updates

Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>

deepmd/dpmodel/descriptor/dpa2.py

@@ -387,7 +387,7 @@ def __init__(
         use_tebd_bias: bool = False,
         type_map: Optional[list[str]] = None,
     ) -> None:
-        r"""The DPA-2 descriptor. see https://arxiv.org/abs/2312.15492.
+        r"""The DPA-2 descriptor[1]_.
 
         Parameters
         ----------
@@ -434,6 +434,11 @@ def __init__(
         sw:                 torch.Tensor
             The switch function for decaying inverse distance.
 
+        References
+        ----------
+        .. [1] Zhang, D., Liu, X., Zhang, X. et al. DPA-2: a
+           large atomic model as a multi-task learner. npj
+           Comput Mater 10, 293 (2024). https://doi.org/10.1038/s41524-024-01493-2
         """
 
         def init_subclass_params(sub_data, sub_class):

deepmd/pt/model/descriptor/dpa2.py

@@ -100,7 +100,7 @@ def __init__(
         use_tebd_bias: bool = False,
         type_map: Optional[list[str]] = None,
     ) -> None:
-        r"""The DPA-2 descriptor. see https://arxiv.org/abs/2312.15492.
+        r"""The DPA-2 descriptor[1]_.
 
         Parameters
         ----------
@@ -147,6 +147,11 @@ def __init__(
         sw:                 torch.Tensor
             The switch function for decaying inverse distance.
 
+        References
+        ----------
+        .. [1] Zhang, D., Liu, X., Zhang, X. et al. DPA-2: a
+           large atomic model as a multi-task learner. npj
+           Comput Mater 10, 293 (2024). https://doi.org/10.1038/s41524-024-01493-2
         """
         super().__init__()
 

doc/train/finetuning.md

@@ -94,7 +94,7 @@ The model section will be overwritten (except the `type_map` subsection) by that
 
 #### Fine-tuning from a multi-task pre-trained model
 
-Additionally, within the PyTorch implementation and leveraging the flexibility offered by the framework and the multi-task training process proposed in DPA2 [paper](https://arxiv.org/abs/2312.15492),
+Additionally, within the PyTorch implementation and leveraging the flexibility offered by the framework and the multi-task training process proposed in DPA2 [paper](https://doi.org/10.1038/s41524-024-01493-2),
 we also support more general multitask pre-trained models, which includes multiple datasets for pre-training. These pre-training datasets share a common descriptor while maintaining their individual fitting nets,
 as detailed in the paper above.
 

doc/train/multi-task-training.md

@@ -26,7 +26,7 @@ and the Adam optimizer is executed to minimize $L^{(t)}$ for one step to update
 In the case of multi-GPU parallel training, different GPUs will independently select their tasks.
 In the DPA-2 model, this multi-task training framework is adopted.[^1]
 
-[^1]: Duo Zhang, Xinzijian Liu, Xiangyu Zhang, Chengqian Zhang, Chun Cai, Hangrui Bi, Yiming Du, Xuejian Qin, Jiameng Huang, Bowen Li, Yifan Shan, Jinzhe Zeng, Yuzhi Zhang, Siyuan Liu, Yifan Li, Junhan Chang, Xinyan Wang, Shuo Zhou, Jianchuan Liu, Xiaoshan Luo, Zhenyu Wang, Wanrun Jiang, Jing Wu, Yudi Yang, Jiyuan Yang, Manyi Yang, Fu-Qiang Gong, Linshuang Zhang, Mengchao Shi, Fu-Zhi Dai, Darrin M. York, Shi Liu, Tong Zhu, Zhicheng Zhong, Jian Lv, Jun Cheng, Weile Jia, Mohan Chen, Guolin Ke, Weinan E, Linfeng Zhang, Han Wang, [arXiv preprint arXiv:2312.15492 (2023)](https://arxiv.org/abs/2312.15492) licensed under a [Creative Commons Attribution (CC BY) license](http://creativecommons.org/licenses/by/4.0/).
+[^1]: Duo Zhang, Xinzijian Liu, Xiangyu Zhang, Chengqian Zhang, Chun Cai, Hangrui Bi, Yiming Du, Xuejian Qin, Anyang Peng, Jiameng Huang, Bowen Li, Yifan Shan, Jinzhe Zeng, Yuzhi Zhang, Siyuan Liu, Yifan Li, Junhan Chang, Xinyan Wang, Shuo Zhou, Jianchuan Liu, Xiaoshan Luo, Zhenyu Wang, Wanrun Jiang, Jing Wu, Yudi Yang, Jiyuan Yang, Manyi Yang, Fu-Qiang Gong, Linshuang Zhang, Mengchao Shi, Fu-Zhi Dai, Darrin M. York, Shi Liu, Tong Zhu, Zhicheng Zhong, Jian Lv, Jun Cheng, Weile Jia, Mohan Chen, Guolin Ke, Weinan E, Linfeng Zhang, Han Wang, DPA-2: a large atomic model as a multi-task learner. npj Comput Mater 10, 293 (2024). [DOI: 10.1038/s41524-024-01493-2](https://doi.org/10.1038/s41524-024-01493-2) licensed under a [Creative Commons Attribution (CC BY) license](http://creativecommons.org/licenses/by/4.0/).
 
 Compared with the previous TensorFlow implementation, the new support in PyTorch is more flexible and efficient.
 In particular, it makes multi-GPU parallel training and even tasks beyond DFT possible,