--- dataset_info: - config_name: tloc_sg_streets_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 80139065 num_examples: 95 download_size: 78394896 dataset_size: 80139065 - config_name: tloc_sgpccsc_long_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 461763833 num_examples: 512 download_size: 450313398 dataset_size: 461763833 - config_name: tloc_yodas2_en_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 866390375 num_examples: 1000 download_size: 846616368 dataset_size: 866390375 - config_name: tloc_yodas2_id_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 841099481 num_examples: 1000 download_size: 827864029 dataset_size: 841099481 - config_name: tloc_yodas2_th_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 868683702 num_examples: 1000 download_size: 849480761 dataset_size: 868683702 - config_name: tloc_yodas2_th_60 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 1595309043.0 num_examples: 1000 download_size: 1558501892 dataset_size: 1595309043.0 - config_name: tloc_yodas2_vi_120 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 515445529 num_examples: 163 download_size: 504642783 dataset_size: 515445529 - config_name: tloc_yodas2_vi_180 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 316303389 num_examples: 62 download_size: 311145516 dataset_size: 316303389 - config_name: tloc_yodas2_vi_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 418032624 num_examples: 489 download_size: 409055243 dataset_size: 418032624 - config_name: tloc_yodas2_vi_60 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 563712163 num_examples: 341 download_size: 550455076 dataset_size: 563712163 - config_name: tloc_yodas2_zh_120 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 756265938 num_examples: 271 download_size: 744389765 dataset_size: 756265938 - config_name: tloc_yodas2_zh_180 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 242337197 num_examples: 49 download_size: 237790999 dataset_size: 242337197 - config_name: tloc_yodas2_zh_30 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 876703974 num_examples: 1000 download_size: 865964059 dataset_size: 876703974 - config_name: tloc_yodas2_zh_60 features: - name: context dtype: audio: sampling_rate: 16000 - name: instruction dtype: string - name: answer dtype: string - name: audio_length dtype: float64 - name: language dtype: string splits: - name: train num_bytes: 1182729820 num_examples: 754 download_size: 1167428388 dataset_size: 1182729820 configs: - config_name: tloc_sg_streets_30 data_files: - split: train path: tloc_sg_streets_30/train-* - config_name: tloc_sgpccsc_long_30 data_files: - split: train path: tloc_sgpccsc_long_30/train-* - config_name: tloc_yodas2_en_30 data_files: - split: train path: tloc_yodas2_en_30/train-* - config_name: tloc_yodas2_id_30 data_files: - split: train path: tloc_yodas2_id_30/train-* - config_name: tloc_yodas2_th_30 data_files: - split: train path: tloc_yodas2_th_30/train-* - config_name: tloc_yodas2_th_60 data_files: - split: train path: tloc_yodas2_th_60/train-* - config_name: tloc_yodas2_vi_120 data_files: - split: train path: tloc_yodas2_vi_120/train-* - config_name: tloc_yodas2_vi_180 data_files: - split: train path: tloc_yodas2_vi_180/train-* - config_name: tloc_yodas2_vi_30 data_files: - split: train path: tloc_yodas2_vi_30/train-* - config_name: tloc_yodas2_vi_60 data_files: - split: train path: tloc_yodas2_vi_60/train-* - config_name: tloc_yodas2_zh_120 data_files: - split: train path: tloc_yodas2_zh_120/train-* - config_name: tloc_yodas2_zh_180 data_files: - split: train path: tloc_yodas2_zh_180/train-* - config_name: tloc_yodas2_zh_30 data_files: - split: train path: tloc_yodas2_zh_30/train-* - config_name: tloc_yodas2_zh_60 data_files: - split: train path: tloc_yodas2_zh_60/train-* license: cc-by-nc-nd-4.0 task_categories: - question-answering language: - en - zh - vi - th - id tags: - speech - audio-llm - temporal-localization - long-form-audio - benchmark - southeast-asia pretty_name: TLoc — Temporal Localization size_categories: - 1K=4.0`, which decodes audio through `torchcodec` and needs a system FFmpeg (versions 4–7): ```bash pip install "datasets[audio]" ``` ```python from datasets import load_dataset ds = load_dataset( "MERaLiON/sea_audiobench_datasets_TLoc", "tloc_yodas2_en_30", split="train", ) row = ds[0] print(row["instruction"]) print(row["answer"]) print(row["language"], row["audio_length"]) # `context` is a torchcodec AudioDecoder, not a dict samples = row["context"].get_all_samples() waveform = samples.data # torch.Tensor, shape (num_channels, num_samples) sr = samples.sample_rate # 16000 print(waveform.shape, sr) ``` If you need a numpy array — most feature extractors take one — squeeze the channel dimension (all TLoc audio is mono): ```python audio = samples.data.squeeze(0).numpy() ``` ## Subsets All data lives in a single `train` split and is intended for **evaluation only**. Config names follow `tloc_{source}_{language}_{context_length_seconds}`. | Config | Source | Lang | Ctx (s) | Examples | Nominal hrs | |---|---|---|---|---|---| | `tloc_sg_streets_30` | SG Streets | en | 30 | 95 | 0.8 | | `tloc_sgpccsc_long_30` | ASR-SgpCCSC | zh | 30 | 512 | 4.3 | | `tloc_yodas2_en_30` | YODAS2 | en | 30 | 1,000 | 8.3 | | `tloc_yodas2_id_30` | YODAS2 | id | 30 | 1,000 | 8.3 | | `tloc_yodas2_th_30` | YODAS2 | th | 30 | 1,000 | 8.3 | | `tloc_yodas2_vi_30` | YODAS2 | vi | 30 | 489 | 4.1 | | `tloc_yodas2_vi_60` | YODAS2 | vi | 60 | 341 | 5.7 | | `tloc_yodas2_vi_120` | YODAS2 | vi | 120 | 163 | 5.4 | | `tloc_yodas2_vi_180` | YODAS2 | vi | 180 | 62 | 3.1 | | `tloc_yodas2_zh_30` | YODAS2 | zh | 30 | 1,000 | 8.3 | | `tloc_yodas2_zh_60` | YODAS2 | zh | 60 | 754 | 12.6 | | `tloc_yodas2_zh_120` | YODAS2 | zh | 120 | 271 | 9.0 | | `tloc_yodas2_zh_180` | YODAS2 | zh | 180 | 49 | 2.5 | | **Total** | | | | **6,736** | **~80.7** | ## Data fields | Field | Type | Description | |---|---|---| | `context` | `Audio(sampling_rate=16000)` | The audio recording. | | `instruction` | `string` | The instruction given to the model. | | `answer` | `string` | The reference answer. | | `audio_length` | `float64` | Duration of `context`, in seconds. | | `language` | `string` | Language code of the subset. | See the paper for the instruction and answer formats, how segments and annotations were constructed, and the evaluation protocol. ## Source data TLoc introduces no new recordings. All audio comes from existing corpora; the contribution is the segmentation, instructions, and reference answers. | Source | Lang | Upstream license | |---|---|---| | [YODAS2](https://huggingface.co/datasets/espnet/yodas2) | en, id, zh, th, vi | CC BY 3.0 | | SG Streets | en | Not specified | | [ASR-SgpCCSC](https://magichub.com/) | zh | CC BY-NC-ND 4.0 | ## Citation If you use TLoc, please cite our benchmark **and** the source corpora. ### This benchmark ```bibtex @inproceedings{liao2026seaspeechbench, title = {{SEA-SpeechBench}: A Large-Scale Multitask Benchmark for Speech Understanding Across Southeast Asia}, author = {Liao, Jingyi and Zhang, Wenyu and Liu, Zhuohan and He, Yingxu and Lin, Geyu and Zou, Xunlong and Sun, Shuo and Alsagoff, Syed Ali Redha and Aw, Ai Ti}, booktitle = {Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP)}, year = {2026} } ``` ### Source corpora ```bibtex @inproceedings{li2023yodas, title = {YODAS: Youtube-Oriented Dataset for Audio and Speech}, author = {Li, Xinjian and Takamichi, Shinnosuke and Saeki, Takaaki and Chen, William and Shiota, Sayaka and Watanabe, Shinji}, booktitle = {2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)}, pages = {1--8}, year = {2023}, organization = {IEEE} } @misc{sgpccsc, title = {{ASR-SgpCCSC: A Singapore Chinese Conversational Speech Corpus}}, author = {Magic Data Tech}, year = {2024}, url = {https://magichub.com/datasets/[SLUG]/} } @inproceedings{khassanov2019enriching, title = {Enriching Rare Word Representations in Neural Language Models by Embedding Matrix Augmentation}, author = {Khassanov, Yerbolat and Zeng, Zhiping and Pham, Van Tung and Xu, Haihua and Chng, Eng Siong}, booktitle = {Interspeech 2019}, pages = {3505--3509}, year = {2019}, month = sep, publisher = {ISCA}, doi = {10.21437/Interspeech.2019-1858}, url = {http://dx.doi.org/10.21437/Interspeech.2019-1858} } ```