Yee-R1
Collection
3 items โข Updated
# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("sds-ai/Yee-R1-mini-GGUF", device_map="auto")Quick Links
ๅฐ็ ๏ผYee๏ผAI ๆฐๆฎๅฎๅ จไธๅฎถ
็ฑ ๅนฟๅท็ ๆฐไฟกๆฏๆๆฏๆ้ๅ ฌๅธ ๅผๅ๏ผๅบไบๅคง่ฏญ่จๆจกๅๆๆฏๆๅปบ็ๆฐๆฎๅฎๅ จๆบ่ฝๅฉๆใ ่ฏฅไปๅบไธบ Yee-R1-mini ็ GGUF ๆจกๅๆไปถ
ๅฐ็ ๏ผYee๏ผๆฏไธๆฌพไธๆณจไบ ๆฐๆฎๅฎๅ จ้ขๅ ็ AI ไธๅฎถ็ณป็ป๏ผไพๆไบๅ ่ฟ็ Qwen3-1.7B ๅคง่ฏญ่จๆจกๅๆถๆ๏ผๅนถ่ๅไบๆฐๆฎๅ็ฑปๅ็บงใๅฎๅ จๅฎก่ฎกใ้ฒๆคๆฃๆต็ญไธไธ่ฝๅใๅฎไธบๅทฅไธใๆฟๅกใ่ฟ่ฅๅ็ญ่กไธๆไพ่ฝป้ๅใๆบ่ฝๅ็ๆฐๆฎๅฎๅ จ่งฃๅณๆนๆก๏ผๅธฎๅฉ็จๆทๅฎ็ฐโๅ่งใๅฏ่งใๅฏๆงใๅฏ้ฒโ็ๆฐๆฎๅฎๅ จ็ฎๆ ใ
ๅฐ็ ไปฅ AI ๆฐๆฎๅฎๅ จไธๅฎถๅคงๆจกๅ ไธบๆ ธๅฟๆๆฏๅบๅบง๏ผๆๅปบไบๅ จๆ ๅผๆฐๆฎๅฎๅ จๅฎก่ฎกไธๅ จ้พ่ทฏ้ฒๆณ้ฒไฝ็ณป๏ผๅจโไบโใโ็ฎกโใโ็ซฏโไธๅคงๅบๆฏไธญ่ฝๅฐๅบ็จ๏ผๅฉๅไผไธไปๅฎนๅบๅฏนๆฐๅญ็ปๆตๆถไปฃ็ๅฎๅ จๆๆใ
๐ ๆ ธๅฟ็น็น
ๅบไบ Qwen3-1.7B ๆๅปบ
- ไฝฟ็จ้ฟ้ๅทดๅทด้ไนๅ้ฎๆๆฐไธไปฃๅคง่ฏญ่จๆจกๅ Qwen3๏ผๅ ทๅคๅผบๅคง็ๆจ็ใ้ป่พๅคๆญไธๆไปคๆง่ก่ฝๅใ
- ๆฏๆๅจ Thinking Mode ๅ Non-Thinking Mode ไน้ด็ตๆดปๅๆข๏ผ้ๅบไธๅๅบ็จๅบๆฏใ
ๅๆจกๆจ็ๆบๅถ
- ๅจๅคๆ้ป่พไปปๅก๏ผๅฆไปฃ็ ๅๆใๆฐๅญฆ่ฎก็ฎใ็ญ็ฅๅถๅฎ๏ผไธญๅฏ็จ Thinking Modeใ
- ๅจๆฅๅธธๅฏน่ฏใๅฟซ้ๅๅบๅบๆฏไธญไฝฟ็จ Non-Thinking Mode๏ผๆๅๆ็ใ
Agent ๅ่ฝๅ
- ้ๆ Qwen-Agent ๆกๆถ๏ผๆฏๆ่ฐ็จๅค้จๅทฅๅ ท๏ผๅฆๆฐๆฎๅบๆฅๅฃใๆฅๅฟๅๆๅจใAPI ๆฅๅฃ็ญ๏ผ๏ผๅฎ็ฐ่ชๅจๅไปปๅกๆง่กใ
้ซๅ ผๅฎนๆง
- ๆฏๆไธปๆต้จ็ฝฒๆนๅผ๏ผๆฌๅฐ่ฟ่กใDocker ๅฎนๅจใKubernetes ้็พคใSaaS API ๆฅๅฃ็ญใ
- ๅ ผๅฎน HuggingFace TransformersใvLLMใSGLangใOllama ็ญๆจ็ๆกๆถใ
๐ ๆง่ฝๆต่ฏ
ไปฅไธๆฏๅฐ็ ๅจ CS-Eval ไธญๅคไธชๅฎๅ จ้ขๅ็็ปผๅๅพๅๆต่ฏ็ปๆ๏ผๅบไบๆจกๆ็ๅฎไธๅกๅบๆฏ็่ฏไผฐไฝ็ณป็ๆ๏ผ
| ็ปผๅๅพๅ | ็ณป็ปๅฎๅ จๅ่ฝฏไปถๅฎๅ จๅบ็ก | ่ฎฟ้ฎๆงๅถไธ่บซไปฝ็ฎก็ | ๅ ๅฏๆๆฏไธๅฏ้ฅ็ฎก็ | ๅบ็ก่ฎพๆฝๅฎๅ จ | AIไธ็ฝ็ปๅฎๅ จ | ๆผๆด็ฎก็ไธๆธ้ๆต่ฏ | ๅจ่ๆฃๆตไธ้ข้ฒ | ๆฐๆฎๅฎๅ จๅ้็งไฟๆค | ไพๅบ้พๅฎๅ จ | ๅฎๅ จๆถๆ่ฎพ่ฎก | ไธๅก่ฟ็ปญๆงไธๅบๆฅๅๅบๆขๅค | ไธญๆไปปๅก | ่ฑๆไปปๅก |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 77.48 | 78.00 | 79.31 | 71.90 | 78.37 | 84.65 | 75.24 | 78.41 | 73.02 | 86.71 | 80.49 | 71.33 | 77.58 | 76.03 |
๐ฆ ๅฟซ้ๅผๅง
from transformers import AutoTokenizer, AutoModelForCausalLM
# ๅ ่ฝฝ tokenizer ๅๆจกๅ
tokenizer = AutoTokenizer.from_pretrained("sds-ai/Yee-R1-mini")
model = AutoModelForCausalLM.from_pretrained(
"sds-ai/Yee-R1-mini",
torch_dtype="auto",
device_map="auto"
)
# ่พๅ
ฅๆ็คบ
prompt = "่ฏทๅธฎๆๆฃๆฅ่ฟไปฝๆฐๆฎๆฏๅฆๅ
ๅซๆๆๅญๆฎต๏ผ"
# ๅบ็จ่ๅคฉๆจกๆฟๅนถๅๆขๆจกๅผ
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=True # ๅๆข่ณๆ่ๆจกๅผ
)
# ็ผ็ ่พๅ
ฅ
inputs = tokenizer([text], return_tensors="pt").to(model.device)
# ็ๆๅๅบ
response_ids = model.generate(**inputs, max_new_tokens=32768)
response = tokenizer.decode(response_ids[0][len(inputs.input_ids[0]):], skip_special_tokens=True)
print("ๅฐ็ ๏ผ\n", response)
๐ ๏ธ ้จ็ฝฒๆนๅผ
ไฝ ๅฏไปฅ้่ฟไปฅไธไปปๆไธ็งๆนๅผ้จ็ฝฒๅฐ็ ๏ผ
ไฝฟ็จ SGLang ๅฏๅจๆๅก
python -m sglang.launch_server --model-path sds-ai/Yee-R1-mini --reasoning-parser qwen3
ไฝฟ็จ vLLM ๅฏๅจๆๅก
vllm serve sds-ai/Yee-R1-mini --enable-reasoning --reasoning-parser deepseek_r1
ไฝฟ็จ Ollama / LMStudio / llama.cpp / KTransformers
Qwen3 ๅทฒ่ขซไธปๆตๆฌๅฐๅ LLM ๅทฅๅ ทๅนฟๆณๆฏๆ๏ผ่ฏฆๆ ่ฏทๅ่ๅฎๆนๆๆกฃใ
๐ ๆไฝณๅฎ่ทตๅปบ่ฎฎ
ไธบ่ทๅพๆไฝณๆง่ฝ๏ผ่ฏท้ตๅพชไปฅไธๆจ่่ฎพ็ฝฎ๏ผ
| ๅบๆฏ | ๆธฉๅบฆ | TopP | TopK | MinP | Presence Penalty |
|---|---|---|---|---|---|
ๆ่ๆจกๅผ (enable_thinking=True) |
0.6 | 0.95 | 20 | 0 | 1.5 (ๅๅฐ้ๅค่พๅบ) |
้ๆ่ๆจกๅผ (enable_thinking=False) |
0.7 | 0.8 | 20 | 0 | ไธๆจ่ไฝฟ็จ |
- ่พๅบ้ฟๅบฆๅปบ่ฎฎ่ฎพไธบ 32,768 tokens๏ผๅคๆไปปๅกๅฏๆๅ่ณ 38,912 tokensใ
- ๅจๅค่ฝฎๅฏน่ฏไธญ๏ผๅๅฒ่ฎฐๅฝๅบไป ไฟ็ๆ็ป่พๅบ้จๅ๏ผ้ฟๅ ๅผๅ ฅๆ็ปดๅ ๅฎนๅฝฑๅไธไธๆ็่งฃใ
๐ ่็ณปๆไปฌ
ไบ่งฃๆดๅคๅ ณไบๅฐ็ ็ไฟกๆฏ๏ผ่ฏท่ฎฟ้ฎ ็ ๆฐไฟกๆฏๅฎ็ฝ
๐ ่ด่ฐข
ๆ่ฐข้ฟ้้ไนๅฎ้ชๅฎคๅผๆบ Qwen3 ๆจกๅ๏ผไธบๅฐ็ ๆไพไบๅๅฎ็่ฏญ่จ็่งฃๅ็ๆ่ฝๅๅบ็กใ
- Downloads last month
- 371
Hardware compatibility
Log In to add your hardware
2-bit
3-bit
4-bit
5-bit
6-bit
8-bit
16-bit
Model tree for sds-ai/Yee-R1-mini-GGUF
Base model
Qwen/Qwen3-1.7B-Base
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="sds-ai/Yee-R1-mini-GGUF")