cowWhySo/va-tool-calling-dataset

---
license: apache-2.0
tags:
- voice-assistant
- tool-calling
- function-calling
- synthetic
- qwen3.5
- conversational
language:
- en
size_categories:
- n<1K
---

# Voice Assistant Tool-Calling Dataset

Synthetic multi-turn conversations for fine-tuning a voice assistant with tool-calling capabilities.

## Stats

| Metric | Value |
|---|---|
| Total conversations | 1585 |
| Tool-call conversations | 898 |
| Direct-answer conversations | 687 |
| Tools defined | 6 |
| Generator model | anthropic/claude-haiku-4.5 |

## Tools

| Tool | Description |
|---|---|
| `get_weather` | Current weather and forecast |
| `set_timer` | Countdown timers and alarms |
| `create_reminder` | Time-based reminders |
| `control_smart_home` | Lights, thermostat, locks, etc. |
| `play_music` | Music by song, artist, genre, mood |
| `web_search` | Current info, news, scores |

## Format

Each line in `data/train.jsonl` is a JSON object with:
- `conversations`: message array (system, user, assistant, tool)
- `tools`: tool definitions (empty array for no-tool conversations)
- `has_tool_call`: boolean flag

Compatible with Qwen3.5 `apply_chat_template(tools=...)`.

## Usage

```python
from datasets import load_dataset
ds = load_dataset("cowWhySo/va-tool-calling-dataset", split="train")
```