why the precision of the LiSi example is very low?

[sunlight905]

I test the example LiSi https://zhuanlan.zhihu.com/p/555628454, all the parameters is unchanged, but when the training complete, the precision is very low?
Could you give me some advice?
Thanks!
2023-11-07 19:41:16,655 - INFO : system 000 candidate : 228 in 20100 1.13 %
2023-11-07 19:41:16,655 - INFO : system 000 failed : 19352 in 20100 96.28 %
2023-11-07 19:41:16,655 - INFO : system 000 accurate : 520 in 20100 2.59 %
2023-11-07 19:41:16,664 - INFO : system 000 accurate_ratio: 0.0259 thresholds: 0.9970 and 0.9990 eff. task min and max -1 50 number of fp tasks: 50
2023-11-08 03:51:02,648 - INFO : system 000 candidate : 516 in 20100 2.57 %
2023-11-08 03:51:02,648 - INFO : system 000 failed : 18523 in 20100 92.15 %
2023-11-08 03:51:02,648 - INFO : system 000 accurate : 1061 in 20100 5.28 %
2023-11-08 03:51:02,656 - INFO : system 000 accurate_ratio: 0.0528 thresholds: 0.9970 and 0.9990 eff. task min and max -1 50 number of fp tasks: 50
2023-11-08 13:47:30,097 - INFO : system 000 candidate : 525 in 30100 1.74 %
2023-11-08 13:47:30,097 - INFO : system 000 failed : 28498 in 30100 94.68 %
2023-11-08 13:47:30,097 - INFO : system 000 accurate : 1077 in 30100 3.58 %
2023-11-08 13:47:30,110 - INFO : system 000 accurate_ratio: 0.0358 thresholds: 0.9970 and 0.9990 eff. task min and max -1 50 number of fp tasks: 50
2023-11-08 23:13:54,374 - INFO : system 000 candidate : 853 in 30100 2.83 %
2023-11-08 23:13:54,374 - INFO : system 000 failed : 27764 in 30100 92.24 %
2023-11-08 23:13:54,374 - INFO : system 000 accurate : 1483 in 30100 4.93 %
2023-11-08 23:13:54,387 - INFO : system 000 accurate_ratio: 0.0493 thresholds: 0.9970 and 0.9990 eff. task min and max -1 50 number of fp tasks: 50
2023-11-09 10:07:28,022 - INFO : system 000 candidate : 691 in 25100 2.75 %
2023-11-09 10:07:28,022 - INFO : system 000 failed : 23258 in 25100 92.66 %
2023-11-09 10:07:28,022 - INFO : system 000 accurate : 1151 in 25100 4.59 %
2023-11-09 10:07:28,033 - INFO : system 000 accurate_ratio: 0.0459 thresholds: 0.9970 and 0.9990 eff. task min and max -1 50 number of fp tasks: 50
2023-11-09 23:50:19,391 - INFO : system 000 candidate : 1567 in 40100 3.91 %
2023-11-09 23:50:19,391 - INFO : system 000 failed : 36697 in 40100 91.51 %
2023-11-09 23:50:19,391 - INFO : system 000 accurate : 1836 in 40100 4.58 %
2023-11-09 23:50:19,408 - INFO : system 000 accurate_ratio: 0.0458 thresholds: 0.9970 and 0.9990 eff. task min and max -1 50 number of fp tasks: 50

`{
"type_map": ["Li", "Si"],
"mass_map": [6.9, 28.1],
"init_data_prefix": "",
"init_data_sys": [
"POSCAR_LiSi.01x01x01/02.md/sys-0016-0016/deepmd"
],
"sys_configs_prefix": "",
"sys_configs": [
["POSCAR_LiSi.01x01x01/02.md/sys-0016-0016/scale-1.000/00000[0-2]/POSCAR"]
],

"_comment": " 00.train ",
"numb_models": 4,
"model_devi_activation_func" :        [["tanh","tanh"],["tanh","gelu"],["gelu","tanh"],["gelu","gelu"]],
  
"default_training_param": {
   "model": {
          "type_map": ["Li", "Si"],
          "descriptor": {
              "type": "se_e2_a",
              "sel": [100, 100],
              "rcut_smth": 2.0,
              "rcut": 6.0,
              "neuron":[25, 50, 100],
              "resnet_dt": false,
              "type_one_side": true,
              "axis_neuron": 12,
              "seed": 1
          },
          "fitting_net": {
              "neuron": [240, 240, 240],
              "resnet_dt": true,
              "seed": 1
          }
      },
      "learning_rate": {
          "type": "exp",
          "start_lr": 0.001,
          "decay_steps": 4000
      },
      "loss": {
          "start_pref_e": 0.02,
          "limit_pref_e": 1,
          "start_pref_f": 1000,
          "limit_pref_f": 1,
          "start_pref_v": 0.0,
          "limit_pref_v": 0.0
      },
      "training": {
          "stop_batch": 800000,
          "disp_file": "lcurve.out",
          "disp_freq": 2000,
          "numb_test": 4,
          "save_freq": 2000,
          "save_ckpt": "model.ckpt",
          "disp_training": true,
          "time_training": true,
          "profiling": false,
          "profiling_file": "timeline.json",
          "_comment": "that's all"
      }
},
  
"_comment": " 01.model_devi ",
"model_devi_dt": 0.001,
"model_devi_skip": 0,
"model_devi_f_trust_lo": 0.15,
"model_devi_f_trust_hi": 0.35,
"model_devi_clean_traj": false,
"model_devi_jobs": [
   {"sys_idx": [0],"temps": [200, 300, 400, 500],"press": [1.0, 100.0, 1000.0, 10000.0, 50000.0], "trj_freq": 5, "nsteps":   1000, "ensemble": "npt", "_idx":     "00" },
   {"sys_idx": [0],"temps": [200, 300, 400, 500],"press": [1.0, 100.0, 1000.0, 10000.0, 50000.0], "trj_freq": 5, "nsteps":   1000, "ensemble": "npt", "_idx":     "01" },
   {"sys_idx": [0],"temps": [200, 300, 400, 500],"press": [1.0, 100.0, 1000.0, 10000.0, 50000.0], "trj_freq": 10, "nsteps":   3000, "ensemble": "npt", "_idx":     "02" },
   {"sys_idx": [0],"temps": [200, 300, 400, 500],"press": [1.0, 100.0, 1000.0, 10000.0, 50000.0], "trj_freq": 10, "nsteps":   3000, "ensemble": "npt", "_idx":     "03" },
   {"sys_idx": [0],"temps": [200, 300, 400, 500],"press": [1.0, 100.0, 1000.0, 10000.0, 50000.0], "trj_freq": 20, "nsteps":   5000, "ensemble": "npt", "_idx":     "04" },
   {"sys_idx": [0],"temps": [200, 300, 400, 500],"press": [1.0, 100.0, 1000.0, 10000.0, 50000.0], "trj_freq": 20, "nsteps":   8000, "ensemble": "npt", "_idx":     "05" }
],
  
"_comment": " 02.fp ",
"fp_style": "vasp",
"fp_accurate_threshold": 0.999,
"fp_accurate_soft_threshold": 0.997,
"shuffle_poscar": false,
"fp_task_max": 50,
"fp_task_min": 3,
"ratio_failed": 0.05,
"fp_pp_path": "<your-POTCAR-path>",
"fp_pp_files": [
  "POTCAR_Li","POTCAR_Si"
],
"fp_incar": "./INCAR_LiSi"

} `