MergedLlama-3-8B-MS-2

MergedLlama-3-8B-MS-2 is a merge of the following models using LazyMergekit:

• NousResearch/Meta-Llama-3-8B-Instruct
• Weyaxi/Einstein-v6.1-Llama3-8B
• cognitivecomputations/dolphin-2.9-llama3-8b
• nvidia/Llama3-ChatQA-1.5-8B
• Kukedlc/SmartLlama-3-8B-MS-v0.1
• mlabonne/ChimeraLlama-3-8B-v3

🧩 Configuration

models:
  - model: NousResearch/Meta-Llama-3-8B
    # No parameters necessary for base model
  - model: NousResearch/Meta-Llama-3-8B-Instruct
    parameters:
      density: 0.6
      weight: 2
  - model: Weyaxi/Einstein-v6.1-Llama3-8B
    parameters:
      density: 0.55
      weight: 2
  - model: cognitivecomputations/dolphin-2.9-llama3-8b
    parameters:
      density: 0.55
      weight: 2
  - model: nvidia/Llama3-ChatQA-1.5-8B
    parameters:
      density: 0.55
      weight: 2
  - model: Kukedlc/SmartLlama-3-8B-MS-v0.1
    parameters:
      density: 0.66
      weight: 1
  - model: mlabonne/ChimeraLlama-3-8B-v3
    parameters:
      density: 0.66
      weight: 1
merge_method: dare_ties
base_model: NousResearch/Meta-Llama-3-8B
parameters:
  int8_mask: true
dtype: float16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Kukedlc/MergedLlama-3-8B-MS-2"
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"])