Instructions to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF 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 xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF
# Run inference directly in the terminal:
llama cli -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF
# Run inference directly in the terminal:
llama cli -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF
# Run inference directly in the terminal:
./llama-cli -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF
# Run inference directly in the terminal:
./build/bin/llama-cli -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Use Docker

docker model run hf.co/xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

LM Studio
Jan

vLLM

How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Ollama
How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with Ollama:
```
ollama run hf.co/xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF
```
Unsloth Desktop

How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

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": "xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with Docker Model Runner:
```
docker model run hf.co/xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF
```

Lemonade

How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Run and chat with the model

lemonade run user.gpt-oss-20b-Code-xCloud-GGUF-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

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 xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF

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 "xCloudinfo/gpt-oss-20b-Code-xCloud-GGUF" \
  --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"

guessleej commited on Jun 25

Commit

3d688ce

verified ·

1 Parent(s): 92678be

Update model card: function + honest strengths

Browse files

Files changed (1) hide show

README.md +11 -2

README.md CHANGED Viewed

@@ -20,9 +20,18 @@ pipeline_tag: text-generation
 以 `openai/gpt-oss-20b`（21B 總參 / 3.6B 活躍 / MoE / MXFP4 / harmony 推理格式）為基底 的**程式碼能力強化** reasoning 模型。以**執行驗證蒸餾**的程式碼指令資料做 LoRA 微調（LoRA 作用於 attention，MoE 專家維持原生 MXFP4），保留 gpt-oss 原生 reasoning 能力。（GGUF，MXFP4，約 14GB）
 ## 做法
-- **資料**：程式碼指令資料**每筆解法都先在沙箱跑過隱藏測試、通過才收**（語料「每筆都證明會動」）。
-- **定位**：對齊「可驗證、可執行」的程式風格與自家技術堆疊；非以刷公開 benchmark 為目標。
 ## 用法（gpt-oss 是 reasoning 模型，務必加 `--jinja` 套用內建 harmony 模板）
 ### llama.cpp

 以 `openai/gpt-oss-20b`（21B 總參 / 3.6B 活躍 / MoE / MXFP4 / harmony 推理格式）為基底 的**程式碼能力強化** reasoning 模型。以**執行驗證蒸餾**的程式碼指令資料做 LoRA 微調（LoRA 作用於 attention，MoE 專家維持原生 MXFP4），保留 gpt-oss 原生 reasoning 能力。（GGUF，MXFP4，約 14GB）
+**功能**：寫程式——理解需求、產生可執行的 Python／程式碼解法，並保留逐步推理（reasoning）。
+## 厲害在哪（據實、不灌水）
+- **訓練料每一筆都經「執行驗證」**：解法在沙箱跑過隱藏測試、通過才收（rejection sampling）——不是網路爬來的程式碼，**每一筆都證明會動**。這是多數蒸餾模型給不出的資料品質保證。
+- **保留通用 coding 實力**：HumanEval pass@1 **87.2%**（第三方題庫、greedy、164 題）——對 21B 總參 / 3.6B 活躍的開源模型屬頂規檔次。
+- **整個家族的「碼力底層」**：[`TAIDE-zhTW`（繁中）](https://huggingface.co/xCloudinfo/gpt-oss-20b-TAIDE-zhTW) 與 [`Uncensored`（無審查）](https://huggingface.co/xCloudinfo/gpt-oss-20b-Uncensored-xCloud) 兩顆都疊在這顆之上。
+- **輕量好部署**：MXFP4、約 14GB，單張中階 GPU（甚至 CPU）即可本地、可控、離線部署，Apache-2.0 商用友善。
+> 定位是「**可控、可驗證、自架**的實用 coder」，不是去刷贏前沿封閉模型；價值在資料每筆可執行、行為可控、貼合自家技術堆疊。
 ## 做法
+- **資料**：程式碼指令資料**每筆解法都先在沙箱跑過隱藏測試、通過才收**。
+- **方法**：teacher 蒸餾 + 執行驗證閘門（rejection sampling）→ LoRA SFT。
 ## 用法（gpt-oss 是 reasoning 模型，務必加 `--jinja` 套用內建 harmony 模板）
 ### llama.cpp