datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
python-toolcallsLogs from run_python_code tool used for benchmarking.
Trajectory-Stitching-Test-7M
Dataset Creation & Methodology
Building this dataset required a highly optimized pipeline running on a dual-H100 NVL GPU cluster. The stitching process operates autonomously without relying on external LLM calls, using a specialized two-pass algorithm.
1. High-Information Keyword Extraction
Instead of relying on simple word counts, the pipeline dynamically builds a dataset-specific stopword list by analyzing Document Frequency (DF) to banish words appearing in more than… See the full description on the dataset page: https://huggingface.co/datasets/pythonformer/Trajectory-Stitching-Test-7M.SO-Python_QA-Data_Science_and_Machine_Learning_classreddit-python
Python Subreddit
Dataset containing data scraped from the Python subreddit.
SO-Python_QA-Networking_and_APIs_classSO-Python_QA-System_Administration_and_DevOps_classcode.evol.instruct.wiz.oss_python.jsonSO-Python_QA-Database_and_SQL_class
Dataset Card for Dataset Name
Dataset Summary
This dataset card aims to be a base template for new datasets. It has been generated using this raw template.
Supported Tasks and Leaderboards
[More Information Needed]
Languages
[More Information Needed]
Dataset Structure
Data Instances
[More Information Needed]
Data Fields
[More Information Needed]
Data Splits
[More Information Needed]
Dataset Creation… See the full description on the dataset page: https://huggingface.co/datasets/RazinAleks/SO-Python_QA-Database_and_SQL_class.SO-Python_QA-filtered-2023-tanh_score-after_2023_02SO dataset of pythontag data
Question filters:
images
links
code blocks
Q_Score > 0
Answer_count > 0
CreationDate > 2023-02-01
Answers filters:
images
links
code blocks
Scores are tanh applied to scaled with AbsMaxScaler to IQR range of Original SO Answers' scores
god_level_python_dataset_v1
God-Level Python Coder Dataset
A high-quality, synthetic dataset for training LLMs to achieve elite ("god-level") Python programming mastery.
Dataset Summary
This dataset contains 2,502 unique, advanced Python coding examples specifically designed to push large language models beyond basic problem-solving into true expert-level Python engineering.
It focuses on the hardest and most important areas of Python:
Deep metaprogramming
Production-grade asyncio &… See the full description on the dataset page: https://huggingface.co/datasets/11-47/god_level_python_dataset_v1.god_level_python_dataset_25k
God-Level Python Coder Dataset (25K Unique Advanced Examples)
Version: 1.0Size: Exactly 25,000 unique entries delivered. 100% synthetic with strong uniqueness guarantees via careful parameterization and deduplication.Focus: Training LLMs to achieve god-level mastery of Python — not just solving problems, but writing idiomatic, performant, robust, elegant, and deeply understood Python code.
This dataset is designed to push LLMs beyond basic LeetCode-style problems into true… See the full description on the dataset page: https://huggingface.co/datasets/11-47/god_level_python_dataset_25k.Python_Reason_D_Qwen_3_40kSO-Python_QA-filtered-2023-tanh_scoreSO dataset of pythontag data
Question filters:
images
links
Q_Score > 0
Answer_count > 0
Answers filters:
images
links
code blocks
Scores are tanh applied to scaled with AbsMaxScaler to IQR range of Original SO Answers' scores
SO-Python_basics_QA-filtered-2023-tanh_scoreSO dataset of python tag data and "Python basics and Envirinment" subcategory
Question filters:
images
links
code blocks
Q_Score > 0
Answer_count > 0
Answers filters:
images
links
code blocks
Scores are tanh applied to scaled with AbsMaxScaler to IQR range of Original SO Answers' scores
SO-Python_QA-filtered-2023-no_code-tanh_scoreSO dataset of pythontag data
Question filters:
images
links
code blocks
Q_Score > 0
Answer_count > 0
Answers filters:
images
links
code blocks
Scores are tanh applied to scaled with AbsMaxScaler to IQR range of Original SO Answers' scores
SO-Python_basics_QA-filtered-2023-T5_paraphrased-tanh_scoreTrajectory-Stitching-Test-SmallSO_Python_basics_QA_human_prefContrastive dataset for Stack Overflow python basics QA with augmentations:
SO-SO comparisons: 6166
Par-SO comparisons: 0
SO-Par comparisons: 36366
Gen-SO comparisons: 0
SO-Gen comparisons: 87114
Gen-Par comparisons: 0
Par-Gen comparisons: 0
Gen-Gen comparisons: 0
Par-Par comparisons: 55494
Paraphrasing model: humarin/chatgpt_paraphraser_on_T5_base
data-oss_instruct-decontaminated_python.jsonlSO-Python_QA-Other_classSO-Python_QA-API_USAGE_classPython_SO_domainsSO-Python_QA-Web_Development_classSO-Python_QA-GUI_Desktop_Applications_classnemotron-cc-small-subset-decontaminated-4conditions
Nemotron CC Small Subset Decontaminated: Pythonformer 4 Conditions
This dataset contains four serializations of the same source documents for
Pythonformer continued-pretraining experiments:
vanilla
alias_only
tool_only
alias_tool
Each condition is stored as gzipped JSONL shards under data/<condition>/.
from datasets import load_dataset
ds = load_dataset(
"pythonformer/nemotron-cc-small-subset-decontaminated-4conditions",
"alias_tool",
split="train",
)
BlackBeenie__Llama-3.1-8B-pythonic-passthrough-merge-details
Dataset Card for Evaluation run of BlackBeenie/Llama-3.1-8B-pythonic-passthrough-merge
Dataset automatically created during the evaluation run of model BlackBeenie/Llama-3.1-8B-pythonic-passthrough-merge
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/BlackBeenie__Llama-3.1-8B-pythonic-passthrough-merge-details.theprint__phi-3-mini-4k-python-details
Dataset Card for Evaluation run of theprint/phi-3-mini-4k-python
Dataset automatically created during the evaluation run of model theprint/phi-3-mini-4k-python
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/theprint__phi-3-mini-4k-python-details.SO-Python_QA-DS_ML_summ_classreddit-python1