Luna-8B-Instruct-262k

Luna-8B-Instruct-262k is a merge of the following models using LazyMergekit:

🧩 Configuration

slices:
models:
    - model: nbeerbower/llama-3-wissenschaft-8B
    - model: mlabonne/ChimeraLlama-3-8B-v3
    - model: jondurbin/bagel-8b-v1.0
merge_method: model_stock
base_model: gradientai/Llama-3-8B-Instruct-262k
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Ppoyaa/Luna-8B-Instruct-262k"
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"])