How to use from
llama.cpp
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf limloop/MN-12B-Faun-RP-RU-GGUF:
# Run inference directly in the terminal:
llama cli -hf limloop/MN-12B-Faun-RP-RU-GGUF:
Install from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf limloop/MN-12B-Faun-RP-RU-GGUF:
# Run inference directly in the terminal:
llama cli -hf limloop/MN-12B-Faun-RP-RU-GGUF:
Use pre-built binary
# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf limloop/MN-12B-Faun-RP-RU-GGUF:
# Run inference directly in the terminal:
./llama-cli -hf limloop/MN-12B-Faun-RP-RU-GGUF:
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf limloop/MN-12B-Faun-RP-RU-GGUF:
# Run inference directly in the terminal:
./build/bin/llama-cli -hf limloop/MN-12B-Faun-RP-RU-GGUF:
Use Docker
docker model run hf.co/limloop/MN-12B-Faun-RP-RU-GGUF:
Quick Links

MN-12B-Faun-RP-RU

Original model

๐Ÿ‡ท๐Ÿ‡บ ะะฐะถะผะธั‚ะต, ั‡ั‚ะพะฑั‹ ั€ะฐะทะฒะตั€ะฝัƒั‚ัŒ ะพะฟะธัะฐะฝะธะต ะฝะฐ ั€ัƒััะบะพะผ

๐ŸŒŸ ะž ะผะพะดะตะปะธ

MN-12B-Faun-RP-RU โ€” ัƒะปัƒั‡ัˆะตะฝะฝั‹ะน merge ะฝะฐ ะฑะฐะทะต Mistral Nemo 12B, ั€ะฐะทะฒะธะฒะฐัŽั‰ะธะน ะธะดะตะธ Hydra ะธ ะพั€ะธะตะฝั‚ะธั€ะพะฒะฐะฝะฝั‹ะน ะฝะฐ:

  • ๐ŸŽญ ะ‘ะพะปะตะต ัั‚ะฐะฑะธะปัŒะฝั‹ะน ะธ ะฒั‹ั€ะฐะทะธั‚ะตะปัŒะฝั‹ะน roleplay
  • ๐Ÿ“š ะฃะปัƒั‡ัˆะตะฝะฝั‹ะน ั€ัƒััะบะธะน ัะทั‹ะบ
  • ๐Ÿง  ะ ะฐััˆะธั€ะตะฝะฝั‹ะน ัะปะพะฒะฐั€ะฝั‹ะน ะทะฐะฟะฐั, ะฒะบะปัŽั‡ะฐั ัะปะพะถะฝั‹ะต ะธ NSFW-ั‚ะตะผั‹
  • ๐Ÿ”“ ะŸั€ะฐะบั‚ะธั‡ะตัะบะธ ะพั‚ััƒั‚ัั‚ะฒัƒัŽั‰ัƒัŽ ั†ะตะฝะทัƒั€ัƒ

ะœะพะดะตะปัŒ ัะพะฑั€ะฐะฝะฐ ะผะตั‚ะพะดะพะผ TIES-merging ะธ ะฝะต ะฟั€ะพั…ะพะดะธะปะฐ ะดะพะฟะพะปะฝะธั‚ะตะปัŒะฝะพะณะพ ะพะฑัƒั‡ะตะฝะธั ะฟะพัะปะต ัะปะธัะฝะธั.

๐ŸŽฏ ะžัะพะฑะตะฝะฝะพัั‚ะธ

  • ะžัะฝะพะฒะฝะพะน ั„ะพะบัƒั โ€” ั€ัƒััะบะธะน ัะทั‹ะบ
  • ะ›ัƒั‡ัˆะต ัƒะดะตั€ะถะธะฒะฐะตั‚ ะฟะตั€ัะพะฝะฐะถะตะน ะธ ัั‚ะธะปัŒ ะดะธะฐะปะพะณะฐ
  • ะ‘ะพะปะตะต ะฑะพะณะฐั‚ะฐั ะธ ะฒะฐั€ะธะฐั‚ะธะฒะฝะฐั ะณะตะฝะตั€ะฐั†ะธั
  • ะฃะปัƒั‡ัˆะตะฝะฝะฐั ัั‚ะฐะฑะธะปัŒะฝะพัั‚ัŒ ะฝะฐ ะดะปะธะฝะฝั‹ั… ะบะพะฝั‚ะตะบัั‚ะฐั… (ะฟั€ะพะฒะตั€ะตะฝะพ ะดะพ ~8192 ั‚ะพะบะตะฝะพะฒ)
  • ะกะปะตะดัƒะตั‚ ะธะฝัั‚ั€ัƒะบั†ะธัะผ, ะฝะพ ะผะพะถะตั‚ ะดะพะฑะฐะฒะปัั‚ัŒ ะดะธัะบะปะตะนะผะตั€ั‹ ะฝะฐ ั‡ัƒะฒัั‚ะฒะธั‚ะตะปัŒะฝั‹ะต ะทะฐะฟั€ะพัั‹

โš ๏ธ ะ’ะฐะถะฝะพ

ะœะพะดะตะปัŒ ัะพั…ั€ะฐะฝัะตั‚ uncensored-ั…ะฐั€ะฐะบั‚ะตั€, ะพะดะฝะฐะบะพ ะฒ ะฝะตะบะพั‚ะพั€ั‹ั… ัะปัƒั‡ะฐัั… ะผะพะถะตั‚ ะดะพะฑะฐะฒะปัั‚ัŒ ะฟั€ะตะดัƒะฟั€ะตะถะดะตะฝะธั ะพ ะฝะตะฟะพะดั…ะพะดัั‰ะตะผ ะบะพะฝั‚ะตะฝั‚ะต ะฟั€ะธ ะฟั€ัะผั‹ั… ะทะฐะฟั€ะพัะฐั…. ะŸั€ะธ ัั‚ะพะผ ะณะตะฝะตั€ะฐั†ะธั ะฝะต ะฑะปะพะบะธั€ัƒะตั‚ัั ะธ ะฟั€ะพะดะพะปะถะฐะตั‚ัั ะฟะพัะปะต ะดะธัะบะปะตะนะผะตั€ะฐ.

High-quality TIES merge based on Mistral Nemo 12B, focused on improved Russian fluency, stronger roleplay, richer vocabulary, and stable long-context performance.


๐ŸŒ Overview

MN-12B-Faun-RP-RU is an evolution of the Hydra-style merge, designed to push further in roleplay quality, language richness, and generation stability.

Key improvements include:

  • ๐Ÿ“š Better Russian
  • ๐ŸŽญ More consistent and immersive roleplay behavior
  • ๐Ÿง  Expanded vocabulary, including expressive and NSFW domains
  • ๐Ÿ” More stable handling of long conversations (tested up to ~8k tokens)

The model may occasionally produce safety disclaimers when prompted directly for sensitive content, but generation continues normally afterward.

Built using TIES merging, which minimizes destructive interference between merged model weights.


๐ŸŽฏ Key Features

Feature Description
Languages Russian, English
Censorship Mostly uncensored (with occasional disclaimers)
Roleplay Improved consistency and immersion
Instruction Following Strong
Vocabulary Expanded, including NSFW domains
Context Length Stable up to ~8192 tokens
Architecture Mistral Nemo 12B

๐Ÿงฉ Model Composition

The merge combines the following models:

Model Role in merge Weight
MN-12B-Hydra-RP-RU Base / foundation 0.60
Impish_Bloodmoon_12B RP + style boost 0.25
Forgotten-Safeword-12B-v4.0 Uncensored behavior 0.10

Weights shown before normalization.


๐Ÿ’ก Usage Example

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_name = "limloop/MN-12B-Faun-RP-RU"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)

prompt = "You are a mysterious forest faun speaking in poetic Russian."
messages = [{"role": "user", "content": prompt}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)

outputs = model.generate(inputs, max_new_tokens=512, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

โš™๏ธ Merge Details

Built using mergekit with the TIES method (Trim, Elect Sign, Merge).

Core mechanism:

  1. Trim low-magnitude deltas via density
  2. Resolve sign conflicts
  3. Weighted averaging of aligned parameters

Merge Configuration

models:
  - model: limloop/MN-12B-Hydra-RP-RU
    parameters:
      weight: 0.6

  - model: SicariusSicariiStuff/Impish_Bloodmoon_12B
    parameters:
      weight: 0.25
      density: 0.9

  - model: ReadyArt/Forgotten-Safeword-12B-v4.0
    parameters:
      weight: 0.1
      density: 0.6

merge_method: ties
parameters:
  epsilon: 0.01
  normalize: true

base_model: limloop/MN-12B-Hydra-RP-RU
dtype: bfloat16

tokenizer:
  source: "base"

โš ๏ธ Known Characteristics

  • No post-merge fine-tuning
  • May produce safety disclaimers before sensitive outputs
  • Occasionally switches to English in complex reasoning
  • Stronger stylistic bias in roleplay compared to Hydra
Downloads last month
3,778
GGUF
Model size
12B params
Architecture
llama
Hardware compatibility
Log In to add your hardware

3-bit

4-bit

5-bit

6-bit

8-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for limloop/MN-12B-Faun-RP-RU-GGUF

Quantized
(3)
this model

Space using limloop/MN-12B-Faun-RP-RU-GGUF 1

Collection including limloop/MN-12B-Faun-RP-RU-GGUF