liaozhaorun/ProStock
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(training): 保留 test 集中标签为 NaN 的样本用于预测

Browse Source
This commit is contained in:
liaozhaorun
2026-04-05 23:24:22 +08:00
parent 1fa4ff9544
commit ad8ba8f6ec
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/training/pipeline.py
View File

@@ -374,8 +374,8 @@ class DataPipeline:
split_data[split_name]["X"] = split_df.select(feature_cols) split_data[split_name]["X"] = split_df.select(feature_cols)
split_data[split_name]["y"] = split_df[label_name] split_data[split_name]["y"] = split_df[label_name]
# 删除标签为 NaN 的行 # 删除标签为 NaN 的行（仅在 train/val 上执行，test 集保留用于预测）
for split_name in ["train", "val", "test"]: for split_name in ["train", "val"]:
if split_name in split_data: if split_name in split_data:
y_series = split_data[split_name]["y"] y_series = split_data[split_name]["y"]
y_nan_count = y_series.null_count() y_nan_count = y_series.null_count()