StrangeMerges_49-7B-dare_ties
StrangeMerges_49-7B-dare_ties is a merge of the following models using LazyMergekit:
𧩠Configuration
models:
- model: Gille/StrangeMerges_32-7B-slerp
parameters:
weight: 0.4
density: 0.6
- model: AurelPx/Percival_01-7b-slerp
parameters:
weight: 0.4
density: 0.55
- model: louisbrulenaudet/Pearl-7B-slerp
parameters:
weight: 0.2
density: 0.5
base_model: Gille/StrangeMerges_47-7B-dare_ties
merge_method: dare_ties
dtype: bfloat16
π» Usage
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "Gille/StrangeMerges_49-7B-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. | 75.50 |
AI2 Reasoning Challenge (25-Shot) | 72.35 |
HellaSwag (10-Shot) | 88.30 |
MMLU (5-Shot) | 64.31 |
TruthfulQA (0-shot) | 74.70 |
Winogrande (5-shot) | 83.74 |
GSM8k (5-shot) | 69.60 |
- Downloads last month
- 58
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_49-7B-dare_ties
Merge model
this model
Spaces using Gille/StrangeMerges_49-7B-dare_ties 3
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard72.350
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard88.300
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard64.310
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard74.700
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard83.740
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard69.600