Octen-Embedding-0.6B-GGUF

import numpy as np
import torch
from llama_cpp import Llama
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim


model = SentenceTransformer(
    "Octen/Octen-Embedding-0.6B",
    model_kwargs={"dtype": torch.bfloat16},
)
llama = Llama.from_pretrained(
    repo_id="mykor/Octen-Embedding-0.6B-GGUF",
    filename="Octen-Embedding-0.6B-BF16.gguf",
    verbose=False,
    embedding=True,
    n_ctx=0,
)

text = """์ฒ˜์Œ์€ ๋ฏธ์•ฝํ•œ ๋ฌผ๊ฒฐ์ด๊ฒ ์ง€๋งŒ
๊ทธ๋ ‡๊ฒŒ ๋„ˆ์˜ ๊ฟˆ์— ๋‹ฟ์„ ์ˆ˜ ์žˆ๊ฒ ์ง€
๊ฑฐ๋Œ€ํ•œ ๋ฐ”๋‹ค์— ๊ฐ‡ํ˜€๋ฒ„๋ฆฐ ๋„ˆ๋ฅผ ๊ตฌํ•  ๋ฐฉ๋ฒ•์„ ์•Œ๊ณ  ์žˆ์–ด
์ ˆ๋Œ€๋กœ ๋ฉˆ์ถ”์ง€ ๋งˆ, ๋น›๋‚˜๊ณ  ์žˆ์œผ๋‹ˆ

์•„, ์‹ ์ด ๋ณธ๋‹ค๋ฉด, ์•„, ํ›”์น˜๊ณ  ๋ง๊ฑธ
๋ฏฟ์ง€ ์•Š๋Š” ๊ฒƒ๋“ค์€ ๋ฌด์‹œํ•ด
์ปค๋‹ค๋ž€ ํŒŒ๋™์„ ์ผ์œผํ‚ค๋Š”
๊ทธ ํž˜์„ ๋„ˆ๋„ ๊ฐ€์ง€๊ณ  ์žˆ์–ด, ๋‚ด๊ฒŒ ๋ณด์—ฌ์ค„๋ž˜

ํŒŒ๋ž€์„ ์ผ์œผ์ผœ ๋ˆˆ๋ถ€์‹  ๋„ ๋ณด์—ฌ์ค˜
์ด ๋ฐ”๋‹ค๋ฅผ ์ง‘์–ด์‚ผํ‚ค์ž
ํŒŒ๋„์˜ ์ƒ‰์€ ํŒŒ๋ž‘์ด ์•„๋‹ˆ์•ผ, ๋ฌผ๋“ค์ด์ž ์„œ๋กœ์˜
์ƒ‰์œผ๋กœ ๋น›๋‚˜๋Š” ๋ฐ”๋‹ค๋ฅผ (๋ฐ”๋‹ค๋ฅผ) ๋งŒ๋“ค์–ด ๊ฐ€๋ณด์ž

์‰ฟ, ์กฐ์šฉํ•œ ์ด๊ณณ์— ๋ˆ„๊ตฐ๊ฐ€ ์žˆ์–ด
ํ•˜๋Š˜์˜ ํ•ด๋„ ํƒ๋‚ด๊ณ  ์žˆ์–ด
๋ˆˆ์น˜์ฑ˜๋‹ค๋ฉด, ์–ผ๋ฅธ ํ—ค์—„์ณ
ํ•ด๋ฅผ ์‚ผํ‚ฌ ํฐ ๋ฌผ๊ฒฐ์„ ๋งŒ๋“ค์ž

ํƒ๋‚˜๋Š” ๋น›์„ ๊ฐ€์ง„ ๋„ˆ๋งˆ์ €
๋„ ์˜์‹ฌํ•˜๊ฒŒ ๋  ํ…Œ์ง€๋งŒ ๋ฏฟ์–ด
๊ฐ€๋Šฅ์„ฑ ๊ทธ๊ฑฐ๋ฉด ๋ผ
์œ ์ผํ•œ ์ƒ‰์„ ๊ฐ€์ง€๊ณ  ์žˆ์œผ๋‹ˆ

์•„, ์‹ ์ด ์•Œ์•„๋„, ์•„, ๊ฐ€์งˆ ์ˆ˜ ์—†์–ด
๋ฏฟ์ง€ ์•Š๋Š” ๊ฒƒ ๋“ค์€ ๋ฌด์‹œํ•ด
์ปค๋‹ค๋ž€ ํŒŒ๋™์„ ์ผ์œผํ‚ค๋Š”
๊ทธ ํž˜์„ ๋„ˆ๋„ ๊ฐ€์ง€๊ณ  ์žˆ์–ด, ๋‚ด๊ฒŒ ๋ณด์—ฌ์ค„๋ž˜

ํŒŒ๋ž€์„ ์ผ์œผ์ผœ ๋ˆˆ๋ถ€์‹  ๋„ ๋ณด์—ฌ์ค˜
์ด ๋ฐ”๋‹ค๋ฅผ ์ง‘์–ด์‚ผํ‚ค์ž
ํŒŒ๋„์˜ ์ƒ‰์€ ํŒŒ๋ž‘์ด ์•„๋‹ˆ์•ผ, ๋ฌผ๋“ค์ด์ž ์„œ๋กœ์˜
์ƒ‰์œผ๋กœ ๋น›๋‚˜๋Š” ๋ฐ”๋‹ค๋ฅผ (๋ฐ”๋‹ค๋ฅผ) ๋งŒ๋“ค์–ด ๊ฐ€๋ณด์ž

๋„Œ ์ •๋ง๋กœ ์–ด๋ฆฌ์„๊ตฌ๋‚˜ ๋‚ด ํ’ˆ์„ ๋– ๋‚˜
๋” ์ปค๋‹ค๋ž€ ๋ฏธ๋ž˜๋ฅผ ๊ฟˆ๊ฟ€ ์ˆ˜ ์žˆ๋‹ค๊ณ  ๋ฏฟ๋‹ˆ ๋ˆ„๊ฐ€, ๊ทธ๋ž˜?

ํŒŒ๋ž€์„ ์ผ์œผ์ผœ ์‹ฌ์—ฐ์„ ๊นจ๊ณ  ๋‚˜์™€
์ด ๋ฐ”๋‹ค๋ฅผ ์ง‘์–ด์‚ผํ‚ค์ž
ํŒŒ๋„์˜ ์ƒ‰์€ ํŒŒ๋ž‘์ด ์•„๋‹ˆ์•ผ, ๋ฌผ๋“ค์ด์ž ์šฐ๋ฆฌ์˜
์ƒ‰์œผ๋กœ ๋น›๋‚˜๋Š” ๋ฐ”๋‹ค๋ฅผ (๋ฐ”๋‹ค๋ฅผ) ๋ณด์—ฌ์ฃผ๋Ÿฌ ๊ฐ€์ž"""


embed1 = model.encode(text)
embed2 = np.array(llama.embed(text), dtype=np.float32)
print(cos_sim(embed1, embed2).item())
0.9998425245285034
Downloads last month
498
GGUF
Model size
0.6B params
Architecture
qwen3
Hardware compatibility
Log In to add your hardware

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for mykor/Octen-Embedding-0.6B-GGUF

Quantized
(11)
this model