bleagle-7b-v0.1-test

bleagle-7b-v0.1-test is a merge of the following models using LazyMergekit:

• eren23/slerp-test-turdus-beagle
• udkai/Turdus
• 222gate/BrurryDog-7b-v0.1

🧩 Configuration

models:
  - model: eren23/slerp-test-turdus-beagle
    parameters:
      density: [1, 0.7, 0.1] # density gradient
      weight: 1.0
  - model: udkai/Turdus
    parameters:
      density: 0.5
      weight: [0, 0.3, 0.7, 1] # weight gradient
  - model: 222gate/BrurryDog-7b-v0.1
    parameters:
      density: 0.33
      weight:
        - filter: mlp
          value: 0.5
        - value: 0
merge_method: dare_ties
base_model: leveldevai/MarcBeagle-7B
parameters:
  normalize: true
  int8_mask: true
dtype: bfloat16
embed_slerp: true
tokenizer_source: union

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "222gate/bleagle-7b-v0.1-test"
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"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	73.89
AI2 Reasoning Challenge (25-Shot)	72.27
HellaSwag (10-Shot)	88.24
MMLU (5-Shot)	64.37
TruthfulQA (0-shot)	67.83
Winogrande (5-shot)	85.48
GSM8k (5-shot)	65.13