Llama-3.1-8B-Instruct-Chinese-Elite
ๆ่ฆ
ๅบไบ Llama-3.1-8B-Instruct ็ไธญ่ฑๅ่ฏญๆไปคๅพฎ่ฐๆจกๅใๆฒฟ็จ Llama-3.2-3B-Elite ็่ฎญ็ปๆนๆณ๏ผ Qwen-3-235b-a22b-Instruct-2507 ๆๅธๆจกๅ่ธ้ฆ + SFT๏ผ๏ผไฝๆๆๅๅฐ่กจๆ
็ฌฆๅท๏ผๅๆถไฟ็ๅนถๅผบๅ่พๅบๆ ผๅผ็ไธไธๅบฆ๏ผๅฆๅฐๆ ้ขๅ ็ฒใ่ฆ็นๅ่กจใๆธ
ๆฐๆฎต่ฝ๏ผ๏ผไปฅ่ทๅพๆดๅนฒๅใ็จณๅฎใๆ่ฏป็ๅ็ญใ
็ฎๅฝ
- ๆจกๅไบฎ็น
- ๆจกๅๆฆ่ง
- ่ฎญ็ปไธๆฐๆฎ
- ๅฟซ้ๅผๅง
- ้็จไธ้ๅถ
- ้จ็ฝฒไธ้ๅ
- ่ฎธๅฏ
- ่ด่ฐข
- ๅผ็จ
- ็ๆฌ่ฎฐๅฝ
ๆจกๅไบฎ็น
- ไธไธไธๅนฒๅ๏ผ้ป่ฎคๅๅฐ่กจๆ ็ฌฆๅท๏ผ่พๅบไปฅๅ ็ฒๅฐๆ ้ข + ่ฆ็นๅ่กจไธบไธป๏ผไพฟไบๅคๅถไธไบๆฌก็ผ่พใ
- ็ปๆ็จณๅฎ๏ผๅฏนๅ่ๆฅๅใๆญฅ้ชคๆธ ๅใๅฏน็ ง่กจใ่ฆ็นๆ่ฆ็ญ่ฟ่กไบ้ฃๆ ผไธๆ ผๅผๅฏน้ฝใ
- ๅ่ฏญ/ๆททๆไผๅ๏ผไธญๆใ่ฑๆไธไธญ่ฑๆททๆๅบๆฏๅ ทๅค่ฏๅฅฝ็ๆฏ่ฏญไธ่ดๆงไธๅฑๆฌกๆธ ๆฐๅบฆใ
- ๆไปค้ตไปๆดๅผบ๏ผๅฏนโไธ่ฆ่กจๆ ็ฌฆๅท / ๅช่พๅบ่ฆ็น่กจ / ไฟ็ Markdown ๆ ้ขๅฑ็บงโ็ญ็บฆๆ้ตไปๅบฆๆด้ซใ
ๅบๅบง๏ผ
meta-llama/Llama-3.1-8B-Instruct๏ผ่ฎญ็ป่ๅผ๏ผๆๅธ่ธ้ฆ + SFTใ
ๆจกๅๆฆ่ง
- ๅๆฐ่งๆจก๏ผ8B
- ไปปๅก๏ผๆไปค่ท้ / ๅฏน่ฏ็ๆ / ้ฎ็ญ / ๆ่ฆ / ็ปๆๅ่พๅบ
- ่ฏญ่จ๏ผไธญๆ & ่ฑๆ๏ผ่ฏๅฅฝๆฏๆไธญ่ฑๆททๆ๏ผ
- ็ฎๆ ๏ผๅจ่ฝป้็ฎๅไธ๏ผไบงๅบ็ฎๆดไธไธใๆ ผๅผๅๅฅฝ็ๅ ๅฎน๏ผๅๅฐ่กจๆ ็ฌฆๅท๏ผไฟ็ๅ ็ฒๅฐๆ ้ขใ่ฆ็นๅ่กจ็ญๆ ผๅผไผๅ๏ผ
่ฎญ็ปไธๆฐๆฎ
- ๆฐๆฎ่งๆจก๏ผ็บฆ 80,000 ๆก้ซ่ดจ้ๆไปค-ๅๅคๆ ทๆฌ๏ผไธญ่ฑๆททๅ๏ผ่ฆ็้ฎ็ญ/ๆ่ฆ/่ฏดๆๆ/็ปๆๅ่พๅบ/ๆญฅ้ชคๅ่ฏดๆ็ญ๏ผใ
- ่ฎญ็ปๆนๆณ๏ผๆๅธ่ธ้ฆ + SFT๏ผๆพๅผๆงๅถๆ ผๅผ/้ฃๆ ผ๏ผๅฐ่กจๆ ใๅผบ่ฐๆ ้ข/ๅ่กจ/ๅ ็ฒ๏ผใ
- ่ฎก็ฎ่ตๆบ๏ผA100 ๅๅก๏ผLoRA/QLoRA ๅฏๅจ่พ็ญๆถ้ดๅ ๅฎๆ่ฅๅนฒ epochใ
- ้ฃๆ ผไธ็บฆๆ๏ผๅๅฐ่กจๆ ็ฌฆๅท๏ผๅผบๅๅฐๆ ้ขๅ ็ฒใ่ฆ็นๅ่กจใๅ ณ้ฎๆฏ่ฏญๅ ็ฒไธๆฎต่ฝๅฑ็บงใ
ๅฆๅๅธ่ธ้ฆๆฐๆฎๅญ้๏ผ่ฏทๅจๆญคๅค่กฅๅ ้พๆฅไธ็ป่ฎก๏ผๆ ทๆฌๆฐ/่ฏญ็งๅ ๆฏ/่ฟๆปคๆ ๅ๏ผใ
ๅฟซ้ๅผๅง
Transformers๏ผๆจ่๏ผ
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model_id = "Jackrong/Llama-3.1-8B-Instruct-Elite"
tok = AutoTokenizer.from_pretrained(model_id, use_fast=True)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto"
)
messages = [
{"role": "system", "content": "Write clean, professional answers with bolded subheadings and structured lists; avoid emojis."},
{"role": "user", "content": "่ฏท็จ่ฆ็น่ฏดๆๅฆไฝไผๅๅจ่ฎกๅ๏ผไฝฟๅ
ถๆดๅฏๆง่กใ"}
]
prompt = tok.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tok(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
**inputs,
max_new_tokens=512,
temperature=0.7,
top_p=0.9,
)
print(tok.decode(outputs[0], skip_special_tokens=True))
vLLM
from vllm import LLM, SamplingParams
llm = LLM(model="Jackrong/Llama-3.1-8B-Instruct-Chinese-Elite", dtype="bfloat16")
params = SamplingParams(temperature=0.7, top_p=0.9, max_tokens=512)
prompt = "ๅๅบ 5 ๆกๅฏๆง่ก็ๅจ่ฎกๅไผๅๅปบ่ฎฎ๏ผ็จๅ ็ฒๅฐๆ ้ข+่ฆ็นๅ่กจ๏ผใ"
outputs = llm.generate([prompt], params)
print(outputs[0].outputs[0].text)
llama.cpp๏ผGGUF๏ผQ4_K_M๏ผ
./main -m Llama-3.1-8B-Instruct-Chinese-Elite.Q4_K_M.gguf \
-p "ไปฅ่ฆ็น่ฏดๆ๏ผๅฆไฝๅฐๆๆฏๆ็ซ ๆนๅๅพๆดไธไธไธๅนฒๅ๏ผ"
ๆ็คบ่ฏไธ่พๅบ่ง่
- ไฝฟ็จ็ฎๆดๆ ้ขไธๅ ็ฒๅฐๆ ้ข็ป็ป็ปๆ๏ผๅ ณ้ฎๆฏ่ฏญไธ็ป่ฎบ้ๅ ็ฒใ
- ไปฅ่ฆ็นๅ่กจๅ็ฐๆญฅ้ชคไธ่ฆ็น๏ผ้ป่ฎค้ฟๅ ่กจๆ ็ฌฆๅทใ
- ้ๆ ทๅปบ่ฎฎ๏ผ
temperature=0.6โ0.8ใtop_p=0.9โ0.95ใ
้็จไธ้ๅถ
้็จ๏ผไธญ/่ฑๆๆไธญ่ฑๆททๆ็้ฎ็ญใๆ่ฆใ่ฏดๆๆใๆๆฏ/ไธๅกๅไฝ๏ผ็ปๆๅ่พๅบ๏ผ่ฎกๅใๆญฅ้ชคใ่กจๆ ผใFAQใไผ่ฎฎ็บช่ฆ๏ผใ
้ๅถ๏ผๅผบไบๅฎๆงใ้ๆๆฐไฟกๆฏ็ไปปๅกๅปบ่ฎฎ้
ๅๆฃ็ดข๏ผๅป็/ๆณๅพ/ๆ่ต็ญ้ซ้ฃ้ฉ่พๅบ้ไบบๅทฅๆ กๅฏน๏ผไธๅพ็จไบ่ฟๆณๆไผคๅฎณๆง็จ้ใ
่ฎธๅฏ
- ๆจกๅๆ้๏ผ้ตๅพช Llama 3.1 Community License๏ผไธๅบๅบงไธ่ด๏ผใ
- ไปฃ็ /่ๆฌ๏ผๅฏไฝฟ็จ Apache-2.0 ็ญ๏ผไธๆนๅๆ้่ฎธๅฏใ
่ด่ฐข
- Meta ๆไพ Llama-3.1 ไธ็ๆๅทฅๅ ท้พ
- ๅผๆบ็คพๅบๅจ่ธ้ฆใSFTใ่ฏๆตใ้จ็ฝฒๆน้ข็่ดก็ฎ
- ่ฎญ็ปๆนๆณไธๅฎ่ทตๆฒฟ็จ่ช Llama-3.2-3B-Elite
ๅผ็จ
@misc{JackrongL31_8B_Elite,
title = {Jackrong/Llama-3.1-8B-Instruct-Chinese-Elite},
author = {Jackrong},
year = {2025},
url = {https://huggingface.co/Jackrong/Llama-3.1-8B-Instruct-Elite}
}
็ๆฌ่ฎฐๅฝ
- v1.0๏ผ้ฆๆฌกๅๅธใๆฐๆฎ่งๆจก็บฆ 80k๏ผA100 ๅๅก่ฎญ็ป๏ผๆไพ GGUF Q4_K_M๏ผๅๅฐ่กจๆ ็ฌฆๅท๏ผๅผบๅๅฐๆ ้ขๅ ็ฒไธ่ฆ็นๅ่กจ๏ผ่ฎญ็ป้ ๆนไธ 3.2-3B-Elite ไธ่ดใ
- Downloads last month
- 2
Hardware compatibility
Log In to add your hardware
4-bit
Model tree for Jackrong/Llama-3.1-8B-Instruct-Chinese-Elite
Base model
meta-llama/Llama-3.1-8B Finetuned
meta-llama/Llama-3.1-8B-Instruct