Skip to content
Advertisement
Backblaze B2 Genblaze Hackathon with $10K in prizes

Multilingual

575 results

Text

OPUS-100

OPUS-100

10M–100M·Custom / Research-only·Parquet
Text

WMT14

WMT14

10M–100M·Custom / Research-only·Parquet
Text

OpusBooks

OpusBooks

1M–10M·Custom / Research-only·Parquet
Text

PHINC

Abstract Code-mixing is the phenomenon of using more than one language in a sentence. In the multilingual communities,…

10K–100K·CC-BY·CSV
Text

multi30k

Multi30k This dataset contains the "multi30k" dataset, which is the "task 1" dataset from here. Each example consists…

10K–100K·JSON
Image

minimind-v_dataset

Ⅰ 数据集 本轮训练用到的图文数据全部来自 ALLaVA-4V 系列。 相比以往从几份 LLaVA 衍生集拼接得到的数据，ALLaVA-4V 的质量更整齐、中英双语原生对照，细粒度描述也更充分。 它由两个子源构成：一份是 LAION 里挑出来的高质量图片（自然图像为主），一份是 VFLAN…

<1K·Apache-2.0·Images (folder)
ImageMultimodalText

BoundingDocs

BoundingDocs 🔍 The largest spatially-annotated dataset for Document Question Answering Dataset Description BoundingDocs is a unified dataset for…

10K–100K·CC-BY·Parquet
ImageMultimodalText

llava-en-zh-300k

This dataset is composed by 150k examples of English Visual Instruction Data from LLaVA. 150k examples of English…

100K–1M·Apache-2.0·Parquet
MultimodalTextVideo

optimism_train

Optimism Bias World Model Benchmark — Training Data Training data for VLM-as-judge models evaluating world model predictions. Structure…

<1K·MIT
Image

SandThink

SandThink Dataset (v1.0) SandThink 是一个专为具身智能 (Embodied AI) 任务设计的大规模指令微调与偏好对齐数据集。该数据集通过结构化的 Chain-of-Thought (CoT) 推理过程，显著提升了 Vision-Language-Action…

<1K·MIT·Images (folder)
Image

C3B

English 简体中文 C³B: Comics Cross-Cultural Benchmark Culture In a Frame: C³B as a Comic-Based Benchmark for Multimodal Cultural…

1K–10K·CC-BY·Images (folder)
Advertisement
Backblaze B2 Genblaze Hackathon with $10K in prizes