Instructions to use tchbcb/nqwen27 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use tchbcb/nqwen27 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf tchbcb/nqwen27:IQ2_XS # Run inference directly in the terminal: llama cli -hf tchbcb/nqwen27:IQ2_XS
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf tchbcb/nqwen27:IQ2_XS # Run inference directly in the terminal: llama cli -hf tchbcb/nqwen27:IQ2_XS
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf tchbcb/nqwen27:IQ2_XS # Run inference directly in the terminal: ./build/bin/llama cli -hf tchbcb/nqwen27:IQ2_XS
Use Docker
docker model run hf.co/tchbcb/nqwen27:IQ2_XS
- Llama new
- LM Studio
- Jan
- Ollama
How to use tchbcb/nqwen27 with Ollama:
ollama run hf.co/tchbcb/nqwen27:IQ2_XS
- Unsloth Desktop
- Pi
How to use tchbcb/nqwen27 with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf tchbcb/nqwen27:IQ2_XS
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "tchbcb/nqwen27:IQ2_XS" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use tchbcb/nqwen27 with Docker Model Runner:
docker model run hf.co/tchbcb/nqwen27:IQ2_XS
- Lemonade
How to use tchbcb/nqwen27 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull tchbcb/nqwen27:IQ2_XS
Run and chat with the model
lemonade run user.nqwen27-IQ2_XS
List all available models
lemonade list
- Hermes Agent
How to use tchbcb/nqwen27 with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf tchbcb/nqwen27:IQ2_XS
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default tchbcb/nqwen27:IQ2_XS
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use tchbcb/nqwen27 with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf tchbcb/nqwen27:IQ2_XS
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "tchbcb/nqwen27:IQ2_XS" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
nqwen27 โ Qwen3.8-27B GSQ-RCO IQ2_XS ยท T4 ไธ้ฎ้จ็ฝฒๅฅไปถ
ๆจกๅ: ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF๏ผๅฅฅๅฐๅฉ IST-DASLab๏ผGSQ + RCO ๆขฏๅบฆๆ็ดขๅจๆ้ๅๅ้ๅ๏ผ็ IQ2_XS๏ผ8.42 GB๏ผ ้ๅ็๏ผๅบๅบง Qwen/Qwen3.8-27Bใ
ๅผๆ: llama.cpp b11175๏ผๆบ็ ็ผ่ฏ๏ผGGML_CUDA=ON๏ผCUDA_ARCHITECTURES=75๏ผใ
ไธบไปไนไธๆฏ NInfer๏ผNInfer ๆฏ้ขๅ RTX 5090๏ผsm_120a๏ผ็ๅๅกๅผๆ๏ผๆๅปบๅผบๅถ Blackwell๏ผไธๅชๅ ่ฝฝ็งๆ
.ninferๆ ผๅผ๏ผไธๅ ่ฝฝ GGUF๏ผ๏ผT4๏ผsm_75๏ผๆ ๆณไฝฟ็จ๏ผGSQ-RCO ๅฎๆน README ไบฆๅฃฐๆๅ ถ GGUF "runs unmodified in llama.cpp, Ollama, and LM Studio"ใ
ๆๅก: llama-server ๆไพ OpenAI ๅ
ผๅฎน API๏ผ/v1/chat/completions ็ญ๏ผ๏ผๆจกๅๅ nqwen27๏ผAPI Key 1234๏ผไธไธๆ 128k๏ผCTX=131072๏ผๆจกๅๅ็ๆฏๆ 262144๏ผ๏ผKV cache q8_0 ้ๅ + FlashAttention๏ผๅนถ้่ฟ aitun.cc ๅญๅๅ้ง้ๆด้ฒๅ
ฌ็ฝใ
ๆไปถ
| ๆไปถ | ่ฏดๆ |
|---|---|
deploy_oneclick.sh |
ไธ้ฎ้จ็ฝฒ๏ผๅผๆ๏ผ้ข็ผ่ฏไผๅ ๏ผGLIBC ไธๅ ผๅฎน่ชๅจๆบ็ ็ผ่ฏ๏ผโ ๆ้ โ API โ ้ง้ |
start_server.sh |
ๅ็ฌๅฏๅจ/้ๅฏ API ๆๅก๏ผkey=1234๏ผ128k๏ผๆพๅญไธ่ถณ่ชๅจ้็บง๏ผ |
start_tunnel.sh |
ๅ็ฌๅฏๅจ aitun ๅญๅๅ้ง้ |
rescue_restart.sh |
็ฐๅบๆๆด๏ผไธ้ฎ้ๅฏ 128k API + ๅ ฌ็ฝ้ง้๏ผๆๅก/้ง้ๆ็บฟๆถ็จ๏ผ |
kaggle_oneclick_cell.py |
Kaggle Notebook ๅ cell ๅผๅฏผ่ๆฌ |
engine/llama-b11175-cu128-sm75-t4.tar.gz |
T4 ไธ็จ้ข็ผ่ฏๅผๆๅ ๏ผ54MB๏ผ๏ผllama.cpp b11175 (4de0926), CUDA 12.8, sm_75, glibc 2.35๏ผๅซๅ จ้จ .so๏ผๅ 40 ๅ้ๆบ็ ็ผ่ฏ |
weights/Qwen3.8-27B-GSQ-RCO-IQ2_XS.gguf |
ๆจกๅๆ้๏ผ8.42 GB๏ผๆ ๅ GGUF๏ผ |
Kaggle T4 ๅฟซ้ๅผๅง
ๆฐๅปบ Kaggle Notebook๏ผๅผ GPU T4 ร2ใInternet ๅผๅฏ๏ผ๏ผ็ฒ่ดด่ฟ่ก kaggle_oneclick_cell.py ๅ
ๅฎน๏ผๆไปปไธ Linux + CUDA ๆบๅจ๏ผ
git clone https://huggingface.co/tchbcb/nqwen27 && cd nqwen27
AITUN_API_KEY=<ไฝ ็aitun_key> bash deploy_oneclick.sh
่ๆฌ่ชๅจๅฎๆ๏ผllama.cpp ๅผๆ๏ผ้ป่ฎคไผๅ
ๆๆฌไปๅบ T4 ้ข็ผ่ฏๅ
๏ผ54MB ๆฐๅ็ง๏ผไธๅ
ผๅฎนๆถๅฎๆน้ข็ผ่ฏๅ
โ ๆบ็ ็ผ่ฏ็บฆ 40 ๅ้๏ผโ ไปๆฌไปๅบ weights/ ไธ่ฝฝๆ้๏ผๅคฑ่ดฅ่ชๅจๅ้ ISTA-DASLab ๅไปๅบ๏ผโ ๅฏๅจ API๏ผ127.0.0.1:8080๏ผkey=1234๏ผ128k ไธไธๆ๏ผโ aitun ้ง้่พๅบๅ
ฌ็ฝ https://<ๅญๅๅ>.aitun.ccใๅ
จ็จ็บฆ 10-15 ๅ้๏ผๆ้ไธ่ฝฝๅ ๅคงๅคด๏ผใ
API ่ฐ็จ
ๅทฒ้จ็ฝฒ็คบไพ๏ผๆๅกๅญๆดปๆ้ด๏ผ๏ผhttps://qwenapi.aitun.cc
curl https://qwenapi.aitun.cc/v1/chat/completions \
-H "Authorization: Bearer 1234" \
-H "Content-Type: application/json" \
-d '{"model":"nqwen27","messages":[{"role":"user","content":"็จไธๅฅ่ฏไป็ป็ปดไน็บณ"}]}'
from openai import OpenAI
client = OpenAI(base_url="https://<ไฝ ็ๅญๅๅ>.aitun.cc/v1", api_key="1234")
r = client.chat.completions.create(
model="nqwen27",
messages=[{"role": "user", "content": "ไฝ ๅฅฝ"}],
)
print(r.choices[0].message.content)
- ๅ
ณ้ญๆ็ปด้พ๏ผQwen3.8 ้ป่ฎคๅผๅฏ thinking๏ผ: ่ฏทๆฑไฝๅ
"chat_template_kwargs": {"enable_thinking": false} - ๆตๅผ่พๅบ:
"stream": true - ่ง่งๅคๆจกๆ: ๆฌๅฅไปถ้ป่ฎค็บฏๆๆฌ๏ผๅฆ้ๅพ็็่งฃ๏ผๅฆ่กไธ่ฝฝๅไปๅบ
mmproj-Qwen3.8-27B-BF16.ggufๅนถ็ป llama-server ๅ--mmprojๅๆฐ
็ฏๅขๅ้
| ๅ้ | ้ป่ฎค | ่ฏดๆ |
|---|---|---|
AITUN_API_KEY |
็ฉบ | aitun ๆณจๅ็จๆท key๏ผๅญๅๅ้ง้ๅฟ ้๏ผ |
API_KEY |
1234 |
OpenAI ๅ ผๅฎน API ็ Bearer ้ดๆ key |
PORT |
8080 |
ๆฌๆบๆๅก็ซฏๅฃ |
CTX |
131072 |
ไธไธๆ้ฟๅบฆ๏ผ128k๏ผๆๅคงๅฏ่ณๆจกๅๅ็ 262144๏ผไฝ T4 ๆพๅญๆ้๏ผ |
P1_CTK/P1_CTV ็ญ |
q8_0/q4_0 |
KV cache ้ๅ้ข่ฎพ้พ๏ผ่งไธๆน FAQ |
SPEC_FLAGS |
็ฉบ | llama-server ้ขๅคๅๆฐ้ไผ ๏ผๆๆบ่งฃ็ ๅฎ้ช็ญ๏ผ๏ผ่งไธๆนๆง่ฝ็ซ ่ |
WORK |
/root/nqwen27 |
้จ็ฝฒ็ฎๅฝ |
SKIP_PREBUILT |
0 |
็ฝฎ 1 ่ทณ่ฟ้ข็ผ่ฏๅฐ่ฏ๏ผ็ดๆฅๆบ็ ็ผ่ฏ |
FAQ๏ผ128k ไธไธๆๆไนๆพ่ฟ 16G ๆพๅญ๏ผ
llama.cpp ็ KV cache ๆฏๆพๅญๅคงๅคด๏ผstart_server.sh ็จ KV cache ้ๅ๏ผ-ctk/-ctv๏ผ+ FlashAttention๏ผ-fa on๏ผ ๅ็ผฉ KV๏ผๅนถๆๆพๅญ่ชๅจ้ๆกฃ๏ผ
| ้ข่ฎพ | ๅฑๅธ่ฝฝ | KV ้ๅ | ้็จ |
|---|---|---|---|
| P1๏ผ้ป่ฎค๏ผ | -ngl 99 ๅ
จๅ
ฅ GPU |
q8_0 |
2รT4๏ผๅฑๅๅๅ KV ็บฆ 8-9G/ๅก๏ผ |
| P2 | -ngl 99 ๅ
จๅ
ฅ GPU |
q4_0 |
ๆพๅญ็ดงๅผ |
| P3 | -ngl 35 ้จๅๅธ่ฝฝ |
q4_0 |
ๅๅก 16G ๅ ๅบ |
ๅฏๅจๆถๅ ่ฏ P1๏ผๆฃๆตๅฐ OOM ่ชๅจๆข P2ใP3๏ผๆๅ็ๆกฃไฝไผๆๅฐๅจๆฅๅฟ้ใKV q8_0 ๆฅ่ฟๆ ๆ๏ผq4_0 ๆ่ฝปๅพฎ็ฒพๅบฆๆๅคฑไฝๅฎๅ จๅฏ็จใ
ๆง่ฝๅบๅไธๆ้่ฏดๆ๏ผ2รT4 ๅฎๆต๏ผ
็จ llama-bench๏ผp1024/tg128๏ผb11175๏ผๅฎๆตไบ็งๆนๆกๅ็็ป่ฎบโโๅฝๅ้ป่ฎค้
็ฝฎ๏ผ-sm layer + ้ป่ฎค ubatch 512๏ผๅฐฑๆฏ 2รT4 ไธ็ๆไผ่งฃ๏ผ
| ๆนๆก | prefill tok/s | decode tok/s | ็ป่ฎบ |
|---|---|---|---|
-sm layer๏ผๅฝๅ้ป่ฎค๏ผ |
313 | 11.5 | ๅๅกๅฑๅๅๆฅๅ๏ผT4 ๆ NVLink ไธๆๅฟซ |
-sm none๏ผๅๅกๅ
จๆพ๏ผ |
223 | 8.0 | ไธคๅกๅ่ฎกๅธฆๅฎฝ > ๅๅก๏ผๅณไฝฟๆๆฅๅๅผ้ |
-sm tensor๏ผๅผ ้ๅนถ่ก๏ผ |
181 | 5.0 | ้ๅฑ PCIe ๅๆญฅๅผ้่ฟ่ถ ๆถ็ |
-sm row |
โ | โ | ๆฌ้ๅ + FA ็ปๅไธๅฏๅจๅคฑ่ดฅ |
-ub 2048 |
224 | 10.9 | ๅคง ubatch ๅ่้้๏ผไฟๆ้ป่ฎค 512 |
ๅฎๆต่งฃ็ ็บฆ 11-13 tok/sใprefill ็บฆ 275-310 tok/s๏ผๅณ 27B ๆจกๅ @ 2bpw ๅจ 2รT4๏ผ2018 ๅนดๆฐๆฎไธญๅฟๅก๏ผๆ NVLink๏ผไธ็็กฌไปถๆ้ๆฐดๅนณใ
ๅ
ณไบ MTP๏ผb11175 ๅผๆๆฏๆ --spec-type draft-mtp๏ผไฝ GSQ-RCO ้ๅ็ GGUF ไธๅซ MTP ๅฑ๏ผๅฏๅจๆถๆฅ context type MTP requested but model doesn't contain MTP layers๏ผ๏ผๅ ๆญคๆ ๆณๅผๅฏ๏ผ้่ฆๆขๅธฆ MTP ๅคด็ๆ้ๆ่กใ
ๆๆบ่งฃ็ ๅฎ้ช๏ผๅทฒๅฎๆต๏ผ้ป่ฎคไธๅผๅฏ๏ผ๏ผ
ngram-simple๏ผๅ ่็จฟๆจกๅ๏ผ๏ผๆฎ้ๆๆฌ +5% ไปฅๅ ๏ผ้ๅคๆๆฌ ยฑ0๏ผไธๅผๅพ๏ผdraft-simple+ unsloth/Qwen3.5-2B-GGUF ่็จฟ๏ผๆฎ้ๅฏน่ฏ 8.8 tok/s๏ผๅๆ ข 20%๏ผ๏ผ้ซๅบฆ้ๅคๆๆฌ 12.2 tok/s๏ผ+12%๏ผ๏ผไป ๅฏน้ๅคๅ็ๆ๏ผๆน้ๆจกๆฟใไปฃ็ ่กฅๅ จ๏ผๆๅพฎๅผฑๆถ็ใ
ๅฆ้ๅจ้ๅคๅๅทฅไฝ่ด่ฝฝไธ่ฏ้ช๏ผ
# ่็จฟๆจกๅๅทฒๅฏ่ชๅจไธ่ฝฝ๏ผ็บฆ1.3GB๏ผ๏ผๆๆๅจๆพๅ
ฅ $WORK/draft/
SPEC_FLAGS="--spec-type draft-simple -md /root/nqwen27/draft/Qwen3.5-2B-Q4_K_M.gguf -ngld 99 -devd CUDA0 -ctkd q4_0 -ctvd q4_0" \
bash start_server.sh
# ๆๅ
่็จฟ็ ngram ๆจกๅผ
SPEC_FLAGS="--spec-type ngram-simple" bash start_server.sh
ๅฏๅจๅคฑ่ดฅไผ่ชๅจๆฒฟ P1โP2โP3 ้ๆกฃๅ
ๅบ๏ผๅปๆ SPEC_FLAGS ้่ทๅณๆขๅค้ป่ฎค้
็ฝฎใ
ๅฎๅ จๆ็คบ
ไปๅบๅ
่ๆฌไธๅ
ๅซไปปไฝ็ๅฎๅฏ้ฅ๏ผaitun key / HF token ๅ้่ฟ็ฏๅขๅ้ไผ ๅ
ฅ๏ผใๅ
ฌ็ฝๆๅก่ฏทๅกๅฟ
ไฟ็ API_KEY ้ดๆ๏ผๅฟ
่ฆๆถๆ้ป่ฎค็ 1234 ๆนๆๅผบ keyใ
ๆฅๆบไธ่ด่ฐข
- ๅบๅบงๆจกๅ: Qwen/Qwen3.8-27B๏ผApache-2.0๏ผ
- ้ๅ: ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF๏ผGSQ: arXiv 2604.18556๏ผRCO: arXiv 2605.00649๏ผ
- ๅผๆ: ggml-org/llama.cpp
- Downloads last month
- 312
2-bit
Model tree for tchbcb/nqwen27
Base model
Qwen/Qwen3.8-27B