OmniBeagle-7B
OmniBeagle-7B is a merge of the following models using LazyMergekit:
𧩠Configuration
models:
- model: mistralai/Mistral-7B-v0.1
# no parameters necessary for base model
- model: shadowml/BeagleSempra-7B
parameters:
density: 0.65
weight: 0.4
- model: shadowml/BeagSake-7B
parameters:
density: 0.6
weight: 0.35
- model: shadowml/WestBeagle-7B
parameters:
density: 0.6
weight: 0.35
merge_method: dare_ties
base_model: mistralai/Mistral-7B-v0.1
parameters:
int8_mask: true
dtype: float16
π» Usage
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "mlabonne/OmniBeagle-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"])
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
Metric | Value |
---|---|
Avg. | 75.66 |
AI2 Reasoning Challenge (25-Shot) | 72.61 |
HellaSwag (10-Shot) | 88.93 |
MMLU (5-Shot) | 64.80 |
TruthfulQA (0-shot) | 74.45 |
Winogrande (5-shot) | 83.11 |
GSM8k (5-shot) | 70.05 |
- Downloads last month
- 184
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for mlabonne/OmniBeagle-7B
Merge model
this model
Space using mlabonne/OmniBeagle-7B 1
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard72.610
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard88.930
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard64.800
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard74.450
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard83.110
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard70.050