The accuracy of DP Gen based on DPA-2 pre trained model is 0 for each iteration

[JiangXiaoMingSan]

According to the notebook on Bohrium:DP-Gen based on a DPA-2 pretrained model(https://bohrium.dp.tech/notebooks/62585747598)
When I marked TiTaAlCrFeNiCo according to the notebook, the accuracy of each iteration was 0, and the specific status is as follows

stage id_stg. iter. accu. cand. fail. lvl_f_lo lvl_f_hi

Stage 0 --------------------

    0        0        0     0.0000     1.0000     0.0000     0.0035   0.5000
    0        1        1     0.0000     1.0000     0.0000     0.0127   0.5000
    0        2        2     0.0000     1.0000     0.0000     0.0122   0.5000
    0        3        3     0.0000     1.0000     0.0000     0.0139   0.5000

I reproduced the example of water in the notebook and obtained an iteration accuracy of 0, and the specific status is as follows

stage id_stg. iter. accu. cand. fail. lvl_f_lo lvl_f_hi

Stage 0 --------------------

    0        0        0     0.0000     1.0000     0.0000     0.0056   0.5000
  Each iteration has downloadable lable_data, which can be found in the folder called 'prep-run-fp/output/labeled_data'. But the accuracy of 0 confuses me. Is this a normal phenomenon?