Skip to content

🌟 The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming

FE prompt: FE通常不会make changes而是加新的特征 DI prompt: 要求让predictions最终结果一致,并提供例子

Yizhou Chi1cdffc3d8550
diff --git a/expo/data/dataset.py b/expo/data/dataset.py
index 9748cb8c2..28bd26d2e 100644
--- a/expo/data/dataset.py
+++ b/expo/data/dataset.py
@@ -40,6 +40,7 @@ DI_INSTRUCTION = """
1. Save the prediction results of BOTH the dev set and test set in `dev_predictions.csv` and `test_predictions.csv` respectively in the output directory.
- Both files should contain a single column named `target` with the predicted values.
2. Make sure the prediction results are in the same format as the target column in the training set.
+- For instance, if the target column is categorical, the prediction results should be categorical as well.
## Output Performance
Print the train and dev set performance in the last step.
diff --git a/metagpt/prompts/task_type.py b/metagpt/prompts/task_type.py
index 6b230fc9e..599d437c5 100644
--- a/metagpt/prompts/task_type.py
+++ b/metagpt/prompts/task_type.py
@@ -26,7 +26,7 @@ The current task is about feature engineering. when performing it, please adhere
- Avoid creating redundant or excessively numerous features in one step.
- Exclude ID columns from feature generation and remove them.
- Each feature engineering operation performed on the train set must also applies to the dev/test separately at the same time.
-- **ATTENTION** Do NOT use the label column to create features or make any changes to the label column, except for cat encoding.
+- **ATTENTION** Do NOT use the label column to create features, except for cat encoding.
- Use the data from previous task result if exist, do not mock or reload data yourself.
- Always copy the DataFrame before processing it and use the copy to process.
"""