datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
CodeFeedback-Filtered-Instruction OpenCodeInterpreter: Integrating Code Generation with Execution and Refinement
[🏠Homepage]
|
[🛠️Code]
OpenCodeInterpreter
OpenCodeInterpreter is a family of open-source code generation systems designed to bridge the gap between large language models and advanced proprietary systems like the GPT-4 Code Interpreter. It significantly advances code generation capabilities by integrating execution and iterative refinement functionalities.
For further information and… See the full description on the dataset page: https://huggingface.co/datasets/m-a-p/CodeFeedback-Filtered-Instruction.python-codes-25k
License
MIT
This is a Cleaned Python Dataset Covering 25,000 Instructional Tasks
Overview
The dataset has 4 key features (fields): instruction, input, output, and text.It's a rich source for Python codes, tasks, and extends into behavioral aspects.
Dataset Statistics
Total Entries: 24,813
Unique Instructions: 24,580
Unique Inputs: 3,666
Unique Outputs: 24,581
Unique Texts: 24,813
Average Tokens per example: 508
Features… See the full description on the dataset page: https://huggingface.co/datasets/flytech/python-codes-25k.Code-Feedback OpenCodeInterpreter: Integrating Code Generation with Execution and Refinement
[🏠Homepage]
|
[🛠️Code]
Introduction
OpenCodeInterpreter is a family of open-source code generation systems designed to bridge the gap between large language models and advanced proprietary systems like the GPT-4 Code Interpreter. It significantly advances code generation capabilities by integrating execution and iterative refinement functionalities.
For further information and related… See the full description on the dataset page: https://huggingface.co/datasets/m-a-p/Code-Feedback.R1-Code-Interpreter-Data
R1-Code-Interpreter: Training LLMs to Reason with Code via Supervised and Reinforcement Learning
Our code is based on Llama-factory/VeRL/Search-R1 for the SFT and RL training and SymBench/BIG-Bench-Hard/reasoning-gym for datasets/benchmarks of reasoning/planning tasks.
📝 Introduction
R1-Code-Interpreter is the first framework to train LLMs for step-by-step code reasoning using multi-turn supervised fine-tuning and reinforcement learning. By curating 144 diverse… See the full description on the dataset page: https://huggingface.co/datasets/yongchao98/R1-Code-Interpreter-Data.mimo-claude-code-traces-1k
MIMO Claude Code Traces
MIMO Claude Code Traces is a collection of coding-agent trajectories in a Claude Code-style environment. Each record contains a user coding task, the full multi-turn message trace, available tool schemas, assistant reasoning fields, tool calls, tool outputs, and metadata such as model name, category, duration, cost, token usage, and whether the trace used tools.
The traces were generated with mimo-v2.5-pro, MiMo's most capable model at the time of… See the full description on the dataset page: https://huggingface.co/datasets/choucsan/mimo-claude-code-traces-1k.math-code-science-deepseek-r1-en
R1 Dataset Collection
Aggregated high-quality English prompts and model-generated responses from DeepSeek R1 and DeepSeek R1-0528.
Dataset Summary
The R1 Dataset Collection combines multiple public DeepSeek-generated instruction-response corpora into a single, cleaned, English-only JSONL file. Each example consists of a <|user|> prompt and a <|assistant|> response in one "text" field. This release includes:
~21,000 examples from the DeepSeek-R1-0528 Distilled Custom… See the full description on the dataset page: https://huggingface.co/datasets/Hugodonotexit/math-code-science-deepseek-r1-en.github-code-fontend-lang
github-code fontend code
Dwonload
方式一
huggingface-cli download --resume-download LiXiang12/github-code-fontend-lang --include "*/*.zip" --repo-type dataset --local-dir github_code
方式二
进入Files and versions/data直接下载zip文件
数据统计
CodeArena
Dataset Summary
To bridge the gap between the model-generated response and human preference, we present a rigorous human-curated benchmark CodeArena to emulate the complexity and diversity of real-world coding tasks, where 397 high-quality samples spanning 40 categories and 40 languages, carefully curated from user queries.
Data Example
An example of 'validation' looks as follows:
{
"id": "60670a8d9b1e39dd845fb1639d0d8b86",
"messages": "[{'role': 'user'… See the full description on the dataset page: https://huggingface.co/datasets/CSJianYang/CodeArena.us-legal-code
Dataset Card for United States Code (Cornell LII) — Hierarchical Sections
Dataset Summary
This dataset is purpose-built for the Prime Intellect U.S. legal evaluation environment.
This dataset contains the text of the United States Code scraped from the Legal Information Institute at Cornell Law School. Each record corresponds to a navigable section (“U.S. Code” tab only) together with its hierarchy path—title, subtitle, division, part, subpart, chapter, subchapter, and so… See the full description on the dataset page: https://huggingface.co/datasets/emre570/us-legal-code.StackPulse_778K_QnA_Code_dataset
💻 StackOverflow-778K: Multi-Year Developer Q&A Dataset
Dataset Summary
A large-scale Stack Overflow question dataset containing 778,929 unique
questions sampled across 7 years (2015–2022). Each question includes the
raw HTML body, plain-text version, tags, score, view count, answer count, and
a rich set of derived features for immediate ML use.
Collected across 8 sampling runs on Feb 27 2026, deduplicated to
778,929 unique questions with only 2 duplicates removed.… See the full description on the dataset page: https://huggingface.co/datasets/Omarrran/StackPulse_778K_QnA_Code_dataset.swiss-code-of-obligations
Swiss Code of Obligations (OR) and Swiss Civil Code
(Part Five: The Code of Obligations) of 30 March 1911 (Status as of 1 September 2023)
Files generated from the Swiss publication platform for federal law
Swiss Code of Obligations
Format
Each article has the following type definition:
With vector embeddings by Xenova/paraphrase-multilingual-mpnet-base-v2
swiss-civil-code-de-paraphrase-multilingual-mpnet-base-v2.jsonl… See the full description on the dataset page: https://huggingface.co/datasets/brunnolou/swiss-code-of-obligations.Stable-Code-Python-SFT
Stable Code Python SFT
The Stable Code Python SFT dataset is a high-quality synthetic dataset derived from the
stabilityai/stable-code-instruct-3b model for the purpose of supervised fine-tuning (SFT). Please refer to the
Versioning section for dataset versions.
Note: If you would like to contribute to this repository,
please read the CONTRIBUTING first.
TableofContents
Features
File Structure
Metadata
Usage
Versioning
License
TeamContact
Reference
Citation… See the full description on the dataset page: https://huggingface.co/datasets/bunyaminergen/Stable-Code-Python-SFT.code_leak_qaCodeGen-Diverse-5K
CodeGen-Diverse-5K: Broad Coverage for Competitive Programming
Part of the CodeGen suite | CodeGen-Deep-5K (sister dataset)
Dataset Description
CodeGen-Diverse-5K is a broad coverage dataset designed for training code generation models across a wide variety of competitive programming problems. This dataset prioritizes problem diversity over solution diversity, covering 5,000 unique problems with consistent, high-quality solutions.
Key Statistics
Total samples:… See the full description on the dataset page: https://huggingface.co/datasets/Naholav/CodeGen-Diverse-5K.code-gradients
Gradients code comparison
This public copy contains 18,718 training rows and 998 test rows. Each row has
exactly one user turn followed by one assistant turn under conversations. The local
source uses the equivalent top-level key messages; only that key was renamed for
compatibility with Gradients baseline preparation.
llama-python-codes-30k
Python Codes - 30k examples, Llama1&2 tokenized dataset
Author
FlyTech
For general guide on how to create, quantize, merge or inference the model and more, visit:
hackmd.io/my_first_ai
Overview
This dataset serves as a rich resource for various Natural Language Processing tasks such as:
Question Answering
Text Generation
Text-to-Text Generation
It primarily focuses on instructional tasks in Python, tokenized specifically for the Llama architecture.… See the full description on the dataset page: https://huggingface.co/datasets/flytech/llama-python-codes-30k.code-training-il
Code-Training-IL
A 40,330-example instruction-tuning dataset for code: 20K Python (NVIDIA OpenCodeInstruct, test-filtered) + 20K TypeScript + 330 hand-written bilingual identity examples.
Overview
code-training-il is a curated, filtered instruction-tuning corpus for training small coding assistants. It is the dataset used to fine-tune code-il-E4B, a 4B on-device model.
The dataset was designed around a thesis: less data, better filtered, beats more data. The… See the full description on the dataset page: https://huggingface.co/datasets/BrainboxAI/code-training-il.TILO.RA_CODER_Dataset
TILO.RA CODER Dataset
Объединённый русско-английский датасет для обучения и поиска по коду.
Формат — пары question / code: вопрос на естественном языке → готовый код-ответ.
Датасет собран для локального ассистента TILO.RA CODER — офлайн-помощника по программированию
Скачать по ссылке
https://github.com/thetemirbolatov/TILO.RA_CODER_Dataset/releases/download/v1.0.0/tilora_knowledge_merged.jsonl
Состав
Источник
Язык
Записей
English coding… See the full description on the dataset page: https://huggingface.co/datasets/thetemirbolatov/TILO.RA_CODER_Dataset.CodeGen-Deep-5K
CodeGen-Deep-5K: Deep Reasoning for Competitive Programming
Part of the CodeGen suite | CodeGen-Diverse-5K (sister dataset)
Dataset Description
CodeGen-Deep-5K is a deep reasoning dataset designed for training code generation models with enhanced problem-solving capabilities. Unlike traditional datasets, this generates multiple distinct solutions for each problem, providing varied reasoning traces and approaches.
Key Statistics
Total samples: 5,000
Unique… See the full description on the dataset page: https://huggingface.co/datasets/Naholav/CodeGen-Deep-5K.Python_GOD_Coder_Omniforge_AI_12k
Python GOD Coder Omniforge AI 12k
Creator: Within Us AI
A 12,000-row mixed-format Python coding dataset designed as a sharpening corpus for building a small but dangerous Python specialist.
This dataset is intentionally focused on the practical behaviors that matter for a modern Python coding model:
implementation with tests
strict code-only instruction following
debugging and repair
refactoring for readability and production readiness
next-token code completion… See the full description on the dataset page: https://huggingface.co/datasets/WithinUsAI/Python_GOD_Coder_Omniforge_AI_12k.worker-safety-qa-eval
Dataset Card for Worker Safety Question and Answer Eval
This dataset contains the worker-safety-qa-eval benchmark. This benchmark is used to evaluate question answering tasks in the domain of worker safety and health.
The focus of the benchmark is to answer queries about worker safety practices and regulations based on laws in Singapore.
For correct answers we refer to the resources from Workplace Safety and Health Council.
Dataset Details
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/codelion/worker-safety-qa-eval.Genesis_AI_Code_50k
Genesis AI Code 50K (Expert)
Developed by: Within Us AI
Expert dataset with diff supervision, failure→reflection→correction, MoE labels, and governance flags.
Splits
train: 49,000
validation: 1,000
Highlights
Tests-as-truth supervision patterns
Diff-first patching
Agentic loops (plan→edit→test→reflect) with bounded budgets
Tool-call trace supervision (where present)
Governance/audit & policy-gate awareness
Storage format
Parquet… See the full description on the dataset page: https://huggingface.co/datasets/WithinUsAI/Genesis_AI_Code_50k.Code-Feedback-decontaminationA decontaminated version of m-a-p/Code-Feedback.
The excluded (28) files are "contaminated" with only two code segments:
simple GCD function: while b: a, b = b, a % b return a
sum_to_n solution: return sum(range(n + 1))
And reformated to sharegpt.
Decontamination is done in the same way as Magicoder (ie., bigcode decontamination process), which uses a substring-match-finding method to find overlaps between a target dataset and the following standard benchmarks:
HumanEval
MBPP… See the full description on the dataset page: https://huggingface.co/datasets/Leon-Leee/Code-Feedback-decontamination.gsm8k-synth
GSM8K-Synth
117,955 grade-school math word problems in the style of GSM8K, LLM-generated (Claude and
Gemini) as training data for small math-word-problem models. Every problem is
round-trip validated (its program re-executes to the stated answer) and
decontaminated against the GSM8K test set — 0% 8-gram overlap.
Built for and used by codelion/sprog-9m, a
9.37M-parameter LLM-free GSM8K solver.
Schema
field
type
description
question
string
the word… See the full description on the dataset page: https://huggingface.co/datasets/codelion/gsm8k-synth.cleand_microsoft_rStar-Coder元データ: https://huggingface.co/datasets/microsoft/rStar-Coder
データ件数: 269,863
平均トークン数: 11674
最大トークン数: 31,184
合計トークン数: 3,150,447,484
ファイル形式: JSONL
ファイルサイズ: 不明
加工内容
synthetic_sftを使用
トークン処理が重たいので、文字数でフィルター
seed_question < 6000
generation < 80000
thinkタグ除去 が中途半端なものを除外
トークナイズ処理（速度向上アップデート
繰り返し除去
code-debugging-sft-50k
Code Debugging SFT (50K)
50,000 ShareGPT-format conversations where the user presents buggy code and the assistant provides root-cause analysis and a corrected solution. Covers Python, JavaScript, Go, TypeScript, and SQL across 14 bug categories.
Motivation
Debugging is one of the most frequent developer tasks — and one of the hardest to train. Most coding datasets focus on writing code from scratch. This dataset trains models to:
Identify the precise root cause… See the full description on the dataset page: https://huggingface.co/datasets/stindardlogic/code-debugging-sft-50k.LiveCodeBench-CodeGenerationGenesis_AI_Code_100k
Genesis AI Code 100K (Frontier)
Developed by: Within Us AI
Frontier dataset with tool-call traces, self-grading, budgets, and audit orientation.
Splits
train: 98,000
validation: 2,000
Highlights
Tests-as-truth supervision patterns
Diff-first patching
Agentic loops (plan→edit→test→reflect) with bounded budgets
Tool-call trace supervision (where present)
Governance/audit & policy-gate awareness
Storage format
Parquet unavailable (No module named… See the full description on the dataset page: https://huggingface.co/datasets/gravermistakes/Genesis_AI_Code_100k.verisci-verified-science-math-code
VeriSci Verified Science Math Code
Verifier-grounded dataset for the Adaption AutoScientist Challenge Part 2, targeting the Science category with secondary Math and Code coverage.
Summary
VeriSci trains models to solve scientific computations, finite-difference PDE updates, numerical ODE steps, unit-checked mechanics, thermodynamics, circuits, chemistry stoichiometry, molarity, unit conversion, vector decomposition, two-point linear modeling, small Python… See the full description on the dataset page: https://huggingface.co/datasets/TheJackBright/verisci-verified-science-math-code.python-codes-25k
License
MIT
This is a Cleaned Python Dataset Covering 25,000 Instructional Tasks
Overview
The dataset has 4 key features (fields): instruction, input, output, and text.It's a rich source for Python codes, tasks, and extends into behavioral aspects.
Dataset Statistics
Total Entries: 24,813
Unique Instructions: 24,580
Unique Inputs: 3,666
Unique Outputs: 24,581
Unique Texts: 24,813
Average Tokens per example: 508
Features… See the full description on the dataset page: https://huggingface.co/datasets/xphillyx/python-codes-25k.