datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
comma_v0.1_training_dataset
Comma v0.1 dataset
This repository contains the dataset used to train Comma v0.1-1T and Comma v0.1-2T.
It is a slightly modified and consolidated version of the Common Pile v0.1 "filtered" data.
If you are looknig for the raw Common Pile v0.1 data, please see this collection.
You can learn more about Common Pile in our paper.
Mixing rates and token counts
The Comma v0.1 models were trained in two stages, a "main" stage and a "cooldown" stage.
During each stage, we… See the full description on the dataset page: https://huggingface.co/datasets/common-pile/comma_v0.1_training_dataset.firefly-train-1.1M本数据应用于项目：Firefly（流萤）: 中文对话式大语言模型 ，训练后得到的模型firefly-1b4
如果您觉得此数据集对您有帮助，请like此数据集并在Github项目中star我们。
我们收集了23个常见的中文数据集，对于每个任务，由人工书写若干种指令模板，保证数据的高质量与丰富度，数据量为115万 。数据分布如下图所示：
每条数据的格式如下，包含任务类型、输入、目标输出：
{
"kind": "ClassicalChinese",
"input": "将下面句子翻译成现代文：\n石中央又生一树，高百余尺，条干偃阴为五色，翠叶如盘，花径尺余，色深碧，蕊深红，异香成烟，著物霏霏。",
"target": "大石的中央长着一棵树，一百多尺高，枝干是彩色的，树叶有盘子那样大，花的直径有一尺宽，花瓣深蓝色，花中飘出奇异的香气笼罩着周围，如烟似雾。"
}
训练数据集的token长度分布如下图所示，绝大部分数据的长度都小于600：
SWE-Fixer-Train-110K
SWE-Fixer: Training Open-Source LLMs for Effective and Efficient GitHub Issue Resolution
📃 Paper |
🚀 GitHub
SWE-Fixer is a simple yet effective solution for addressing real-world GitHub issues by training open-source LLMs. It features a streamlined retrieve-then-edit pipeline with two core components: a code file retriever and a code editor.
This repo holds the data SWE-Fixer-Train-110K we curated for SWE-Fixer training.
For more information, please visit our project page.… See the full description on the dataset page: https://huggingface.co/datasets/internlm/SWE-Fixer-Train-110K.LEMAS-Dataset-train
Overview
This dataset is part of LEMAS-Project (lemas-project.github.io/LEMAS-Project).
It contains a large-scale training set (150k+ hours) and a curated evaluation set
(500 utterances per language) covering 10 languages, all with word-level alignment.
Fields
key: unique utterance identifier; the first two characters indicate the language ID
audio: relative path to the MP3 audio file (in the eval set, this key is renamed to "file_name" for compatibility with the viewer)… See the full description on the dataset page: https://huggingface.co/datasets/LEMAS-Project/LEMAS-Dataset-train.Nemotron-Image-Training-v3
Nemotron Image Training v3
Versions
Date
Commit
Changes
2026-04-28
HEAD
Initial commit.
Dataset Description
Nemotron Image Training v3 is a collection of image-centric multimodal training data for vision–language models. Similar to Nemotron-VLM-Dataset v2, it was curated as a large-scale, multi-subdataset release where each subset ships a standardized conversation JSONL alongside a dataset card describing sources, licensing, and media layout.… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-Image-Training-v3.Llama-Nemotron-Post-Training-Dataset
Llama-Nemotron-Post-Training-Dataset-v1.1 Release
Update [4/8/2025]:
v1.1: We are releasing an additional 2.2M Math and 500K Code Reasoning Data in support of our release of Llama-3.1-Nemotron-Ultra-253B-v1. 🎉
Data Overview
This dataset is a compilation of SFT and RL data that supports improvements of math, code, general reasoning, and instruction following capabilities of the original Llama instruct model, in support of NVIDIA’s release of… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Llama-Nemotron-Post-Training-Dataset.train_1M_CN
内容
包含约100万条由BELLE项目生成的中文指令数据。
样例
{
"instruction": "给定一个文字输入，将其中的所有数字加1。\n“明天的会议在9点开始，记得准时到达。”\n",
"input": "",
"output": "“明天的会议在10点开始，记得准时到达。”"
}
字段：
instruction: 指令
input: 输入（本数据集均为空）
output: 输出
使用限制
仅允许将此数据集及使用此数据集生成的衍生物用于研究目的，不得用于商业，以及其他会对社会带来危害的用途。
本数据集不代表任何一方的立场、利益或想法，无关任何团体的任何类型的主张。因使用本数据集带来的任何损害、纠纷，本项目不承担任何责任。
SII_self_evovling_02_training_datasettrain_0.5M_CN
内容
包含约50万条由BELLE项目生成的中文指令数据。
样例
{
"instruction": "给定一个文字输入，将其中的所有数字加1。\n“明天的会议在9点开始，记得准时到达。”\n",
"input": "",
"output": "“明天的会议在10点开始，记得准时到达。”"
}
字段：
instruction: 指令
input: 输入（本数据集均为空）
output: 输出
使用限制
仅允许将此数据集及使用此数据集生成的衍生物用于研究目的，不得用于商业，以及其他会对社会带来危害的用途。
本数据集不代表任何一方的立场、利益或想法，无关任何团体的任何类型的主张。因使用本数据集带来的任何损害、纠纷，本项目不承担任何责任。
train_3.5M_CN
内容
包含约350万条由BELLE项目生成的中文指令数据。
样例
{
"id":"66623028",
"conversations":[
{
"from":"human",
"value":"归纳介绍三种不同类型的电影题材"
},
{
"from":"assistant",
"value":"1. 动作片：动作片通常强调动作和战斗场面，故事情节紧凑，充满了紧张和兴奋的氛围。主角一般是具有超凡实力或特殊技能的英雄人物，往往会与反派角色进行一系列激烈的格斗，追逐和爆破等酷炫场面。例如《速度与激情》和《复仇者联盟》等。\n2. 爱情片：爱情片是以爱情为主题的电影，着重呈现角色之间的感情纠葛，展现了爱情的美好和悲壮。情节与角色之间的发展令人深思，增加了观众的感性体验。这种电影的故事情节主要以人类与人类之间的情感变化为主题。例如《泰坦尼克号》和《让子弹飞》等。\n3.… See the full description on the dataset page: https://huggingface.co/datasets/BelleGroup/train_3.5M_CN.GPT-Training-Datacodeparrot-clean-train
CodeParrot 🦜 Dataset Cleaned (train)
Train split of CodeParrot 🦜 Dataset Cleaned.
Dataset structure
DatasetDict({
train: Dataset({
features: ['repo_name', 'path', 'copies', 'size', 'content', 'license', 'hash', 'line_mean', 'line_max', 'alpha_frac', 'autogenerated'],
num_rows: 5300000
})
})
Inter-Edit-Train
Inter-Edit-Train
Inter-Edit-Train is the official large-scale training set released for the CVPR 2026 paper Inter-Edit: First Benchmark for Interactive Instruction-Based Image Editing.
This dataset is designed for the Interactive Instruction-based Image Editing (I^3E) task, where a model performs localized image edits from a concise textual instruction together with imprecise spatial guidance.
Highlights
1,099,964 image editing pairs
610,186 unique source images
Four… See the full description on the dataset page: https://huggingface.co/datasets/a1557811266/Inter-Edit-Train.train_2M_CN
内容
包含约200万条由BELLE项目生成的中文指令数据。
样例
{
"instruction": "将以下三个句子组合成一个有意义的段落。\n狗是人类最好的朋友。它们非常聪明，可以进行各种活动。如果你喜欢散步，狗可以成为你一起散步的伙伴。",
"input": "",
"output": "狗是人类最好的朋友，它们非常聪明，可以进行各种活动。如果你喜欢散步，狗可以成为你一起散步的伙伴。出门散步是一种良好的锻炼方式，而有狗的陪伴会让散步变得更有趣，并且有狗在身边也能给你带来安全感。所以，拥有一只狗作为你的伙伴，可以帮助你变得更加积极主动和健康。"
}
字段：
instruction: 指令
input: 输入（本数据集均为空）
output: 输出
使用限制
仅允许将此数据集及使用此数据集生成的衍生物用于研究目的，不得用于商业，以及其他会对社会带来危害的用途。… See the full description on the dataset page: https://huggingface.co/datasets/BelleGroup/train_2M_CN.VideoKR-Train
VideoKR-Train
📄 ArXiv
｜ 💻 Code
｜ 🤗 Collection
About
This repository contains the VideoKR training data presented in VideoKR: Towards Knowledge- and Reasoning-Intensive Video Understanding (ICML 2026 Spotlight).
VideoKR is the first large-scale training corpus specifically designed for knowledge- and reasoning-intensive video understanding. It contains 315K video reasoning examples over 145K newly collected, CC-licensed… See the full description on the dataset page: https://huggingface.co/datasets/minuzero/VideoKR-Train.lingshu_training_data_medical_domain
Website
🤖 7B Model
🤖 8B Model based on InternVL3
🤖 32B Model
MedEvalKit
Technical Report
Lingshu MCP
Lingshu Medical MLLM Training Data (Medical Domain)
This dataset contains the medical-domain training data used in the multi-stage training of the Lingshu Medical Multimodal Large Language Model (MLLM). General-domain data has been removed; only medical data is included.
The training… See the full description on the dataset page: https://huggingface.co/datasets/lingshu-medical-mllm/lingshu_training_data_medical_domain.OLA-Embed-Trainingtemp-decoder-train-tokenizednemotron-post-training-v2-qwen-3.5-9b-regen
Dataset Card for Nemotron Post Training v2 Qwen 3.5 9B Regen
Regenerated responses from nvidia/Nemotron-Post-Training-Dataset-v2 dataset using Qwen3.5 9B model.
Parameter
Value
Max Tokens
4096
Temperature
1.0
Top-k
20
Top-p
0.95
Repetition Penalty
1.5
Dataset consists only the english samples from the Nemotron Post Training Dataset. 85% of the chat prompts have reasoning enabled, every other category has reasoning disabled.
Category
Value
math… See the full description on the dataset page: https://huggingface.co/datasets/Dogacel/nemotron-post-training-v2-qwen-3.5-9b-regen.codeparrot-ds-trainresume-training-dataset
Resume Training Dataset
Dataset Summary
This dataset contains 22,855 curated resume samples designed for training AI models on resume analysis, generation, and career development tasks. Each entry includes structured conversations between users seeking resume help and AI assistants providing feedback, making it ideal for training models to understand professional writing patterns, critique resumes, and suggest improvements.
Dataset Details
Supported… See the full description on the dataset page: https://huggingface.co/datasets/MikePfunk28/resume-training-dataset.Nemotron-Image-Training-v3
Nemotron Image Training v3
Versions
Date
Commit
Changes
2026-04-28
HEAD
Initial commit.
Dataset Description
Nemotron Image Training v3 is a collection of image-centric multimodal training data for vision–language models. Similar to Nemotron-VLM-Dataset v2, it was curated as a large-scale, multi-subdataset release where each subset ships a standardized conversation JSONL alongside a dataset card describing sources, licensing, and media layout.… See the full description on the dataset page: https://huggingface.co/datasets/liyah1616/Nemotron-Image-Training-v3.Nemotron-RL-Ultra-Training-Blends
Dataset Description:
This dataset provides Reinforcement Learning (RL) and Multi-teacher On-Policy Distillation (MOPD) training-data blends used by the public Nemotron-3-Ultra post-training recipe. The blends are consumed by the NeMo RL training recipes through the NeMo Gym agent framework, in which each prompt is paired with an agent/environment that returns a verifiable or judge-based reward. Each subset is a separate blend; see the recipe for how the blends are used.
The… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-RL-Ultra-Training-Blends.sat-vl-sft-training-ready-v1
Dataset Summary
NuTonic/sat-bbox-metadata-sft-v1 is a metadata-first, procedural VLM SFT dataset built from an existing “sat-bbox” style dataset tree (Sentinel‑2 chips + per-tile JSON metadata sidecars, optionally paired Mapbox stills).
The goal is to create high-signal, production-shaped supervision for multimodal chat models:
Captioning for satellite chips
Grounding (bounding boxes in normalized coordinates) for land-cover regions
Class-focused captions and absence checks for… See the full description on the dataset page: https://huggingface.co/datasets/NuTonic/sat-vl-sft-training-ready-v1.personal-trainer-ausbildung-ki-datensatz
SNFA Personal Trainer Ausbildung KI-Datensatz
Ein deutschsprachiger Wissensdatensatz der SNF Academy zu Personal Training, Fitnessausbildung, Berufspraxis, Coaching, Selbstständigkeit und regionalen Angeboten in der Schweiz.
Inhalt
Die Datei snfa_personal_trainer_dataset.jsonl enthält thematisch abgegrenzte Abschnitte aus den Dokumenten dieses Repositorys. Jeder Datensatz besitzt eine eindeutige ID sowie Angaben zu Titel, Abschnitt, Inhalt, Kategorie, Quelldatei… See the full description on the dataset page: https://huggingface.co/datasets/snfacademy/personal-trainer-ausbildung-ki-datensatz.training_setbird23-train-filtered
BIRD-SQL Train (Filtered)
A high-quality subset of the original BIRD train split for text-to-SQL finetuning.
Overview
Over the past year the community has shared many observations about data quality in BIRD. We performed a rigorous data quality check process to retain examples that are consistent with schema and faithfully answer the question. The resulting set keeps 6,601 instances out of 9,428 (≈70%), and serves as a drop-in replacement for training.
Original Train: 9… See the full description on the dataset page: https://huggingface.co/datasets/birdsql/bird23-train-filtered.c4-en-html-with-training_metadata_alltrain-of-thought
Train of Thought Dataset
Overview
This dataset readapts agentlans/think-more
into the Alpaca-style instruction tuning format for training language models in direct answering and chain-of-thought reasoning.
Dataset Structure
Each original example was randomly assigned to be thinking on or off:
Thinking off: Outputs only the final answer.
Thinking on:
Outputs a chain-of-thought (CoT) reasoning process wrapped in <think>...</think>, followed by the final answer… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/train-of-thought.rlbenchfail_train_dataset
Guardian: RLBench-Fail Dataset
This dataset is part of the Guardian project: Detecting Robotic Planning and Execution Errors with Vision-Language Models. It contains annotated robotic manipulation failure data generated in the RLBench simulator for training and evaluating Vision-Language Models (VLMs) on failure detection tasks.
Failures are produced by an automated pipeline that procedurally perturbs successful scripted trajectories in simulation, generating diverse planning… See the full description on the dataset page: https://huggingface.co/datasets/paulpacaud/rlbenchfail_train_dataset.