Instructions to use byeolki/Llama-KoEmpathy with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use byeolki/Llama-KoEmpathy with Transformers:

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

pipe = pipeline("text-generation", model="byeolki/Llama-KoEmpathy")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("byeolki/Llama-KoEmpathy")
model = AutoModelForCausalLM.from_pretrained("byeolki/Llama-KoEmpathy")

llama-cpp-python

How to use byeolki/Llama-KoEmpathy with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="byeolki/Llama-KoEmpathy",
	filename="unsloth.Q8_0.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Inference
Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use byeolki/Llama-KoEmpathy with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf byeolki/Llama-KoEmpathy:Q8_0
# Run inference directly in the terminal:
llama-cli -hf byeolki/Llama-KoEmpathy:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf byeolki/Llama-KoEmpathy:Q8_0
# Run inference directly in the terminal:
llama-cli -hf byeolki/Llama-KoEmpathy:Q8_0

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf byeolki/Llama-KoEmpathy:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf byeolki/Llama-KoEmpathy:Q8_0

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf byeolki/Llama-KoEmpathy:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf byeolki/Llama-KoEmpathy:Q8_0

Use Docker

docker model run hf.co/byeolki/Llama-KoEmpathy:Q8_0

LM Studio
Jan

vLLM

How to use byeolki/Llama-KoEmpathy with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/byeolki/Llama-KoEmpathy:Q8_0

SGLang

How to use byeolki/Llama-KoEmpathy 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 "byeolki/Llama-KoEmpathy" \
    --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": "byeolki/Llama-KoEmpathy",
		"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 "byeolki/Llama-KoEmpathy" \
        --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": "byeolki/Llama-KoEmpathy",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use byeolki/Llama-KoEmpathy with Ollama:
```
ollama run hf.co/byeolki/Llama-KoEmpathy:Q8_0
```

Unsloth Studio new

How to use byeolki/Llama-KoEmpathy with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for byeolki/Llama-KoEmpathy to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for byeolki/Llama-KoEmpathy to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for byeolki/Llama-KoEmpathy to start chatting

Docker Model Runner
How to use byeolki/Llama-KoEmpathy with Docker Model Runner:
```
docker model run hf.co/byeolki/Llama-KoEmpathy:Q8_0
```

Lemonade

How to use byeolki/Llama-KoEmpathy with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull byeolki/Llama-KoEmpathy:Q8_0

Run and chat with the model

lemonade run user.Llama-KoEmpathy-Q8_0

List all available models

lemonade list

🤗 Llama-KoEmpathy 💝

Llama-KoEmpathy는 Llama-3.1-8B를 AIHub 공감형대화 데이터셋으로 파인튜닝한 언어 모델입니다. 이 모델은 사용자의 감정을 이해하고 공감하는 대화를 생성하는 것을 목표로 합니다. 💭✨

Built with Llama

✨ Features

LLaMA 아키텍처 기반 한국어 챗봇
감정 인식 및 공감 능력
LoRA를 활용한 효율적인 파인튜닝

⚠️ Licensing Notice

이 모델은 Llama 3.1 Community License에 따라 제공됩니다. 모델 사용 시 다음 사항을 준수해야 합니다:

Meta의 Acceptable Use Policy를 따라야 합니다
월간 활성 사용자가 7억명을 초과하는 제품/서비스에 사용할 경우 Meta의 별도 라이선스가 필요합니다

📋 Model Description

🦙 Base Model: unsloth/Meta-Llama-3.1-8B
📚 Dataset: AIHub 공감형대화
🔧 Training Method: LoRA (r=16, alpha=16)

⚙️ Training Configuration

📏 Max Sequence Length: 2048
📦 Batch Size: 128
🔄 Gradient Accumulation Steps: 4
📈 Learning Rate: 2e-4
🛠️ Optimizer: AdamW 8bit
💾 Quantization: GGUF q8_0
🔁 Training Epochs: 3

TIP: 데이터의 크기나 Epoch을 늘리면 좀 더 좋은 결과를 가져올 수 있을거에요!

💻 Usage

모델을 사용하는 방법은 다음과 같습니다:

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
max_length = 512
temperture = 0.7

model_name = "byeolki/Llama-KoEmpathy"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)
prompt = """아래는 작업을 설명하는 지시사항입니다. 입력된 내용을 바탕으로 적절한 응답을 작성하세요.
### 지시사항:
입력에 대해서 공감해주세요.
### 입력:
{input}
### 응답:
"""
input_text = "안녕하세요. 저 오늘 기분이 안 좋아요."

inputs = tokenizer(prompt.format(input=input_text), return_tensors="pt").to(model.device)
outputs = model.generate(
    **inputs,
    max_length=max_length,
    temperature=temperture,
    top_p=0.9,
    do_sample=True,
    pad_token_id=tokenizer.pad_token_id,
    bos_token_id=tokenizer.bos_token_id,
    eos_token_id=tokenizer.eos_token_id,
    num_return_sequences=1
)
generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
try:
    response = generated_text.split("### 응답:\n")[-1].strip()
except:
    response = generated_text
print(response)

💬 Example Conversations

👤 User: 나 오늘 속상한 일이 있었어.
🤖 Assistant: 무슨 일인데 그래? 기분이 많이 안 좋았나 보네.

👤 User: 나 살이 너무 많이 찐거 같아서 속상해.
🤖 Assistant: 살이 쪄서 속상하다니. 왜 그래?

⚖️ License

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

📝 Notice

Downloads last month: 22

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for byeolki/Llama-KoEmpathy

Base model

meta-llama/Llama-3.1-8B

Quantized

(324)

this model

Quantizations

1 model