Instructions to use dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach with Transformers:

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

pipe = pipeline("text-generation", model="dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach")
model = AutoModelForCausalLM.from_pretrained("dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach

SGLang

How to use dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach 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 "dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach with Docker Model Runner:
```
docker model run hf.co/dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Explanation

With the base model, applied DPO to the small amount of layers with the open dataset , saved just the adapter part
Merged the base model and the tuned adapter together

Base Model

beomi/OPEN-SOLAR-KO-10.7B

Used Corpus

maywell/ko_Ultrafeedback_binarized

Score

Average	Ko-ARC	Ko-HellaSwag	Ko-MMLU	Ko-TruthfulQA	Ko-CommonGen V2
52.83	50	60.55	48.8	71.51	43.65

Log

2024.01.25: Initial version Upload
2024.02.10: Readme updated
2024.02.11: Score updated

LICENSE

Apache 2.0

Citation

beomi/OPEN-SOLAR-KO-10.7B

@misc {solar_ko_junbum_2023,
    author       = { {L. Junbum} },
    title        = { Solar-Ko-10.7b },
    year         = 2024,
    url          = { https://huggingface.co/beomi/SOLAR-KO-10.7B },
    publisher    = { Hugging Face }
}

Downloads last month: 10

Safetensors

Model size

11B params

Tensor type

F16

Model tree for dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach

Quantizations

1 model

dddsaty
/

OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach

Model tree for dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach

Dataset used to train dddsaty/OPEN_SOLAR_KO_10.7B_DPO_Adapter_Attach