StrangeMerges_9-7B-dare_ties
StrangeMerges_9-7B-dare_ties is a merge of the following models using LazyMergekit:
🧩 Configuration
models:
- model: Gille/StrangeMerges_8-7B-slerp
# no parameters necessary for base model
- model: leveldevai/TurdusBeagle-7B
parameters:
density: 0.5
weight: 0.4
- model: samir-fama/FernandoGPT-v1
parameters:
density: 0.5
weight: 0.6
merge_method: dare_ties
base_model: Gille/StrangeMerges_8-7B-slerp
parameters:
normalize: true
dtype: float16
💻 Usage
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "Gille/StrangeMerges_9-7B-dare_ties"
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. | 73.32 |
AI2 Reasoning Challenge (25-Shot) | 70.31 |
HellaSwag (10-Shot) | 87.46 |
MMLU (5-Shot) | 65.08 |
TruthfulQA (0-shot) | 65.08 |
Winogrande (5-shot) | 81.37 |
GSM8k (5-shot) | 70.58 |
- Downloads last month
- 72
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 Gille/StrangeMerges_9-7B-dare_ties
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard70.310
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard87.460
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard65.080
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard65.080
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard81.370
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard70.580