datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
model-benchmark-outputsThis repo contains the outputs of various models on the test set of the UofA-LINGO/text_to_triplets dataset.
Model
Filename
Training Dataset
vicuna-7b-gpt-explanations (benchmark with 1024 cutoff length)
vicuna-gpt-explanations.pickle
UofA-LINGO/webnlg-gpt-explanations
vicuna-7b-combined
vicuna-combined.pickle
UofA-LINGO/webnlg-combined
vicuna-7b-combined-with-reflection
vicuna-with-reflections.pickle
UofA-LINGO/webnlg-combined-with-reflections
vicuna-7b-newins… See the full description on the dataset page: https://huggingface.co/datasets/UofA-LINGO/model-benchmark-outputs.Delighting-Benchmark-v1-Canonical-Outputs
Delighting Benchmark v1 — Canonical Outputs
Canonical Paint3D and Hunyuan3D-2.0 texture-generation outputs for 30 test
cases containing visible illumination and specular effects:
10 ABO matte cases
10 ABO glossy cases
10 Poly Haven PBR cases
Layout
Results are grouped by method:
paint3d/<case_id>/
hunyuan3d/<case_id>/
Every case contains exactly seven files:
input_image.png
input_mesh.obj
textured.glb
albedo_render.mp4
albedo_texture.png
gt_lit_render.mp4… See the full description on the dataset page: https://huggingface.co/datasets/ZeyuLing/Delighting-Benchmark-v1-Canonical-Outputs.asr-benchmark-outputs
SaarAI ASR Benchmark Outputs
Raw per-utterance model outputs (transcription manifests) produced by the gsma-asr-bench runners on SaarAI/asr-leaderboard-datasets.
files: 508
utterances: 4390208
languages: 7
models: 47
Layout
data/<language_name>/<split>__<dataset_config>__<model_slug>.jsonl
index.jsonl # one record per file (language, split, model, rows, sha256, ...)
index.csv
Directories categorise by language name; the file name begins with the split name… See the full description on the dataset page: https://huggingface.co/datasets/SaarAI/asr-benchmark-outputs.gametime-outputs
Gametime Outputs
Model outputs (stereo full-duplex mix of user prompt + model response) for the Gametime benchmark.
Load
from datasets import load_dataset
ds = load_dataset("gametime-benchmark/gametime-outputs", "moshi", split="basic")
ex = next(iter(ds))
wav = ex["audio"]["array"] # numpy float array, shape=(n, 2) stereo
sr = ex["audio"]["sampling_rate"] # int (24000 for most, 48000 for gpt-realtime)
print(ex["id"], sr, wav.shape, ex["dataset"])… See the full description on the dataset page: https://huggingface.co/datasets/gametime-benchmark/gametime-outputs.business-model-kg-benchmark-outputs
Business Model KG Benchmark and Outputs
This repository contains the public evaluation package for Business Model KG:
gold benchmark triples and final generated graph outputs.
Business Model KG turns SEC 10-K business descriptions into standardized
business-model knowledge graphs. This package is meant for inspecting and
reproducing the evaluation side of the project without needing the full training
workspace or runtime model assets.
Layout
benchmarks/
dev/clean/… See the full description on the dataset page: https://huggingface.co/datasets/WindyITS/business-model-kg-benchmark-outputs.ocr-benchmark-outputs
Document OCR using LightOnOCR-2-1B
This dataset contains OCR results from images in byczong/pl-insurance-terms-struct using LightOnOCR-2, a fast and compact 1B OCR model trained with RLVR.
Processing Details
Source Dataset: byczong/pl-insurance-terms-struct
Model: lightonai/LightOnOCR-2-1B
Number of Samples: 109
Processing Time: 5.3 min
Processing Date: 2026-04-17 00:28 UTC
Configuration
Image Column: image
Output Column: markdown
Dataset Split: train… See the full description on the dataset page: https://huggingface.co/datasets/j4xfu2mm/ocr-benchmark-outputs.