "repinit_neuron": [2, 4, 8],
         "repinit_axis_neuron": 4,
         "repinit_activation": "tanh",