Easy Ko Llama3 8b Instruct V1

🚀 Easy-Systems/easy-ko-Llama3-8b-Instruct-v1

Easy-Systems/easy-ko-Llama3-8b-Instruct-v1 is the first LLM model of Easy Systems. It is based on the English model meta-llama/Meta-Llama-3-8B-Instruct and fine-tuned for Korean. The LLM model will be continuously updated in the future.

Image generated by DALL-E.

📚 Documentation

Data

• Data from AI hub (https://www.aihub.or.kr/) was processed for various tasks (QA, Summary, Translation, etc.) and used for fine-tuning.
• In-house processed data was also used for fine-tuning.

💻 Usage Examples

Basic Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Easy-Systems/easy-ko-Llama3-8b-Instruct-v1"

model = AutoModelForCausalLM.from_pretrained(model_id,                                            
                                             attn_implementation="flash_attention_2",
                                             torch_dtype=torch.bfloat16,
                                             device_map="auto")

tokenizer = AutoTokenizer.from_pretrained(model_id, add_special_tokens=True)

prompt="리눅스 프로세스를 강제로 종료하는 방법은?"
messages = [  
    {"role": "system", "content": "당신은 친절한 AI chatbot 입니다. 요청에 대해서 step-by-step 으로 간결하게 한국어(Korean)로 답변해주세요."},
    {"role": "user", "content": f"\n\n### 명령어: {prompt}\n\n### 응답:"}   
]

input_ids = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

terminators = [
    tokenizer.eos_token_id,
    tokenizer.convert_tokens_to_ids("<|eot_id|>")
]

outputs = model.generate(
    input_ids,
    max_new_tokens=1024,
    eos_token_id=terminators,
    pad_token_id=tokenizer.eos_token_id,
    do_sample=True,
    temperature=0.2,
    repetition_penalty = 1.3,
    top_p=0.9,
    top_k=10,
)
response = outputs[0][input_ids.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True).strip())

Example Output

In Linux, you can use the `kill` or `pkill` commands to forcefully terminate a specific process.

Step 1: Use `ps -ef | grep <process_name>` to display all currently running processes.
Step 2: Enter `kill <process_ID>` to immediately terminate the corresponding process.

Alternatively, you can specify `-9` (SIGKILL signal) to forcefully terminate the process, which will kill it immediately without giving it a last chance to terminate normally in the operating system:

Step 3: Enter `kill -9 <process_ID>`.

Note that if there are files or services necessary for the system's stability, you should not delete them directly and should handle them according to appropriate permissions and instructions. Also, some programs may have issues such as data loss when forcefully terminated, so please check your saved work before terminating. 

📄 License

• Creative Commons Attribution-NonCommercial-ShareAlike 4.0 (CC-BY-NC-SA-4.0)
• For commercial use, please contact us at the following address.

Contact

• If you want to contact us for commercial use or other inquiries, please send an email to the following address.
• Hyung-Koo Kang: hkkang@easy.co.kr