| license: cc-by-nc-nd-4.0 | |
| base_model: skt/kogpt2-base-v2 | |
| tags: | |
| - gpt2 | |
| - lora | |
| - korean | |
| - chatbot | |
| language: | |
| - ko | |
| # ๋ชจ๋ธ ์ด๋ฆ | |
| skt/kogpt2-base-v2 | |
| ## ๋ชจ๋ธ ์ค๋ช | |
| ํ๊ตญ์ด๋ก ํ์ต๋ ์คํ์์ค ๊ธฐ๋ฐ GPT-2 ๋ชจ๋ธ | |
| ## ๋ชจ๋ธ ์์ธ | |
| ์ฑ๋ด ๊ตฌ์ถ, ํ ์คํธ ๊ฐ์ฑ ์์ธก, ํ ์คํธ ๋ถ์ ๊ธฐ๋ฐ ์๋ต ์์ฑ์ ์ฌ์ฉ๋ ์ ์์ผ๋ฉฐ, ๊ด์ฌ ์๋ ๊ฐ๋ฐ์๋ ๋ชจ๋ธ๊ณผ ๊ด๋ จ ์์ค๋ฅผ ๋ค์ด๋ก๋ํด ํ๋ก์ ํธ์ ์ ์ฉํ๊ฑฐ๋ ์์ ํ๋ฉด ๋๋ค. | |
| ## LoRA ์ค์ | |
| lora_config = LoraConfig( | |
| r=16, | |
| lora_alpha=32, | |
| target_modules=["c_attn", "c_proj", "c_fc"], | |
| lora_dropout=0.05, | |
| bias="none", | |
| task_type=TaskType.CAUSAL_LM | |
| ) | |
| ## ํ์ต ์ค์ | |
| ## ํ์ต ๊ฒฐ๊ณผ | |
| ## ์ฌ์ฉ ๋ฐฉ๋ฒ | |
| ```python | |
| ``` | |
| ## ๋ชจ๋ธ ์ ๋ณด | |