datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
InstructCoder
Paper |
Code |
Blog
InstructCoder (CodeInstruct): Empowering Language Models to Edit Code
Updates
May 23, 2023: Paper, code and data released.
Overview
InstructCoder is the first dataset designed to adapt LLMs for general code editing. It consists of 114,239 instruction-input-output triplets and covers multiple distinct code editing scenarios, generated by ChatGPT. LLaMA-33B finetuned on InstructCoder performs on par with ChatGPT on a… See the full description on the dataset page: https://huggingface.co/datasets/likaixin/InstructCoder.code_contests_instruct
Dataset Card for "code_contests_instruct"
The deepmind/code_contests dataset formatted as markdown-instruct for text generation training.
There are several different configs. Look at them. Comments:
flesch_reading_ease is computed on the description col via textstat
hq means that python2 (aka PYTHON in language column) is dropped, and keeps only rows with flesch_reading_ease 75 or greater
min-cols drops all cols except language and text
possible values for language are {'CPP'… See the full description on the dataset page: https://huggingface.co/datasets/BEE-spoke-data/code_contests_instruct.ds-coder-instruct-v1
Dataset Card for DS Coder Instruct Dataset
DS Coder is a dataset for instruction fine tuning of language models. It is a specialized dataset focusing only on
data science (eg. plotting, data wrangling, machine learnig models, deep learning, and numerical computations). The dataset contains code examples both in R and Python.
The goal of this dataset is to enable creation of small-scale, specialized language model assistants for data science projects.
Dataset Details… See the full description on the dataset page: https://huggingface.co/datasets/ed001/ds-coder-instruct-v1.code_contest_instruct_cppds-coder-instruct-v2
Dataset Card for DS Coder Instruct v2 Dataset
Changes from v1:
Added WizardLM evol data science samples
Removed R samples from v2
DS Coder is a dataset for instruction fine tuning of language models. It is a specialized dataset focusing only on
data science (eg. plotting, data wrangling, machine learnig models, deep learning, and numerical computations). The dataset contains code examples both in Python (R samples were removed in v2).
The goal of this dataset is to enable… See the full description on the dataset page: https://huggingface.co/datasets/ed001/ds-coder-instruct-v2.Vibe-Coding-InstructCodeMaster-Phi-Instruct
Code Master Phi is a compiled dataset designed for training Phi3 instruct models. This dataset is focused on code-based data and integrates multiple high-quality sources to ensure a robust training foundation. The sources include:
Replete-AI/code_bagel: A diverse collection of code snippets and examples.
nickrosh/Evol-Instruct-Code-80k-v1: A dataset featuring evolved instructions for code generation tasks.
iamtarun/python_code_instructions_18k_alpaca: A compilation of Python code… See the full description on the dataset page: https://huggingface.co/datasets/thesven/CodeMaster-Phi-Instruct.gemma4-code-review-instruct
gemma4-code-review-instruct
197K code review examples — 58K with chain-of-thought <think> reasoning traces.
Built to train models that don't just flag issues, but explain their reasoning before delivering a review. Drop-in ready for SFT with any chat model.
Why This Dataset
Most code review datasets give you diff → comment. This one gives you diff → think → comment for 30% of examples — reasoning traces that show how to analyze a diff before writing the review.… See the full description on the dataset page: https://huggingface.co/datasets/liodon-ai/gemma4-code-review-instruct.python-github-code-instruct-filtered-5k
Dataset Card for "python-github-code-instruct-filtered-5k"
This fine dataset tomekkorbak/python-github-code, filtered by scores greater than 0.03.
Feedback and additional columns generated through OpenAI and Cohere responses.
SynthUI-Code-Instruct-2k-v1Synth UI 🎹
https://www.synthui.design
Dataset details
This dataset aims to provide a diverse collection of NextJS code snippets, along with their corresponding instructions, to facilitate the training of language models for NextJS-related tasks. It is designed to cover a wide range of NextJS functionalities, including UI components, routing, state management, and more.
This dataset consists of:
Note: The dataset is seperated into two main parts:
raw Contains only the… See the full description on the dataset page: https://huggingface.co/datasets/JulianAT/SynthUI-Code-Instruct-2k-v1.Instruct-Python-Code-Turkish
Dataset Card for Instruct-Python-Code-Turkish
Language: Turkish
Dataset Description
The translation was performed using the Google translation model to ensure high-quality, accurate translation.
Dataset Details
Size: ≈5K
Translation tool: Google Translate
Data format: Instruct, Output
CodeChat-Instruct-v1
CodeChat-Instruct-v1
CodeChat-Instruct-v1 is a synthetic coding instruction dataset designed for supervised fine-tuning of language models on programming-related conversations. It includes diverse coding tasks such as code review, code improvement, complexity analysis, edge-case discussion, code explanation, library/API usage, refactoring guidance. The dataset is suitable for training coding assistants, educational programming tutors, and general-purpose code LLMs with strong… See the full description on the dataset page: https://huggingface.co/datasets/kd13/CodeChat-Instruct-v1.code.evol.instruct.wiz.oss_python.jsoncode-instruct-mixed
Description
Filtered/normalised subsets of public code-instruction datasets (Magicoder OSS-Instruct & Evol-Instruct, CodeFeedback, Glaive). The source column attributes each row to its origin; each source retains its upstream licence.
Derived dataset. Source material retains its original per-item licence (see source/repo columns); treat as other / mixed. Provided as-is.
Usage
from datasets import load_dataset
ds = load_dataset("PotatoHD/code-instruct-mixed")
gene-code-generation-instruct
code-generation-instruct v2
Gate-passed instruction data for code-generation — published when 50 fresh examples cleared the quality bar
Kind: synthetic
Domain: code-generation
Records: 96
Created: 2026-06-20T19:02:16+00:00
SHA-256: a3f6a919356ea6d71f365ea93c8ea06cb7dc19f22cad57210dedb1f327caed90
Pipeline: v2.0.0
Filters: {"min_quality": 0.55, "limit": 1000, "source": null, "backend": "llama", "min_judge": 0.7}
Generated by: Qwen3-4B-Instruct-2507-Q4_K_M.gguf (backend:… See the full description on the dataset page: https://huggingface.co/datasets/Gene829/gene-code-generation-instruct.instruct_code_cleaning
SFT code dataset building
Contain a list of tasks useful when building a iniitial dataset source:
reverse_translation
Given a history of conversations, what would the human ask next?
reverse_translation_first_round
Suppose you already have a response, the LLM must predict what question does the human asked
clean_code
Given a code snippet, it determines whether its useful and atomic enough to be use for a response by LLM
gen_code_question
Generates a question given a… See the full description on the dataset page: https://huggingface.co/datasets/syntaxsynth/instruct_code_cleaning.Evol-Instruct-Code-80k-v1
Evol-Instruct-Code-80k-v1
This is a cleansed version of nickrosh/Evol-Instruct-Code-80k-v1
Usage
from datasets import load_dataset
dataset = load_dataset("Sharathhebbar24/Evol-Instruct-Code-80k-v1", split="train")
amenokaku-code-instruct-python-mit-450kunishou/amenokaku-code-instructを以下の条件で絞り込んだものです。
MITライセンス (licence: 'MIT')
python (source: ['gasyori_100_knocks', 'datascience_100_knocks_python', 'bifi', 'python_for_begginers_solve_50_exercises', 'nlp_100_knocks')
source: 'bifi'をランダムに100件に絞り込み
Bangla-Code-Instruct
🐯 Bangla-Code-Instruct: A Comprehensive Bangla Code Instruction Dataset
Accepted at LREC 2026
Nishat Raihan, Antonios Anastasopoulos, Marcos Zampieri
George Mason University, Fairfax, VA, USA
The first large-scale Bangla code instruction dataset (300K examples) for training Code LLMs in Bangla.
⚠️ Note: The dataset will be released after the LREC 2026 conference. Stay tuned!
Overview
Bangla-Code-Instruct is a comprehensive… See the full description on the dataset page: https://huggingface.co/datasets/md-nishat-008/Bangla-Code-Instruct.aihub-korean-education-instruct-sample
Korean Education Instruction Dataset (Sample)
Note: 이 데이터셋은 전체 데이터셋의 샘플 버전입니다 (카테고리별 최대 1,000건).
개요
AI Hub의 한국어 교육 데이터셋 13종을 sLLM 지시학습(Instruction Tuning)용으로 변환한 데이터셋입니다.
초등학교부터 고등학교까지의 다양한 교육 콘텐츠를 포함합니다.
데이터셋 통계
카테고리
데이터 수
math (수학)
1000
korean (국어)
1000
writing (글쓰기)
1000
career (진로)
1000
curriculum (교과)
1000
tutor (튜터링)
1000
총계
6000
사용 방법
from datasets import load_dataset
# 데이터셋 로드
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/neuralfoundry-coder/aihub-korean-education-instruct-sample.maplestory-worlds-creator-code-instruct
MapleStory Worlds Creator Code (mlua)
Instruction-style code dataset for mlua, the scripting language of
MapleStory Worlds. Built from the
official Creator Center example code: each example is grounded in its source
document and paired with a natural-language task, reasoning, a self-contained
explanation, and commented mlua code. Intended to teach LLMs to write mlua game
scripts.
The example code is preserved from the official source (a code-preservation check
rejects any record… See the full description on the dataset page: https://huggingface.co/datasets/msw-ai-tf/maplestory-worlds-creator-code-instruct.Evol-Instruct-Code-80k-v1-koTranslated nickrosh/Evol-Instruct-Code-80k-v1 using nayohan/llama3-instrucTrans-enko-8b.
This is a raw translation dataset. It needs to be filtered for repetitions generated by the model.
amenokaku-code-instruct-python-mitkunishou/amenokaku-code-instructを以下の条件で絞り込んだものです。
MITライセンス (licence: 'MIT')
python (source: ['gasyori_100_knocks', 'datascience_100_knocks_python', 'bifi', 'python_for_begginers_solve_50_exercises', 'nlp_100_knocks')
WizardLM_evol_instruct_V2_only_codefiltered from (WizardLM/WizardLM_evol_instruct_V2_196k)[https://huggingface.co/datasets/WizardLM/WizardLM_evol_instruct_V2_196k] using "```"
opc_regen_Qwen3-Coder-30B-A3B-Instruct
OPC Regenerated Dataset (Qwen3-Coder-30B-A3B-Instruct)
This dataset is a regenerated version of the OPC training dataset, where assistant responses have been regenerated using Qwen3-Coder-30B-A3B-Instruct as the target model.
Purpose
Regenerating training data with the target model helps better align the draft model with the target model's output distribution, improving acceptance rates and overall speculative decoding performance in SpecForge.
Dataset Statistics… See the full description on the dataset page: https://huggingface.co/datasets/JinnP/opc_regen_Qwen3-Coder-30B-A3B-Instruct.CodeDebug-Instruct-v2-Reasoning
CodeDebug-Instruct-v2-Reasoning
CodeDebug-Instruct-v2-Reasoning is a synthetic debugging instruction dataset designed for supervised fine-tuning of language models with enhanced code reasoning capabilities. It covers diverse debugging scenarios including import errors, syntax errors, runtime errors, performance bottlenecks, time limit exceeded (TLE) issues, and general code repair tasks with step-by-step reasoning and corrected solutions. The dataset is suitable for training… See the full description on the dataset page: https://huggingface.co/datasets/kd13/CodeDebug-Instruct-v2-Reasoning.evol-instruct-code-cot-80k
evol-instruct-code-cot-80k
COT distilled dataset with 63,007 examples.
Source
Base: nickrosh/Evol-Instruct-Code-80k-v1
Model: Mistral-7B-Instruct-v0.2-AWQ
Format
instruction: Task
thinking: <think>...</think> reasoning
response: Solution
en-appealcourt-coded-instruct_v02
Dataset Card for JuDDGES/en-appealcourt-coded-instruct_v02
Dataset Summary
The raw data was acquired from publicly available judgments from the Court of Appeal (Criminal Division) (link) of England and Wales. These judgments are available in HTML format on the national archives website for online reading. They can be downloaded as XML or PDF files under the crown copyright license (link): and the Open Government license (see Appendix 6 in the paper). These licenses… See the full description on the dataset page: https://huggingface.co/datasets/JuDDGES/en-appealcourt-coded-instruct_v02.deepseek-ai-deepseek-coder-v2-lite-instruct__llm-quality-persona-consistency-mini__019e3b6fdda4
deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct on llm.quality.persona-consistency-mini (NVIDIA H100 80GB HBM3)
Back to leaderboard
Headline metrics
Metric
Value
Unit
N Samples
5
N Ok
5
Ok Rate
1
Persona Consistency Mean
0.88
Accuracy
0.88
Persona Consistency P50
0.8
Persona Consistency P95
1
Accuracy P05
0.8
Accuracy P50
0.8
Accuracy P95
1
Drift Rate
0.6
Mean Drift Turn
2.6667
TTFT P50
57.5324
ms
Total P50 Ms
2672.8064… See the full description on the dataset page: https://huggingface.co/datasets/Yobitel/deepseek-ai-deepseek-coder-v2-lite-instruct__llm-quality-persona-consistency-mini__019e3b6fdda4.code-instruct-ka
code-instruct-ka
Georgian code instruction dataset for training coding capabilities.
Dataset Summary
Property
Value
Examples
61,288
Size
937 MB
Language
Georgian
Data Fields
id: Unique identifier
conversation: Conversation with code instructions
Usage
from datasets import load_dataset
ds = load_dataset("tbilisi-ai-lab/code-instruct-ka")
Citation
@misc{tbilisi2025codeinstructka,
title = {code-instruct-ka:… See the full description on the dataset page: https://huggingface.co/datasets/tbilisi-ai-lab/code-instruct-ka.