The model normalizes imperfect audio transcriptions, applies domain-specific glossaries, and translates character dialogue while retaining emotion and nuances.

This variant preserves the upstream Qwen3.5 MTP / speculative-decoding head in GGUF format so it can be used with MTP-capable llama.cpp builds.

Standard Mode

The traditional output format where only the translated text is returned.

Available Quantizations

Quantization	Filename	Size	Description
q4_k_m	`asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-q4_k_m.gguf`	5.4 GB	Good balance of quality and size
q6_k	`asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-q6_k.gguf`	7.0 GB	Higher quality, moderate size
q8_0	`asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-q8_0.gguf`	9.1 GB	Near-lossless quality
bf16	`asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-bf16.gguf`	17.1 GB	Full BF16, no quantization loss

Prompt Example

text

将以下日语ASMR逐字稿翻译成简体中文。

音轨：track01_示例音轨
场景说明：主角与青梅竹马在校园下午的对话...

术语表（请严格使用zh栏位的译名）：
{
  "cvs": [],
  "characters": [],
  "terms": [{"ja": "放課後", "zh": "放学后"}]
}

翻译前请静默修正以下Whisper识别错误：
- 重复片语（连续3次以上且无变化）：仅保留一次
- 错字／同音异字：依上下文修正
- 字幕版权行（字幕：／翻訳：／QQ／LINE水印）：text设为null
- 错误专有名词：依术语表修正

翻译规则：
- 呻吟与气息声（あ、ん、はあ）→ 自然对应（啊、嗯、哈、呼）
- 拟声词：日语形式翻译（パンパン→啪啪）；中文形式保留原样
- 保留角色语气与口吻
- text字段只输出译文，不加注释或括号说明

输入：逐字稿JSON数组 — {"id": <n>, "text": "<日文>", "start": <ms>, "end": <ms>}

输出：将连续构成同一句话的片段合并，JSON数组格式：
{"ids": [<n>, ...], "text": "<简体中文>", "start": <最早ms>, "end": <最晚ms>}

字幕版权行：{"ids": [<n>], "text": null, "start": <ms>, "end": <ms>}
每个输入id必须恰好出现在一个输出项中。

逐字稿：
[
  {"id": 1, "text": "ねぇ、放課後、", "start": 3000, "end": 5000},
  {"id": 2, "text": "一緒に帰らない?", "start": 5000, "end": 7000}
]

Example Output:

[{"ids": [1, 2], "text": "呐，放学后，要不要一起回去？", "start": 3000, "end": 7000}]

Usage

llama-server

llama-server -m asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-q4_k_m.gguf -c 4096 --port 8080

llama-cli

llama-cli -m asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-q4_k_m.gguf -p "<your prompt>" -n 2048

llama-cli with MTP

llama-cli -m asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1-q4_k_m.gguf --spec-type draft-mtp --spec-draft-n-max 6 -p "<your prompt>" -n 2048

Structured Decoding (Recommended)

This model outputs JSON arrays. Using structured decoding (e.g. GBNF grammar or JSON schema constraints) avoids wasted computation on malformed output and guarantees valid JSON on every generation.

JSON Schema:

json

{
  "type": "array",
  "items": {
    "type": "object",
    "properties": {
      "ids": {
        "type": "array",
        "items": {
          "type": "integer"
        },
        "minItems": 1
      },
      "text": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ]
      },
      "start": {
        "type": "integer"
      },
      "end": {
        "type": "integer"
      }
    },
    "required": [
      "ids",
      "text",
      "start",
      "end"
    ],
    "additionalProperties": false
  },
  "minItems": 1
}

Supported by llama.cpp (--json-schema), vLLM, and outlines.

Training Details

Base model: unsloth/Qwen3.5-9B
Method: LoRA (r=16, alpha=16)
Target modules: q_proj, v_proj, up_proj, gate_proj, k_proj, down_proj, o_proj
Locale: zh-cn (Simplified Chinese)
Mode: Standard Mode
Max sequence length: 4096
Precision: bf16

mmis1000/asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1

asmr-qwen3.5-9b-zh-cn-mtp-gguf-v0.1