datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
thinking_droid_lerobot_output_qwen3vlMMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking
MMFineReason-Full-2.3M
The Complete Pre-Selection Dataset — Before Quality Filtering
📖 Overview
MMFineReason-Full-2.3M is the complete pre-selection dataset containing 2.3M samples and 8.8B solution tokens, generated through our reasoning distillation pipeline before the data selection stage. This dataset includes all samples that passed basic template and length validation, but have not undergone correctness verification filtering.
🎯 Key Characteristics… See the full description on the dataset page: https://huggingface.co/datasets/OpenDataArena/MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking.thinking_furniture_bench_dataset_lerobot_output_qwen3vlFineReason-1.8M-Qwen3-VL-235B-Thinking
MMFineReason
Closing the Multimodal Reasoning Gap via Open Data-Centric Methods
Average score across mathematical reasoning and multimodal understanding benchmarks.
📖 Overview
MMFineReason is a large-scale, high-quality multimodal reasoning dataset comprising 1.8M samples and 5.1B solution tokens, featuring detailed reasoning annotations distilled from Qwen3-VL-235B-A22B-Thinking.
🎯 Key Highlights
1.8M High-Quality Samples with 5.1B Solution Tokens… See the full description on the dataset page: https://huggingface.co/datasets/NarsAI/FineReason-1.8M-Qwen3-VL-235B-Thinking.MMFineReason-1.8M-Qwen3-VL-235B-Thinking
MMFineReason
Closing the Multimodal Reasoning Gap via Open Data-Centric Methods
Average score across mathematical reasoning and multimodal understanding benchmarks.
📖 Overview
MMFineReason is a large-scale, high-quality multimodal reasoning dataset comprising 1.8M samples and 5.1B solution tokens, featuring detailed reasoning annotations distilled from Qwen3-VL-235B-A22B-Thinking.
🎯 Key Highlights
1.8M High-Quality Samples with 5.1B Solution Tokens… See the full description on the dataset page: https://huggingface.co/datasets/OpenDataArena/MMFineReason-1.8M-Qwen3-VL-235B-Thinking.MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking
MMFineReason-Full-2.3M
The Complete Pre-Selection Dataset — Before Quality Filtering
📖 Overview
MMFineReason-Full-2.3M is the complete pre-selection dataset containing 2.3M samples and 8.8B solution tokens, generated through our reasoning distillation pipeline before the data selection stage. This dataset includes all samples that passed basic template and length validation, but have not undergone correctness verification filtering.
🎯 Key Characteristics… See the full description on the dataset page: https://huggingface.co/datasets/ericktwo/MMFineReason-Full-2.3M-Qwen3-VL-235B-Thinking.MMFineReason-1.8M-Qwen3-VL-235B-Thinking
MMFineReason
Closing the Multimodal Reasoning Gap via Open Data-Centric Methods
Average score across mathematical reasoning and multimodal understanding benchmarks.
📖 Overview
MMFineReason is a large-scale, high-quality multimodal reasoning dataset comprising 1.8M samples and 5.1B solution tokens, featuring detailed reasoning annotations distilled from Qwen3-VL-235B-A22B-Thinking.
🎯 Key Highlights
1.8M High-Quality Samples with 5.1B Solution Tokens… See the full description on the dataset page: https://huggingface.co/datasets/Sandeepthakur/MMFineReason-1.8M-Qwen3-VL-235B-Thinking.VLAA-Thinking
SFT or RL? An Early Investigation into Training R1-Like Reasoning Large Vision-Language Models
🌐 Project Page
• 📄 Arxiv
• 💻 Code
🤗 VLAA-Thinker Family
• 🤔 VLAA-Thinking Dataset
🤗 VLAA-Thinker-Qwen2.5-3B
• 🤗 VLAA-Thinker-Qwen2.5-7B
Both VLAA-Thinker-Qwen2.5-3B and VLAA-Thinker-Qwen2.5-7Bachieve SOTA performance on OpenCompass Multimodal Reasoning Leaderboard as of April 7th, 2025.
Contents
Quick Start 🚀… See the full description on the dataset page: https://huggingface.co/datasets/UCSC-VLAA/VLAA-Thinking.thinking_fmb_dataset_lerobot_output_qwen3vlmhlc-training-qwen3vl-qwen3_vl_2b_thinking_hard_mixed_sources_120k
Multi Head Latent Control Training Data - Qwen3-VL 2B Thinking hard Mixed Sources 120k
Dataset Description
This repository contains verified training data for the Multi Head Latent Control paper release. It is part of the Multi Head Latent Control training data Hugging Face collection.
Paper
https://arxiv.org/abs/2607.14277
Code
https://github.com/Amirhosein-gh98/Multi-Head-Latent-Control
Dataset Summary
Field… See the full description on the dataset page: https://huggingface.co/datasets/AmirhoseinGH/mhlc-training-qwen3vl-qwen3_vl_2b_thinking_hard_mixed_sources_120k.MMFineReason-1.8M-Qwen3-VL-235B-Thinking
MMFineReason
Closing the Multimodal Reasoning Gap via Open Data-Centric Methods
Average score across mathematical reasoning and multimodal understanding benchmarks.
📖 Overview
MMFineReason is a large-scale, high-quality multimodal reasoning dataset comprising 1.8M samples and 5.1B solution tokens, featuring detailed reasoning annotations distilled from Qwen3-VL-235B-A22B-Thinking.
🎯 Key Highlights
1.8M High-Quality Samples with 5.1B Solution Tokens… See the full description on the dataset page: https://huggingface.co/datasets/dans25275/MMFineReason-1.8M-Qwen3-VL-235B-Thinking.MMFineReason-SFT-586K-Qwen3-VL-235B-Thinking
MMFineReason-SFT-586K
The Hardest 33% — Less Data, More Reasoning
📖 Overview
MMFineReason-SFT-586K is a difficulty-filtered subset of MMFineReason-1.8M, containing the hardest 33% of samples where Qwen3-VL-4B-Thinking do not consistently succeed. (pass rate ≠ 1).
Specifically, this subset removes all easy samples (pass rate = 1) under Qwen3-VL-4B-Thinking, retaining only instances that require non-trivial multimodal reasoning.
🎯 Key Highlights
586K… See the full description on the dataset page: https://huggingface.co/datasets/OpenDataArena/MMFineReason-SFT-586K-Qwen3-VL-235B-Thinking.finevision-mini-thinking
FineVision-mini Thinking
FineVision-mini is a slice I made of HuggingFaceM4/FineVision:
169 of its image subsets, 101,321 rows (~40 GB) out of FineVision's 24.2M rows / 4.65 TB (about 0.4% of
the rows, 0.9% of the bytes), sampled with a fixed seed. The 16 text-only subsets were left out.
This dataset is that slice, fully translated and augmented with reasoning, published in increments:
each batch processes more rows of FineVision-mini and is appended here, until the whole slice… See the full description on the dataset page: https://huggingface.co/datasets/olob0/finevision-mini-thinking.MMFineReason-SFT-123K-Qwen3-VL-235B-Thinking
MMFineReason-SFT-123K
The Hardest 7% — Less Data, More Reasoning
📖 Overview
MMFineReason-SFT-123K is a difficulty-filtered subset of MMFineReason-1.8M, containing only the hardest 7% of samples where Qwen3-VL-4B-Thinking consistently fails (pass rate = 0).
🎯 Key Highlights
123K Challenging Samples: Only instances where a 4B thinking model fails all 4 inference attemptsEfficient Training: Comparable performance to full 1.8M dataset with only 7% of… See the full description on the dataset page: https://huggingface.co/datasets/OpenDataArena/MMFineReason-SFT-123K-Qwen3-VL-235B-Thinking.mhlc-training-qwen3vl-qwen3_vl_4b_thinking_hard_mixed_sources_120k
Multi Head Latent Control Training Data - Qwen3-VL 4B Thinking hard Mixed Sources 120k
Dataset Description
This repository contains verified training data for the Multi Head Latent Control paper release. It is part of the Multi Head Latent Control training data Hugging Face collection.
Paper
https://arxiv.org/abs/2607.14277
Code
https://github.com/Amirhosein-gh98/Multi-Head-Latent-Control
Dataset Summary
Field… See the full description on the dataset page: https://huggingface.co/datasets/AmirhoseinGH/mhlc-training-qwen3vl-qwen3_vl_4b_thinking_hard_mixed_sources_120k.shotpath-qwen3vl-thinking-cot-20260714PP
SteelBench: A Diagnostic Benchmark for Vision-Language Models in Industrial Safety Monitoring
SteelBench is a diagnostic benchmark of densely annotated CCTV clips from an
operating integrated steel plant. It is designed to evaluate vision-language
models (VLMs) on real-world industrial action recognition, PPE assessment,
and safety-violation detection — under naturally occurring degradation
(dust, glare, steam, low light), at distances and crowdedness levels that
curated… See the full description on the dataset page: https://huggingface.co/datasets/ThinkingHub/PP.medraN-thinking-1024thinking_berkeley_autolab_ur5_lerobot_output_qwen3vlThinking-in-Video-DataThinking in Video: Can Video Generators Really Reason About the Real World?
This repository contains the official implementation of Causal-Generative Dual-Judge (CGDJ) for auditing world-model consistency of video generative models — the official codebase of the Thinking in Video paradigm.
🌟 Overview
Thinking in Video is a reasoning paradigm in which a video generative model is used not merely to synthesize pixels, but to simulate, predict, and verify causal… See the full description on the dataset page: https://huggingface.co/datasets/BRZ911/Thinking-in-Video-Data.pusht_96_int1_visual_nomarker_allstep_thinking_trickiness_cot
PushT int1 Visual Nomarker All-Step Thinking Trickiness COT
This dataset is derived from successful PushT visual-nomarker trajectories in novastar112/pusht_96_int1_visual_nomarker.
Each row contains one full successful trajectory from the first move through the final stop action.
Main files:
training/pusht_allstep_thinking_cot.jsonl.gz: 500,000 train rows.
testing/pusht_allstep_thinking_cot.jsonl.gz: 200 test rows.
Message format:
Each user turn is the PushT prompt text plus one… See the full description on the dataset page: https://huggingface.co/datasets/novastar112/pusht_96_int1_visual_nomarker_allstep_thinking_trickiness_cot.synthetic-self-correction-and-thinking-samples
Self Correction and Thinking
A seed library for training language models to reason with self-correction.
Teaches three reasoning behaviors -- catching your own errors, verifying correct answers, and rejecting false doubts -- across four domains, three difficulty tiers, and three reasoning modes. Also includes multi-turn user-correction conversations where the user actively corrects or challenges the assistant.
The structure at a glance
graph TB… See the full description on the dataset page: https://huggingface.co/datasets/sbussiso/synthetic-self-correction-and-thinking-samples.nycc-thinkingbghira_pseudo-camera-10k-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT
bghira_pseudo-camera-10k-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT
bghira/pseudo-camera-10k with responses/captions generated with gemini-2.0-flash-thinking-exp-1219.
The format should be similar to that of liuhaotian/LLaVA-Instruct-150K.
Images can be found in the images.zip folder. The zip also contains .txt captions for ease of use in non-VQA tasks.
Generation Details
If BlockedPromptException, StopCandidateException, or InvalidArgument was returned, the… See the full description on the dataset page: https://huggingface.co/datasets/PJMixers-Images/bghira_pseudo-camera-10k-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT.thinking_cmu_play_fusion_lerobot_output_qwen3vlHandpicked-Images-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT
Handpicked-Images-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT
Some random images with responses/captions generated with gemini-2.0-flash-thinking-exp-1219.
The format should be similar to that of liuhaotian/LLaVA-Instruct-150K.
Images can be found in the images.zip folder. The zip also contains .txt captions for ease of use in non-VQA tasks.
Generation Details
If BlockedPromptException, StopCandidateException, or InvalidArgument was returned, the sample was… See the full description on the dataset page: https://huggingface.co/datasets/PJMixers-Images/Handpicked-Images-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT.mathvision-thinkingr_portraitphotography-gemini-2.0-flash-thinking-exp-1219-CustomShareGPTGeneration-Reviewvlaa-thinking-grpo
VLAA-Thinking-SFT-126K
Large-scale vision-language dataset with 126K instruction-following samples featuring chain-of-thought reasoning
Dataset Description
This dataset contains vision-language samples with instruction-following conversations. Each sample includes:
image: PIL Image object
question: Question or instruction text
answer or gt: Response with thinking process (SFT dataset) or ground truth answer (GRPO dataset)
caption: Image caption (may be empty for some… See the full description on the dataset page: https://huggingface.co/datasets/penfever/vlaa-thinking-grpo.