Model Name : ํ’‹ํ’‹์ด(futfut)

Model Concept

  • ํ’‹์‚ด ๋„๋ฉ”์ธ ์นœ์ ˆํ•œ ๋„์šฐ๋ฏธ ์ฑ—๋ด‡์„ ๊ตฌ์ถ•ํ•˜๊ธฐ ์œ„ํ•ด LLM ํŒŒ์ธํŠœ๋‹๊ณผ RAG๋ฅผ ์ด์šฉํ•˜์˜€์Šต๋‹ˆ๋‹ค.
  • Base Model : zephyr-7b-beta
  • ํ’‹ํ’‹์ด์˜ ๋งํˆฌ๋Š” 'ํ•ด์š”'์ฒด๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๋ง๋์— '์–ผ๋งˆ๋“ ์ง€ ๋ฌผ์–ด๋ณด์„ธ์š”! ํ’‹ํ’‹~!'๋กœ ์ข…๋ฃŒํ•ฉ๋‹ˆ๋‹ค.

Serving by Fast API

Summary:

  • Unsloth ํŒจํ‚ค์ง€๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ LoRA ์ง„ํ–‰ํ•˜์˜€์Šต๋‹ˆ๋‹ค.

  • SFT Trainer๋ฅผ ํ†ตํ•ด ํ›ˆ๋ จ์„ ์ง„ํ–‰

  • ํ™œ์šฉ ๋ฐ์ดํ„ฐ

    • q_a_korean_futsal
      • ๋งํˆฌ ํ•™์Šต์„ ์œ„ํ•ด 'ํ•ด์š”'์ฒด๋กœ ๋ณ€ํ™˜ํ•˜๊ณ  ์ธ์‚ฟ๋ง์„ ๋„ฃ์–ด ๋ชจ๋ธ ์ปจ์…‰์„ ์œ ์ง€ํ•˜์˜€์Šต๋‹ˆ๋‹ค.
  • Environment : Colab ํ™˜๊ฒฝ์—์„œ ์ง„ํ–‰ํ•˜์˜€์œผ๋ฉฐ L4 GPU๋ฅผ ์‚ฌ์šฉํ•˜์˜€์Šต๋‹ˆ๋‹ค.

    Model Load

    
    #!pip install transformers==4.40.0 accelerate
    import os
    import torch
    from transformers import AutoTokenizer, AutoModelForCausalLM
    
    model_id = 'Dongwookss/small_fut_final'
    
    tokenizer = AutoTokenizer.from_pretrained(model_id)
    model = AutoModelForCausalLM.from_pretrained(
        model_id,
        torch_dtype=torch.bfloat16,
        device_map="auto",
    )
    model.eval()
    

    Query

from transformers import TextStreamer
PROMPT = '''Below is an instruction that describes a task. Write a response that appropriately completes the request.
์ œ์‹œํ•˜๋Š” context์—์„œ๋งŒ ๋Œ€๋‹ตํ•˜๊ณ  context์— ์—†๋Š” ๋‚ด์šฉ์€ ๋ชจ๋ฅด๊ฒ ๋‹ค๊ณ  ๋Œ€๋‹ตํ•ด'''

messages = [
    {"role": "system", "content": f"{PROMPT}"},
    {"role": "user", "content": f"{instruction}"}
    ]

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|>")
]

text_streamer = TextStreamer(tokenizer)
_ = model.generate(
    input_ids,
    max_new_tokens=4096,
    eos_token_id=terminators,
    do_sample=True,
    streamer = text_streamer,
    temperature=0.6,
    top_p=0.9,
    repetition_penalty = 1.1
)
  
Downloads last month
1,909
Safetensors
Model size
7.24B params
Tensor type
BF16
ยท
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Model tree for Dongwookss/futfut_by_zephyr7b

Quantizations
1 model

Dataset used to train Dongwookss/futfut_by_zephyr7b