A single-turn Vietnamese speech dataset of work instructions used on shipbuilding and manufacturing sites. Scripts are the instructions actually issued on the floor — equipment inspection, safety briefings, tool handover and storage, measurement recording, progress reporting, consumable distribution — phrased as a supervisor speaking to a worker, and recorded by Vietnamese speakers so that models can recognize commands as they are spoken by the foreign-national workforce common to these sites.
Equipment and tool names are capitalized mid-sentence so domain terminology is marked inside the transcript. Because on-site speech is dense with equipment vocabulary and delivered as short imperatives, the dataset keeps that phrasing rather than normalizing it into written prose.
| filename | transcript | gender |
|---|---|---|
| 00074_1 | Hãy thu gom toàn bộ vật tư còn lại trước khi Kết thúc công việc, hoàn trả về kho vật tư và ghi vào sổ theo dõi số lượng. | female |
| 00074_3 | Kiểm tra xem mọi người đã nắm vững các quy tắc an toàn liên quan đến Dây đai an toàn chưa, ai chưa nắm rõ thì huấn luyện lại đi. | female |
| 00074_5 | Hôm nay xong việc giũa lưỡi cưa ở đuôi tàu thì kiểm tra tình trạng lưỡi cưa rồi báo cáo kết quả. | female |
| 00074_10 | Hãy kiểm tra kỹ tình trạng Máy hút bụi cầm tay ở khu vực thi công ống, nếu có gì bất thường thì báo ngay cho tổ trưởng. | female |
Work instruction text is gathered from shipbuilding and manufacturing sites and normalized into single-utterance scripts, with equipment names, tool names, and measurement terms preserved rather than simplified. Scripts are balanced across instruction types — inspection, safety, preparation and placement, recording and reporting, cleanup and return — so that the set is not dominated by any one command pattern. Each script is then recorded by a Vietnamese speaker, and the speaker's gender is stored on the record.
Every recording is transcribed and checked against the source script for omissions, substitutions, and mispronounced equipment terms, since a single misheard tool name changes the instruction. Reviewers additionally confirm that recording level, clipping, and background noise fall within spec, because deployment is in a noise-heavy environment. Utterances failing transcript match or audio spec are re-recorded rather than corrected in post-processing.