How to use from
Hermes Agent
Start the llama.cpp server
# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf TAOTAO777/ai-girlfriend-natsume:Q4_K_S
Configure Hermes
# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default TAOTAO777/ai-girlfriend-natsume:Q4_K_S
Run Hermes
hermes
Quick Links

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

AI Girlfriend — Shiki Natsume and atri· Model Repository

This is the model distribution repository for AI Girlfriend.

📖 English | 中文


⚠️ Disclaimer

All model files in this repository are existing community open-source models. I am only mirroring and redistributing them for convenience. I claim no ownership over any of these models.

Original authors and licenses:

Model Author Source License
Qwen3.6-35B-A3B-APEX mudler HuggingFace Apache 2.0 (Qwen)
WAI-Nsfw-Illustrious-17 WAI0731 CivitAI Community model
miaomiaoHarem_v20 miaomiao0226 CivitAI Community model
GPT-SoVITS weights Self-trained Free to use

This project is purely non-profit and personal. No commercial use intended. If any original author objects to mirror distribution, please contact me to take it down.

Download

# Download everything
huggingface-cli download TAOTAO777/ai-girlfriend-natsume --local-dir ./models

# Or selectively:
huggingface-cli download TAOTAO777/ai-girlfriend-natsume llm/ --local-dir ./models
huggingface-cli download TAOTAO777/ai-girlfriend-natsume comfyui-checkpoints/ --local-dir ./checkpoints
huggingface-cli download TAOTAO777/ai-girlfriend-natsume gpt-sovits-weights/ --local-dir ./gpt-sovits-weights

Directory Mapping

Place downloaded files according to the source project's expected paths:

llm/
└── Qwen3.6-35B-A3B-APEX-I-Compact.gguf
    → Expected at: <your vllm>/models/

comfyui-checkpoints/
├── WAI-Nsfw-Illustrious-17.safetensors
└── miaomiaoHarem_v20.safetensors
    → Expected at: <ComfyUI>/models/checkpoints/

gpt-sovits-weights/
├── SoVITS_weights_v2Pro/xxx-e30.ckpt
└── GPT_weights_v2Pro/xxx_e20_s6240.pth
    → Expected at: <GPT-SoVITS>/SoVITS_weights_v2Pro/ and GPT_weights_v2Pro/

Main Project Repo

Source code and configuration: TAOTAO777/ai-girlfriend-natsume (GitHub)


中文章节

⚠️ 免责声明

本仓库中的模型文件均为社区开源模型,本人仅做收集和镜像分发,不声称任何所有权。本项目完全非盈利,纯个人兴趣,不涉及任何商业用途。若原作者认为镜像分发不妥,请联系我删除。

原作者列表

模型 原作者 原始来源 许可证
Qwen3.6-35B-A3B-APEX mudler HuggingFace Apache 2.0 (Qwen)
WAI-Nsfw-Illustrious-17 WAI0731 CivitAI 社区模型
miaomiaoHarem_v20 miaomiao0226 CivitAI 社区模型
GPT-SoVITS 权重 本人训练 自由使用

下载方式

# 全量下载
huggingface-cli download TAOTAO777/ai-girlfriend-natsume --local-dir ./models

目录对应

llm/           → <你的vllm>/models/
comfyui-checkpoints/  → <ComfyUI>/models/checkpoints/
gpt-sovits-weights/   → <GPT-SoVITS>/SoVITS_weights_v2Pro/ + GPT_weights_v2Pro/

主项目

代码和配置:TAOTAO777/ai-girlfriend-natsume (GitHub)

Downloads last month
1,968
GGUF
Model size
27B params
Architecture
qwen35
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support