README.md

metadata

tags:
  - merge
  - mergekit
  - lazymergekit
  - udkai/Turdus
  - CultriX/MergeTrix-7B-v2
  - mlabonne/NeuralBeagle14-7B
base_model:
  - udkai/Turdus
  - CultriX/MergeTrix-7B-v2
  - mlabonne/NeuralBeagle14-7B
license: cc-by-nc-4.0

Open-LLM Benchmark Results:

TurdusTrixBeagle-DARETIES-7B 📑 Average: 75.2 ARC: 73.46 HellaSwag: 88.61 MMLU: 64.89 TruthfulQA: 68.81 Winogrande: 85.16 GSM8K: 70.28

TurdusTrixBeagle-DARETIES-7B

TurdusTrixBeagle-DARETIES-7B is a merge of the following models using LazyMergekit:

• udkai/Turdus
• CultriX/MergeTrix-7B-v2
• mlabonne/NeuralBeagle14-7B

🧩 Configuration

models:
  - model: mistralai/Mistral-7B-v0.1
    # No parameters necessary for base model
  - model: udkai/Turdus
    parameters:
      density: 0.65
      weight: 0.4
  - model: CultriX/MergeTrix-7B-v2
    parameters:
      density: 0.45
      weight: 0.3
  - model: mlabonne/NeuralBeagle14-7B
    parameters:
      density: 0.55
      weight: 0.3
merge_method: dare_ties
base_model: mistralai/Mistral-7B-v0.1
parameters:
  int8_mask: true
dtype: float16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "jsfs11/TurdusTrixBeagle-DARETIES-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"])