Datasets:
| license: apache-2.0 | |
| task_categories: | |
| - audio-classification | |
| - automatic-speech-recognition | |
| language: | |
| - hi | |
| - te | |
| - or | |
| size_categories: | |
| - 10K<n<100K | |
| This is a domain-specific, multilingual agricultural speech dataset with a primary focus on Hindi, Telugu, Oriya designed for speech-to-text and automatic speech recognition tasks, featuring human-annotated transcriptions and intended for benchmarking ASR model performance in real-world agricultural scenarios | |
| ## Reference | |
| If you use this dataset, please cite: | |
| - Benchmarking Automatic Speech Recognition for Indian Languages in Agricultural Contexts | |
| arXiv:2602.03868 – https://arxiv.org/abs/2602.03868 |