Another issue is that my dataset has a huge number of files, making it difficult to use auto kaggle mode.

[HuaizhengZhang]

Project Structure:
.
├── DOCUMENTATION.md
├── README.md
├── data_analysis.py
├── sample_submission.csv
├── test
│ ├── C
│ └── D
├── test.csv
└── train
├── TRAIN_A.csv
└── TRAIN_B.csv

4 directories, 7 files
\n=== Random 10 samples from test/C: ===
TEST_C_1552.csv
TEST_C_0105.csv
TEST_C_1141.csv
TEST_C_0179.csv
TEST_C_1494.csv
TEST_C_2346.csv
TEST_C_1823.csv
TEST_C_0539.csv
TEST_C_0546.csv
TEST_C_1876.csv
\n=== Random 10 samples from test/D: ===
TEST_D_1957.csv
TEST_D_1396.csv
TEST_D_1067.csv
TEST_D_0797.csv
TEST_D_0204.csv
TEST_D_0306.csv
TEST_D_1668.csv
TEST_D_1316.csv
TEST_D_2591.csv
TEST_D_2208.csv