datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
UltraData-SFT-2605
UltraData-SFT-2605
📦 UltraData Collection |
🌐 UltraData |
🤗 MiniCPM5 Series
English |
中文
📚 Introduction
UltraData-SFT-2605 is the full set of core-domain SFT data used in the post-training of MiniCPM5-1B-SFT within the MiniCPM5-1B series, and a key representative of L3 refined data in the UltraData L0-L4 tiered data management framework. It covers math, code, knowledge, instruction following, and other core domains, containing over 15 million Deep… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/UltraData-SFT-2605.UltraData-SFT-Agent-2609
UltraData-SFT-Agent-2609
📦 UltraData Collection |
🌐 UltraData |
🤗 MiniCPM5 Series
English |
中文
📚 Introduction
UltraData-SFT-Agent-2609 is the L3 refined data for Agent instruction-tuning within UltraData's L0-L4 tiered data management framework. Built for the post-training of MiniCPM5-2B, it complements UltraData-SFT-2605 (core-domain SFT) with executable Agent trajectories. The release contains approximately 500,000 samples spanning tool use… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/UltraData-SFT-Agent-2609.UltraFeedback
Introduction
GitHub Repo
UltraRM-13b
UltraCM-13b
UltraFeedback is a large-scale, fine-grained, diverse preference dataset, used for training powerful reward models and critic models. We collect about 64k prompts from diverse resources (including UltraChat, ShareGPT, Evol-Instruct, TruthfulQA, FalseQA, and FLAN). We then use these prompts to query multiple LLMs (see Table for model lists) and generate 4 different responses for each prompt, resulting in a total of 256k samples.
To… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/UltraFeedback.UltraChat
Dataset Card for Dataset Name
Dataset Description
An open-source, large-scale, and multi-round dialogue data powered by Turbo APIs. In consideration of factors such as safeguarding privacy, we do not directly use any data available on the Internet as prompts.
To ensure generation quality, two separate ChatGPT Turbo APIs are adopted in generation, where one plays the role of the user to generate queries and the other generates the response.
We instruct the user model with… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/UltraChat.FormalVerse
MathForm: Scaling Mathematical Autoformalization with Knowledge Retrieval and Verification-Guided Refinement
FormalVerse is a verified Lean 4 autoformalization dataset released with the
paper MathForm: Scaling Mathematical Autoformalization with Knowledge Retrieval
and Verification-Guided Refinement.
Every example is produced by the MathForm pipeline, which retrieves relevant
Mathlib knowledge before generation and refines each candidate using Lean
compiler diagnostics and… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/FormalVerse.MA-ProofBench
MA-ProofBench: A Two-Tiered Evaluation of LLMs for Theorem Proving in Mathematical Analysis
English | 中文
We introduce MA-ProofBench, to the best of our knowledge, the first formal benchmark for evaluating large language models (LLMs) on theorem proving in Mathematical Analysis. It contains 200 rigorously formalized theorem-proving problems in Lean 4 + Mathlib (v4.28.0), split into two difficulty tiers:
Tier
Description
Source
Count
Level I
Undergraduate… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/MA-ProofBench.UltraSafety
UltraSafety
UltraSafety derives 1,000 seed instructions on safety from AdvBench and MaliciousInstruct and bootstraps another 2,000 instructions using Self-Instruct. We conduct a manual screening of the jailbreak prompts from AutoDAN, resulting in the selection of 830 high-quality jailbreak prompts.
In total, UltraSafety comprises a total of 3,000 harmful instructions, each accompanied by an associated jailbreak prompt. Each harmful instruction corresponds to our completions… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/UltraSafety.EVisRAG-Trainpaper: 2510.09733
Dataset Description
This is a VQA Training dataset, collected from ChartQA, InfographicVQA, and MMLongBench-Doc.
Load the dataset
import pandas as pd
import os
import sys
data_name = sys.argv[1]
df = pd.read_parquet(f"data/{data_name}/images.parquet", engine="pyarrow")
output_dir = f"data/{data_name}"
os.makedirs(f"{output_dir}/imgs", exist_ok=True)
for idx, row in df.iterrows():
img_bytes = row['image']['bytes']
output_path = os.path.join(output_dir, row["path"])… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/EVisRAG-Train.llava_zhDensingLaw-ScalingBench
DensingLaw-ScalingBench
This dataset was created to enable a more accurate performance scaling law estimation of Large Language Models (LLMs).
This dataset is released as part of our paper, Densing Law of LLMs.
📜 Paper
💡 Overview
This repository contains the open-source dataset used for calculating conditional loss in our LLM density evaluation framework.
LLM density is defined as the ratio of effective parameter size to actual parameter size, where effective… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/DensingLaw-ScalingBench.EVisRAG-Test-InfoVQADataset Description
This is a VQA dataset about reasoning over the document layout, textual content, graphical elements, and data visualizations from InfographicVQA.
Load the dataset
import pandas as pd
import os
import sys
data_name = sys.argv[1]
df = pd.read_parquet(f"data/{data_name}/images.parquet", engine="pyarrow")
output_dir = f"data/{data_name}"
os.makedirs(f"{output_dir}/imgs", exist_ok=True)
for idx, row in df.iterrows():
img_bytes = row['image']['bytes']
output_path =… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/EVisRAG-Test-InfoVQA.EVisRAG-Test-ChartQADataset Description
This is a VQA dataset about Charts with Visual and Logical Reasoning from ChartQA.
Load the dataset
import pandas as pd
import os
import sys
data_name = sys.argv[1]
df = pd.read_parquet(f"data/{data_name}/images.parquet", engine="pyarrow")
output_dir = f"data/{data_name}"
os.makedirs(f"{output_dir}/imgs", exist_ok=True)
for idx, row in df.iterrows():
img_bytes = row['image']['bytes']
output_path = os.path.join(output_dir, row["path"])
with open(output_path… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/EVisRAG-Test-ChartQA.EVisRAG-Test-ViDoSeekDataset Description
This is a VQA dataset designed to evaluate RAG performance on visually rich documents requiring complex reasoning from ViDoRAG.
Load the dataset
import pandas as pd
import os
import sys
data_name = sys.argv[1]
df = pd.read_parquet(f"data/{data_name}/images.parquet", engine="pyarrow")
output_dir = f"data/{data_name}"
os.makedirs(f"{output_dir}/imgs", exist_ok=True)
foridx, row in df.iterrows():
img_bytes = row['image']['bytes']
output_path = os.path.join(output_dir… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/EVisRAG-Test-ViDoSeek.curated-openbmb-code-math
Curated OpenBMB Code/Math Post-Training Data
English code/math-focused post-training data derived from curated OpenBMB UltraData rows.
Contents
Config
Rows
Schema
Purpose
sft_no_think
25,891
prompt, response
Direct code/math SFT plus necessary technical instruction-following/alignment
sft_think
6,018
prompt, response
Code/math reasoning SFT with <think>...</think> traces
Total rows: 31,909.
Curation
The SFT split keeps English code… See the full description on the dataset page: https://huggingface.co/datasets/josephmayo/curated-openbmb-code-math.EVisRAG-Test-SlideVQADataset Description
This is a VQA dataset for Document Visual Question Answering on Multiple Images from SlideVQA.
Load the dataset
import pandas as pd
import os
import sys
data_name = sys.argv[1]
df = pd.read_parquet(f"data/{data_name}/images.parquet", engine="pyarrow")
output_dir = f"data/{data_name}"
os.makedirs(f"{output_dir}/imgs", exist_ok=True)
for idx, row in df.iterrows():
img_bytes = row['image']['bytes']
output_path = os.path.join(output_dir, row["path"])
with… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/EVisRAG-Test-SlideVQA.EVisRAG-Test-DocVQADataset Description
This is a VQA dataset on Document Images from DocVQA.
Load the dataset
import pandas as pd
import os
import sys
data_name = sys.argv[1]
df = pd.read_parquet(f"data/{data_name}/images.parquet", engine="pyarrow")
output_dir = f"data/{data_name}"
os.makedirs(f"{output_dir}/imgs", exist_ok=True)
for idx, row in df.iterrows():
img_bytes = row['image']['bytes']
output_path = os.path.join(output_dir, row["path"])
with open(output_path, "wb") as f:… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/EVisRAG-Test-DocVQA.openbmb__MiniCPM-S-1B-sft-llama-format-details
Dataset Card for Evaluation run of openbmb/MiniCPM-S-1B-sft-llama-format
Dataset automatically created during the evaluation run of model openbmb/MiniCPM-S-1B-sft-llama-format
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… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/openbmb__MiniCPM-S-1B-sft-llama-format-details.