Instructions to use Azazelle/Sina-Odin-7b-Merge with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Azazelle/Sina-Odin-7b-Merge with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Azazelle/Sina-Odin-7b-Merge")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Azazelle/Sina-Odin-7b-Merge")
model = AutoModelForCausalLM.from_pretrained("Azazelle/Sina-Odin-7b-Merge")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Azazelle/Sina-Odin-7b-Merge with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Azazelle/Sina-Odin-7b-Merge"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Azazelle/Sina-Odin-7b-Merge",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Azazelle/Sina-Odin-7b-Merge

SGLang

How to use Azazelle/Sina-Odin-7b-Merge with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Azazelle/Sina-Odin-7b-Merge" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Azazelle/Sina-Odin-7b-Merge",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Azazelle/Sina-Odin-7b-Merge" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Azazelle/Sina-Odin-7b-Merge",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Azazelle/Sina-Odin-7b-Merge with Docker Model Runner:
```
docker model run hf.co/Azazelle/Sina-Odin-7b-Merge
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Model Card for Sina-Odin-7b-Merge

Part of a series of experimental DARE merges.

.yaml file for mergekit

models:
  - model: Mihaiii/Metis-0.3
    # no parameters necessary for base model
  - model: rishiraj/smol-7b #75
    parameters:
      weight: 0.2
      density: 0.41
  - model: SanjiWatsuki/openchat-3.5-1210-starling-slerp #125
    parameters:
      weight: 0.33
      density: 0.54
  - model: Azazelle/Dumb-Maidlet #200
    parameters:
      weight: 0.53
      density: 0.71
merge_method: dare_ties
base_model: Mihaiii/Metis-0.3
parameters:
  int8_mask: true
dtype: bfloat16

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	47.82
AI2 Reasoning Challenge (25-Shot)	52.82
HellaSwag (10-Shot)	68.86
MMLU (5-Shot)	45.54
TruthfulQA (0-shot)	39.20
Winogrande (5-shot)	72.22
GSM8k (5-shot)	8.26

Downloads last month: 188

Safetensors

Model size

7B params

Tensor type

BF16

Model tree for Azazelle/Sina-Odin-7b-Merge

Quantizations

2 models

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

52.820
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

68.860
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

45.540
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

39.200
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

72.220
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

8.260