G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment
Paper • 2303.16634 • Published • 3
How to use Trofish/KULLM-RLHF with Transformers:
# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("text-generation", model="Trofish/KULLM-RLHF") # Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("Trofish/KULLM-RLHF")
model = AutoModelForCausalLM.from_pretrained("Trofish/KULLM-RLHF")How to use Trofish/KULLM-RLHF with vLLM:
# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Trofish/KULLM-RLHF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
-H "Content-Type: application/json" \
--data '{
"model": "Trofish/KULLM-RLHF",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'docker model run hf.co/Trofish/KULLM-RLHF
How to use Trofish/KULLM-RLHF with SGLang:
# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
--model-path "Trofish/KULLM-RLHF" \
--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": "Trofish/KULLM-RLHF",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'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 "Trofish/KULLM-RLHF" \
--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": "Trofish/KULLM-RLHF",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'How to use Trofish/KULLM-RLHF with Docker Model Runner:
docker model run hf.co/Trofish/KULLM-RLHF
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
2023 성균관대 하계집중 산학협력프로젝트 VAIV
GPT-NEOX(Polyglot-ko) 기반 자연스럽고 윤리적인 한국어 기반 일상 대화형 챗봇 모델 구현
Self-Instruct: GPT4를 이용한 데이터 증강
RLHF(Reinforcement Learning from Human Feedback): 사람의 선호도를 반영한 강화학습
DeepSpeed: 대규모 분산 딥러닝을 위한 새로운 메모리 최적화 기술
- 고려대학교 NLP & AI 연구실과 HIAI 연구소가 개발한 한국어 LLM "KULLM" 사용