Guilherme34 / Firefly

Firefly

Firefly is a merge of the following models using LazyMergekit :

• SicariusSicariiStuff/Impish_LLAMA_3B
• Guilherme34/Lumina
• ValiantLabs/Llama3.2-3B-Enigma
• DavidAU/Llama-3.2-3B-Instruct-heretic-ablitered-uncensored

🧩 Configuration

models:
  - model: SicariusSicariiStuff/Impish_LLAMA_3B
    parameters:
      weight: 1.0  

  - model: Guilherme34/Lumina
    parameters:
      weight: 0.75  

  - model: ValiantLabs/Llama3.2-3B-Enigma
    parameters:
      weight: 0.125  

  - model: DavidAU/Llama-3.2-3B-Instruct-heretic-ablitered-uncensored
    parameters:
      weight: 0.125  

merge_method: task_arithmetic
base_model: SicariusSicariiStuff/Impish_LLAMA_3B

dtype: bfloat16

parameters:
  normalize: true   
  int8_mask: true   

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Guilherme34/Firefly"
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"])