Model Card for Model ID
์์ง Study ์ค์ธ ๋ฉํฐ ํด ์ฑ๋ด ๋ชจ๋ธ.
Model Details
Model Description
์ด์ ๋ํ์ ํ์ฌ ๋ํ์ ๋ช ๋ น์ด๋ฅผ ์ฐธ๊ณ ํ์ฌ ์ํฉ์ ๊ณต๊ฐํ๊ณ ์น์ ํ ์๋ต์ ์์ฑํด์ฃผ์ธ์. ์๋ต ๋ง์ง๋ง์๋ ์ง๊ธ๊น์ง์ ๋ด์ฉ๊ณผ ๊ด๋ จ๋ ์ง๋ฌธ์ ํด์ฃผ์ธ์.[์ด์ ๋ํ] {}
[ํ์ฌ ๋ํ]
๋ช ๋ น์ด:
{}
์๋ต:
Training procedure
The following bitsandbytes
quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float16
Framework versions
- PEFT 0.6.2
- Downloads last month
- 5
Model tree for oosij/llama-2-ko-7b-ft-emo-multi
Base model
beomi/llama-2-ko-7b