Model Composition and Strengths

This configuration merges models with complementary strengths across the following areas:

• Uncensored responses
• Roleplay (LimaRP, Samantha)
• Instruction following
• Creative writing
• General assistance

Mystral-Uncensored-RP-7B

Mystral-Uncensored-RP-7B is a merge of the following models using LazyMergekit:

• Naphula/Warlock-7B-v3
• luvGPT/mistral-7b-uncensored
• QuixiAI/samantha-mistral-instruct-7b
• SilverFan/LimaRP-daybreak-7B
• Ppoyaa/Niel-7B
• JoPmt/Trismal-Obrum-v2-7B-Base-Ties
• mpasila/SeaMax-7B

🧩 Configuration

slices:
  - sources:
      - model: Naphula/Warlock-7B-v3
        layer_range: [0, 32]
      - model: luvGPT/mistral-7b-uncensored
        layer_range: [0, 32]
      - model: QuixiAI/samantha-mistral-instruct-7b
        layer_range: [0, 32]
      - model: SilverFan/LimaRP-daybreak-7B
        layer_range: [0, 32]
      - model: Ppoyaa/Niel-7B
        layer_range: [0, 32]
      - model: JoPmt/Trismal-Obrum-v2-7B-Base-Ties
        layer_range: [0, 32]
      - model: mpasila/SeaMax-7B
        layer_range: [0, 32]
merge_method: ties
base_model: Naphula/Warlock-7B-v3
parameters:
  density: 0.6
  weight: 0.4
  normalize: true
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Stormtrooperaim/Mystral-Uncensored-RP-7B"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])