solar-kor-resume
Update @ 2024.05.27: First release of Ocelot-Ko-self-instruction-10.8B-v1.0
This model card corresponds to the 10.8B Instruct version of the Solar-Ko model.
The train wad done on A100-80GB
Resources and Technical Documentation:
Citation
@misc {cpm-ai/Ocelot-Ko-self-instruction-10.8B-v1.0,
author = { {frcp, nebchi, pepperonipizza97} },
title = { solar-kor-resume},
year = 2024,
url = { https://huggingface.co/cpm-ai/Ocelot-Ko-self-instruction-10.8B-v1.0 },
publisher = { Hugging Face }
}
Model Developers: frcp, nebchi, pepperonipizza97
Model Information
Resume Proofreading and evaluation of inputs and outputs.
Description
It has been trained with a large amount of Korean tokens compared to other LLMs, enabling it to generate high-quality Korean text.
Model Architecture Solar is an auto-regressive language model that is scaled using the DUS method.
*You can find dataset list here: https://huggingface.co/datasets/cpm-ai/gpt-self-introduction-all
Inputs and outputs
- Input: Text string, such as a question, a prompt, or a document to be Proofreaded.
- Output: Generated Korea text in response to the input, such as an answer to a question, or a evaluation of a resume.
Running the model on a single / multi GPU
# pip install accelerate, flash_attn, sentencepiece
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("cpm-ai/Ocelot-Ko-self-instruction-10.8B-v1.0")
model = AutoModelForCausalLM.from_pretrained("cpm-ai/Ocelot-Ko-self-instruction-10.8B-v1.0", device_map="auto")
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, max_new_tokens=4096, streamer=streamer)
text = λλ μκΈ°μκ°μ 첨μ μ λ¬Έκ°μΌ.
μ£Όμ΄μ§ μκΈ°μκ°μλ₯Ό 첨μν΄μ λ€μ μμ±ν΄μΌν΄.
μΆλ ₯νμμ λ€μμ μ§μΌμΌν΄.
[첨μ]
λ€μμ΄ μκΈ°μκ°μμΌ :
[μ λ μ΄λ¦° μμ λΆν° μλ²½μ£Όμμ μΈ μ±κ²©μ κ°μ§κ³ μμμ΅λλ€. μ΄λ‘ μΈν΄ νμ μμ μ λ₯λ ₯μ λν λΆμκ°μ λλΌλ©° κ³Όλν μ€νΈλ μ€λ₯Ό λ°μμμ΅λλ€. νμ°½ μμ μλ κ³Όμ λ νλ‘μ νΈλ₯Ό μλ²½νκ² λ§λ¬΄λ¦¬νμ§ λͺ»νλ©΄ μμ‘΄κ°μ΄ ν¬κ² νλ€λ Έμ΅λλ€. μ€νκ΅ μμ μλ ν κ°μ§ λ¬Έμ μ λ무 μ€λ μκ°μ ν¬μνμ¬ λ€λ₯Έ νμ΅ κΈ°νλ₯Ό λμΉκΈ°λ νμ΅λλ€. μ΄λ¬ν κ²½νλ€μ μ μκ² μλ²½ν¨μ μΆκ΅¬νλ κ²μ΄ μ’
μ’
νμ€μ λΆμ ν©νλ€λ κ²μ κΉ¨λ¬κ² νμ΅λλ€.
κ³ λ±νκ΅μ λνκ΅μ μ§ννλ©΄μλ μ΄λ¬ν μλ²½μ£Όμμ μΈ μ±κ²©μ 극볡νκΈ° μν΄ λ
Έλ ₯νμ΅λλ€. νμ§λ§ μ¬μ ν μ€ν¨λ₯Ό λ°μλ€μ΄λ κ²μ΄ μ΄λ ΅κ³ , μμ μ νκ³λ₯Ό μΈμ νλ κ²μ΄ μ΄λ €μ μ΅λλ€. μ΄λ¬ν κ³Όμ μ ν΅ν΄ μλ²½ν¨μ λν κ°λ°μ΄ μ μ μ±μ₯κ³Όμ μ μ μ½νλ μμΈμ΄ λμμμ κΉ¨λ¬μμ΅λλ€.]"""
messages = [
{
"role": "user",
"content": "{}".format(text)
}
]
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipe(
prompt,
temperature=0.2,
add_special_tokens=True
)
print(outputs[0]["generated_text"][len(prompt):])
results
[첨μ]
μ΄λ¦° μμ λΆν° μ λ μλ²½ν κ²°κ³Όλ₯Ό μΆκ΅¬νλ©° μ€μ€λ‘λ₯Ό μλ°ν΄μ¨ μ±κ²©μ΄μμ΅λλ€. μ΄λ νμ
κ³Ό κ΄λ ¨λ μ€νΈλ μ€λ‘ μ΄μ΄μ Έ, κ³Όμ λ₯Ό μμνλλΌλ λ§μ‘±λ³΄λ€λ λΆλ§μ‘±μ κ°μ μ΄ λ μ»Έλ μκΈ°μμ΅λλ€. νΉν μ€νκ΅ λ ν λ¬Έμ λ₯Ό ν΄κ²°νκΈ° μν΄ μ§λμΉκ² μ€λ«λμ 맀λ¬λ € κ²°κ΅ μ€μν μκΈ°λ₯Ό λμΉ κ²½νμ μ μ±μ₯μ ν° μν₯μ λ―Έμ³€μ΅λλ€. μ΄ κ³Όμ μμ μλ²½μ£Όμλ₯Ό μΆκ΅¬νλ κ²μ΄ νμ€μ μ΄μ§ μμ μ μλ€λ μ¬μ€μ κΉ¨λ«κΈ° μμνμ΅λλ€.
κ³ λ±νκ΅μ λνμμλ μ΄λ¬ν μ±ν₯μ κ°μ νκ³ μ λ€μν λ
Έλ ₯μ κΈ°μΈμμ΅λλ€. μλ₯Ό λ€μ΄, λͺ©νλ₯Ό μΈλΆννκ³ λ¨κ³λ³λ‘ μ κ·Όνλ©΄μ μ±μ·¨κ°κ³Ό μμ κ°μ ν€μ°κΈ° μν΄ λ
Έλ ₯νμ΅λλ€. λν, ν νλ‘μ νΈμμ μν μ λΆλ΄νκ³ νλ ₯ν¨μΌλ‘μ¨ κ°μΈμ νκ³λ³΄λ€ μ 체 μ±κ³Όλ₯Ό μ°μ μνλ λ²μ λ°°μ μ΅λλ€. λΉλ‘ μμ§ μλ²½ν¨μ΄λΌλ κ΅΄λ λ‘λΆν° μμ ν μμ λ‘μμ§μ§λ λͺ»νμ§λ§, μ΄λ₯Ό 극볡νκ³ μ±μ₯ν μ μλ λ°©λ²μ μ°Ύμλ€λ μ μμ μλΆμ¬μ λλλλ€.
Evaluation Results - LogicKor
Model | κΈμ°κΈ° | μ΄ν΄ | λ¬Έλ² |
---|---|---|---|
HyperClovaX | 8.50 | 9.50 | 8.50 |
solar-1-mini-chat | 8.50 | 7.00 | 5.21 |
allganize/Llama-3-Alpha-Ko-8B-Instruct | 8.50 | 8.35 | 4.92 |
Synatra-kiqu-7B | 4.42 | 5.71 | 4.50 |
Ocelot-ko-10.8B | 8.57 | 7.00 | 6.57 |
Evaluation Results - Kobest
λͺ¨λΈ λͺ μΉ | Average n=0 n=5 |
HellaSwag n=0 n=5 |
COPA n=0 n=5 |
BooIQ n=0 n=5 |
---|---|---|---|---|
KoGPT | 58.2 63.7 | 55.9 58.3 | 73.5 72.9 | 45.1 59.8 |
Polyglot-ko-13B | 62.4 68.2 | 59.5 63.1 | 79.4 81.1 | 48.2 60.4 |
LLaMA 2-13B | 45.2 60.5 | 41.3 44.0 | 59.3 63.8 | 34.9 73.8 |
Baichuan 2-13B | 52.7 53.9 | 39.2 39.6 | 60.6 60.6 | 58.4 61.5 |
QWEN-14B | 47.8 66.4 | 45.3 46.8 | 64.9 68.9 | 33.4 83.5 |
Orion-14B-Chat | 68.8 73.2 | 47.0 49.6 | 77.7 79.4 | 81.6 90.7 |
Ocelot-ko-10.8B | 72.5 75.9 | 50.0 51.4 | 75.8 82.5 | 91.7 93.8 |
Software
Training was done using QLoRA
- Downloads last month
- 50
Inference Providers
NEW
This model is not currently available via any of the supported third-party Inference Providers, and
the model is not deployed on the HF Inference API.