---
license: apache-2.0
language:
- ar
- en
tags:
- islamic
- arabic
- tool-use
- function-calling
- conversational
- speculative-decoding
- dspark
size_categories:
- 1K<n<10K
---

# DSpark-Muslim Training Conversations

1,594 real, tool-augmented conversations generated by running
[`NightPrince/Muslim-6B-v3`](https://huggingface.co/NightPrince/Muslim-6B-v3) — with its
real production MCP tools wired in (hadith, tafsir, fatwa, recitation validation, audio) —
against a curated bank of religious questions. Built to train a domain-specific
[DSpark](https://arxiv.org/abs/2607.05147) speculative-decoding draft model for
Muslim-6B-v3; see [DSpark-Muslim-Model](https://github.com/NightPrinceY/DSpark-Muslim-Model)
for the full pipeline.

## Why this exists

A generic DSpark draft checkpoint trained on stock Qwen3 doesn't match a domain
fine-tune's actual behavior (persona, tool-calling format, measured rulings) well. The
fix: generate training data by running the *actual* target model through its *actual*
production tools, so a draft model trained on it learns to predict tokens in the
presence of real tool-call/tool-result context — not just freeform Q&A.

## Format

One JSON object per line:
```json
{
  "id": "q_00001",
  "source": "generated",
  "intent": "surah_info",
  "conversations": [
    {"role": "system", "content": "..."},
    {"role": "user", "content": "كم عدد آيات سورة النحل؟"},
    {"role": "assistant", "content": "", "tool_calls": [{"id": "...", "type": "function", "function": {"name": "get_surah_statistics", "arguments": "..."}}]},
    {"role": "tool", "tool_call_id": "...", "name": "get_surah_statistics", "content": "..."},
    {"role": "assistant", "content": "سورة النحل عدد آياتها مئة وثمانية وعشرون آية."}
  ]
}
```

`conversations` is directly usable as the `messages` field for `tokenizer.apply_chat_template`
(Qwen3/Hermes-style tool-calling format).

## Stats

- **1,594 conversations** (of 1,728 seed questions attempted — ~7.8% rejected by an
  automated quality gate: TTS-clean violations (raw digits/markdown in the final answer) or
  malformed role sequencing)
- **80.2%** involve at least one real tool call
- **40 distinct tool names invoked**, across 6 real MCP servers (validator, tafsir/audio,
  hadith, `mcp.tafsir.net`, IslamQA fatwa search, web search)
- Average 4.6 turns per conversation, max 9
- Seed question sources: 1,127 systematically generated (covering the full tool
  taxonomy), 375 hand-curated (from a LoRA fine-tuning dataset), 65 manually-curated real
  production voice-session questions, 27 behavioral probe prompts

## Known limitation

The model occasionally hallucinates a slightly-wrong tool name (e.g. `searchAnswers`
instead of `search_answers`, ~41 occurrences) — these are kept as-is rather than filtered,
since the goal is distilling this specific model's real output distribution (including its
real imperfections) for speculative decoding, not curating a "correct tool-calling" dataset.

## License

Apache 2.0, matching Muslim-6B-v3's own license. The underlying tafsir/hadith/fatwa text
retrieved by tool calls carries its own upstream licensing (Sunnah.com/hadith-json for
hadith, individual tafsir book copyrights, IslamQA for fatwa text) — this dataset
distributes it only as part of these training conversations, same as any other model
training corpus built on public reference text.