gavindiaz/NexQuant
1
0
Fork 0
mirror of https://github.com/NicolasBohn/NexQuant.git synced 2026-07-27 23:47:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: update the documentation for custom dataset (#969)

Browse Source
This commit is contained in:
Yuante Li
2025-06-18 10:58:28 +08:00
committed by GitHub
parent 9f793d58b3
commit 0c219b81e0
1 changed files with 14 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/scens/data_science.rst
+14 -1
View File
@@ -26,7 +26,7 @@ The Data Science Agent is an agent that can automatically perform feature engine
- 📥 **Prepare Competition Data**
- Data Science competition data, contains three parts: competition description file (markdown file), competition dataset and competition evaluation files.
- Data Science competition data typically consists of three components: a competition description file (in Markdown format), the competition dataset, and evaluation scripts. For reference, an example of a custom user-defined dataset is provided in ``rdagent/scenarios/data_science/example``.
- **Correct directory structure (Here is an example of competition data with id custom_data)**
@@ -43,6 +43,7 @@ The Data Science Agent is an agent that can automatically perform feature engine
└── test.csv
└── sample_submission.csv
└── description.md
└── sample.py
- ``ds_data/custom_data/train.csv:`` Necessary training data in csv or parquet format, or training images.
@@ -50,12 +51,24 @@ The Data Science Agent is an agent that can automatically perform feature engine
- ``ds_data/custom_data/sample_submission.csv:`` (Optional) Competition sample submission file.
- ``ds_data/custom_data/sample.py:`` (Optional) Sample code for generating debug data from the competition dataset. If not provided, R&D-Agent will use its default sampling logic. For details, see the ``create_debug_data`` function in ``rdagent/scenarios/data_science/debug/data.py``.
- ``ds_data/eval/custom_data/grade.py:`` (Optional) Competition grade script, in order to calculate the score for the submission.
- ``ds_data/eval/custom_data/valid.py:`` (Optional) Competition validation script, in order to check if the submission format is correct.
- ``ds_data/eval/custom_data/submission_test.csv:`` (Optional) Competition test label file.
- 🔧 **Set up Environment for Custom User-defined Dataset**
.. code-block:: sh
dotenv set DS_SCEN rdagent.scenarios.data_science.scen.DataScienceScen
dotenv set DS_LOCAL_DATA_PATH <your local directory>/ds_data (e.g. rdagent/scenarios/data_science/example)
dotenv set DS_IF_USING_MLE_DATA False
dotenv set DS_CODER_ON_WHOLE_PIPELINE True
dotenv set DS_CODER_COSTEER_ENV_TYPE docker
🔍 MLE-bench Guide: Running ML Engineering via MLE-bench
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~