Finesse Benchmark Database

  Overview

finesse-benchmark-database is a data generation factory for atomic probes in the Finesse benchmark. It generates probes_atomic.jsonl files from Wikimedia Wikipedia datasets, leveraging Hugging Face's datasets library, tokenizers from transformers, and optional PyTorch support. This tool is designed to create high-quality, language-specific probe datasets for benchmarking fine-grained understanding in NLP tasks.… See the full description on the dataset page: https://huggingface.co/datasets/enzoescipy/finesse-benchmark-database.

Dataset enabling organizations to benchmark their data literacy capability globally.

The NIST Computational Chemistry Comparison and Benchmark Database is a collection of experimental and ab initio thermochemical properties for a selected set of gas-phase molecules. The goals are to provide a benchmark set of experimental data for the evaluation of ab initio computational methods and allow the comparison between different ab initio computational methods for the prediction of gas-phase thermochemical properties. The data files linked to this record are a subset of the experimental data present in the CCCBDB.

Dataset Description

This dataset provides a benchmark for automatic data product creation. The task is framed as follows: given a natural language data product request and a corpus of text and tables, the objective is to identify the relevant tables and text documents that should be included in the resulting data product which would useful to the given data product request. The benchmark brings together three variants: HybridQA, TAT-QA, and ConvFinQA, each consisting of:

A corpus… See the full description on the dataset page: https://huggingface.co/datasets/ibm-research/data-product-benchmark.

LLMSQL Benchmark

This benchmark is designed to evaluate text-to-SQL models. For usage of this benchmark see https://github.com/LLMSQL/llmsql-benchmark.

If you want to use this benchmark from huggingface, please see: https://huggingface.co/llmsql-bench.

Arxiv Article: https://arxiv.org/abs/2510.02350

Files

• tables.jsonl — Database table metadata
• questions.jsonl — All available questions
• train_questions.jsonl, val_questions.jsonl, test_questions.jsonl — Data splits for finetuning, see https://github.com/LLMSQL/llmsql-benchmark
• sqlite_tables.db — sqlite db with tables from tables.jsonl, created with the help of create_db_sql.
• create_db.sql — SQL script that creates the database sqlite_tables.db.

Citation

If you use this benchmark, please cite:

@inproceedings{llmsql_bench,
 title={LLMSQL: Upgrading WikiSQL for the LLM Era of Text-to-SQLels},
 author={Pihulski, Dzmitry and Charchut, Karol and Novogrodskaia, Viktoria and Koco{'n}, Jan},
 booktitle={2025 IEEE International Conference on Data Mining Workshops (ICDMW)},
 year={2025},
 organization={IEEE}
}

The publication provides and describes a clean and expert-curated benchmark dataset to be used for machine-learning-based research in ecotoxicology. The package contains several data files associated with the challenges we propose and some supplementary data files to aid in the interpretation of results.

This datasets generated with the LDBC SNB Data generator.

https://github.com/ldbc/ldbc_snb_datagen

It corresponds to Scale Factors 1 and 3. They are used in the following paper:

An early look at the LDBC social network benchmark's business intelligence workload

10.1145/3210259.3210268

Results of the IGUANA Benchmark in 2015/16 for the truncated DBpedia dataset. The dataset is 50% of the initial 100% dataset.

Explore why AI excels at complex math but struggles with SQL queries, with benchmark data showing a 60% accuracy ceiling in database operations across leading models.

A total of 12 software defect data sets from NASA were used in this study, where five data sets (part I) including CM1, JM1, KC1, KC2, and PC1 are obtained from PROMISE software engineering repository (http://promise.site.uottawa.ca/SERepository/), the other seven data sets (part II) are obtained from tera-PROMISE Repository (http://openscience.us/repo/defect/mccabehalsted/).

This dataset represents a newer version of the NQUADS files in RDF from Publication Offices used for benchmarking graph databases.

Benchmark is a Point FeatureClass representing land-surveyed benchmarks in Cupertino. Benchmarks are stable sites used to provide elevation data. It is primarily used as a reference layer. The layer is updated as needed by the GIS department. Benchmark has the following fields:

OBJECTID: Unique identifier automatically generated by Esri type: OID, length: 4, domain: none

ID: Unique identifier assigned to the Benchmark type: Integer, length: 4, domain: none

REF_MARK: The reference mark associated with the Benchmark type: String, length: 10, domain: none

ELEV: The elevation of the Benchmark type: Double, length: 8, domain: none

Shape: Field that stores geographic coordinates associated with feature type: Geometry, length: 4, domain: none

Description: A more detailed description of the Benchmark type: String, length: 200, domain: none

Owner: The owner of the Benchmark type: String, length: 10, domain: none

GlobalID: Unique identifier automatically generated for features in enterprise database type: GlobalID, length: 38, domain: none Operator:

The user responsible for updating this database type: String, length: 255, domain: OPERATOR

last_edited_date: The date the database row was last updated type: Date, length: 8, domain: none

created_date: The date the database row was initially created type: Date, length: 8, domain: none

VerticalDatum: The vertical datum associated with the Benchmarktype: String, length: 100, domain: none

The following dataset includes "Active Benchmarks," which are provided to facilitate the identification of City-managed standard benchmarks. Standard benchmarks are for public and private use in establishing a point in space. Note: The benchmarks are referenced to the Chicago City Datum = 0.00, (CCD = 579.88 feet above mean tide New York). The City of Chicago Department of Water Management’s (DWM) Topographic Benchmark is the source of the benchmark information contained in this online database. The information contained in the index card system was compiled by scanning the original cards, then transcribing some of this information to prepare a table and map. Over time, the DWM will contract services to field verify the data and update the index card system and this online database.This dataset was last updated September 2011. Coordinates are estimated. To view map, go to https://data.cityofchicago.org/Buildings/Elevation-Benchmarks-Map/kmt9-pg57 or for PDF map, go to http://cityofchicago.org/content/dam/city/depts/water/supp_info/Benchmarks/BMMap.pdf. Please read the Terms of Use: http://www.cityofchicago.org/city/en/narr/foia/data_disclaimer.html.

In the last two decades, alignment analyses have become an important technique in quantitative historical linguistics and dialectology. Phonetic alignment plays a crucial role in the identification of regular sound correspondences and deeper genealogical relations between and within languages and language families. Surprisingly, up to today, there are no easily accessible benchmark data sets for phonetic alignment analyses. Here we present a publicly available database of manually edited phonetic alignments which can serve as a platform for testing and improving the performance of automatic alignment algorithms. The database consists of a great variety of alignments drawn from a large number of different sources. The data is arranged in a such way that typical problems encountered in phonetic alignment analyses (metathesis, diversity of phonetic sequences) are represented and can be directly tested.

The BuildingsBench datasets consist of:

• Buildings-900K: A large-scale dataset of 900K buildings for pretraining models on the task of short-term load forecasting (STLF). Buildings-900K is statistically representative of the entire U.S. building stock.
• 7 real residential and commercial building datasets for benchmarking two downstream tasks evaluating generalization: zero-shot STLF and transfer learning for STLF.

Buildings-900K can be used for pretraining models on day-ahead STLF for residential and commercial buildings. The specific gap it fills is the lack of large-scale and diverse time series datasets of sufficient size for studying pretraining and finetuning with scalable machine learning models. Buildings-900K consists of synthetically generated energy consumption time series. It is derived from the NREL End-Use Load Profiles (EULP) dataset (see link to this database in the links further below). However, the EULP was not originally developed for the purpose of STLF. Rather, it was developed to "...help electric utilities, grid operators, manufacturers, government entities, and research organizations make critical decisions about prioritizing research and development, utility resource and distribution system planning, and state and local energy planning and regulation." Similar to the EULP, Buildings-900K is a collection of Parquet files and it follows nearly the same Parquet dataset organization as the EULP. As it only contains a single energy consumption time series per building, it is much smaller (~110 GB).

BuildingsBench also provides an evaluation benchmark that is a collection of various open source residential and commercial real building energy consumption datasets. The evaluation datasets, which are provided alongside Buildings-900K below, are collections of CSV files which contain annual energy consumption. The size of the evaluation datasets altogether is less than 1GB, and they are listed out below:

1. ElectricityLoadDiagrams20112014
2. Building Data Genome Project-2
3. Individual household electric power consumption (Sceaux)
4. Borealis
5. SMART
6. IDEAL
7. Low Carbon London

A README file providing details about how the data is stored and describing the organization of the datasets can be found within each data lake version under BuildingsBench.

Benchmark test databases for IQA.

The ont-open-data registry provides reference sequencing data from Oxford Nanopore Technologies to support, 1) Exploration of the characteristics of nanopore sequence data. 2) Assessment and reproduction of performance benchmarks 3) Development of tools and methods. The data deposited showcases DNA sequences from a representative subset of sequencing chemistries. The datasets correspond to publicly-available reference samples (e.g. Genome In A Bottle reference cell lines). Raw data are provided with metadata and scripts to describe sample and data provenance.

Context

Benchmarks allow for easy comparison between multiple CPUs by scoring their performance on a standardized series of tests, and they are useful in many instances: When buying or building a new PC.

Content

Newest data as of May 2nd, 2022 This dataset contains benchmarks of AMD processors.

Acknowledgements

Data scrapped from userbenchmark.

Inspiration

When Lisa Su became CEO of Advanced Micro Devices in 2014, the company was on the brink of bankruptcy. Since then, AMD's stock has soared—from less than US $2 per share to more than $110. The company is now a leader in high-performance computing. She funneled billions of dollars to research and development, while Intel funneled their R&D funds into executive pay. Now Intel is losing a large portion of the market share they originally dominated in.

If you enjoyed this dataset, here's some similar datasets you may like 😎

Benchmarking results for 47Mio Triples based upon DBpedia dataset using 499 queries on TNT, Fuseki, Virtuoso and N-graphStore with approx. 300GB RAM provided for each Triple store

Ground benchmark datasets are published annually in the standard file formats Text (CSV) and XML based on EPSG code 25833. Depending on the file format, ground benchmark data sets are provided completely for the areas of responsibility of the expert committees and for the state of Brandenburg in a zipped file with a statistical indication and a description of the elements. The CSV file is based on VBORIS2. A key bridge to the old format can be taken from the data. On request, ground benchmark data records for municipal areas can be cut out or provided in shape format. Furthermore, the delivery of soil benchmarks in the form of web-based geoservices is possible.