diff --git a/README.md b/README.md
index 162603099a..bac29b4ddd 100644
--- a/README.md
+++ b/README.md
@@ -23,7 +23,7 @@ For more information, check the [documentation](https://deepmd.readthedocs.io/).
 * [Model compression](doc/getting-started.md#compress-a-model). Accelerate the efficiency of model inference for 4-15 times.
 * [New descriptors](doc/getting-started.md#write-the-input-script). Including [`se_e2_r`](doc/train-se-e2-r.md) and [`se_e3`](doc/train-se-e3.md).
 * [Hybridization of descriptors](doc/train-hybrid.md). Hybrid descriptor constructed from concatenation of several descriptors.
-* Atom type embedding.
+* [Atom type embedding](doc/train-se-e2-a-tebd.md). Enable atom type embedding to decline training complexity and refine performance.
 * Training and inference the dipole (vector) and polarizability (matrix).
 * Split of training and validation dataset.
 * Optimized training on GPUs. 
diff --git a/doc/train-se-e2-a-tebd.md b/doc/train-se-e2-a-tebd.md
new file mode 100644
index 0000000000..e895adc858
--- /dev/null
+++ b/doc/train-se-e2-a-tebd.md
@@ -0,0 +1,42 @@
+# Train a Deep Potential model using `type embedding` approach
+ 
+We generate specific type embedding vector for each atom type, so that we can share one descriptor embedding net and one fitting net in total, which decline training complexity largely. 
+
+The training input script is similar to that of [`se_e2_a`](train-se-e2-a.md#the-training-input-script), but different by adding the `type_embedding` section. 
+
+### Type embedding net
+The `model` defines how the model is constructed, adding a section of type embedding net:
+```json=
+    "model": {
+	"type_map":	["O", "H"],
+	"type_embedding":{
+			...
+	},	    
+	"descriptor" :{
+            ...
+	},
+	"fitting_net" : {
+            ...
+	}
+    }
+```
+Model will automatically apply type embedding approach and generate type embedding vectors. If type embedding vector is detected, descriptor and fitting net would take it as a part of input.
+
+The construction of type embedding net is given by `type_embedding`. An example of `type_embedding` is provided as follows
+```json=
+	"type_embedding":{
+	    "neuron":		[2, 4, 8],
+	    "resnet_dt":	false,
+	    "seed":		1
+	}
+```
+* The `neuron` specifies the size of the type embedding net. From left to right the members denote the sizes of each hidden layer from input end to the output end, respectively. It takes one-hot vector as input and output dimension equals to the last dimension of the `neuron` list. If the outer layer is of twice size as the inner layer, then the inner layer is copied and concatenated, then a [ResNet architecture](https://arxiv.org/abs/1512.03385) is built between them.
+* If the option `resnet_dt` is set `true`, then a timestep is used in the ResNet.
+* `seed` gives the random seed that is used to generate random numbers when initializing the model parameters.
+
+
+
+A complete training input script of this example can be find in the directory. 
+```bash
+$deepmd_source_dir/examples/water/se_e2_a_tebd/input.json
+```
\ No newline at end of file
diff --git a/doc/train-se-e2-a.md b/doc/train-se-e2-a.md
index 03bf33295d..c69b17a574 100644
--- a/doc/train-se-e2-a.md
+++ b/doc/train-se-e2-a.md
@@ -47,6 +47,7 @@ The construction of the descriptor is given by section `descriptor`. An example
 	    "rcut":		6.00,
 	    "sel":		[46, 92],
 	    "neuron":		[25, 50, 100],
+	    "type_one_side":	true,
 	    "axis_neuron":	16,
 	    "resnet_dt":	false,
 	    "seed":		1
@@ -56,6 +57,7 @@ The construction of the descriptor is given by section `descriptor`. An example
 * `rcut` is the cut-off radius for neighbor searching, and the `rcut_smth` gives where the smoothing starts. 
 * `sel` gives the maximum possible number of neighbors in the cut-off radius. It is a list, the length of which is the same as the number of atom types in the system, and `sel[i]` denote the maximum possible number of neighbors with type `i`. 
 * The `neuron` specifies the size of the embedding net. From left to right the members denote the sizes of each hidden layer from input end to the output end, respectively. If the outer layer is of twice size as the inner layer, then the inner layer is copied and concatenated, then a [ResNet architecture](https://arxiv.org/abs/1512.03385) is built between them.
+* If the option `type_one_side` is set to `true`, then descriptor will consider the types of neighbor atoms. Otherwise, both the types of centric and  neighbor atoms are considered.
 * The `axis_neuron` specifies the size of submatrix of the embedding matrix, the axis matrix as explained in the [DeepPot-SE paper](https://arxiv.org/abs/1805.09003) 
 * If the option `resnet_dt` is set `true`, then a timestep is used in the ResNet.
 * `seed` gives the random seed that is used to generate random numbers when initializing the model parameters.