Instructions to use peculiar-ragdoll/Sharp-MiniCPM5-2B-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 peculiar-ragdoll/Sharp-MiniCPM5-2B-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 peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S
# Run inference directly in the terminal:
llama cli -hf peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S
# Run inference directly in the terminal:
llama cli -hf peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

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 peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S
# Run inference directly in the terminal:
./build/bin/llama cli -hf peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

Use Docker

docker model run hf.co/peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

Llama new
LM Studio
Jan

vLLM

How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "peculiar-ragdoll/Sharp-MiniCPM5-2B-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": "peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

Ollama
How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF with Ollama:
```
ollama run hf.co/peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S
```
Unsloth Desktop

How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

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": "peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF with Docker Model Runner:
```
docker model run hf.co/peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S
```

Lemonade

How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

Run and chat with the model

lemonade run user.Sharp-MiniCPM5-2B-GGUF-Q4_K_S

List all available models

lemonade list

Hermes Agent

How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-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 peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

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 peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S

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 "peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q4_K_S" \
  --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"

Maple-Preview or LFM2.5-2.6B Sharps?

by warlocksyno - opened Sep 28

Discussion

warlocksyno

Sep 28

•

edited Sep 28

Hey! Amazing work! I've been using Dirk and Tiel for all my deployments, they've been fantastic. Dirk has been great on the dual GPU systems, Tiel for single GPU and MoE offloaded ones. MiniCPM5 and Spark might be the answer for >8GB systems, or even CPU. But in my testing LFM2.5, Ling Tiny, and Maple have been the fastest on CPU and iGPU, but both LFM and Maple never actually output their answers in any given budget, even though within the first 3-4% of their thinking block the answer is in there. Would it be possible to give one of those the Sharp treatment? Ling does better, but still not great.

warlocksyno changed discussion title from Maple-Preview or LFM2.5-2.6B Swifts? to Maple-Preview or LFM2.5-2.6B Sharps? Sep 28

peculiar-ragdoll

Owner Sep 28

Hey there, thanks for the request and the warm praise!:)

What do you think is the right use cases for these models (Maple, LFM and Ling respecitvely)? Chat assistants? Knowledge retrieval? Agentic tool call? Because when I dynamically quantize stuff the calibration corpus makes a difference, so if it's not for coding then I'd just put a bunch of tool calling, reasoning and other stuff relevant to the use case in there.

Personally I only use LLMs for agentic coding and related tasks, so I'd rather hear what the use case for these models are from someone who uses these than assume what it's for.

warlocksyno

Sep 28

In these cases it's for RAG retrieval and chat. Trying to build a harness for myself to quickly get accurate information from automotive repair manuals. For instance, asking a bot "what are the torque specs for a 2004 Jeep Wrangler 2.4L headbolts", the harness should try to get the information without an agentic model, but if it can't it forwards it to a model to grab. E2B has been great for that, but trash at anything else. Like if you asked it to "generate a invoice for shop order #243" or whatever, it falls on it's face. So a mixture of accurate information retrieval and tool use. What's annoying is pretty much all 3 of the models mentioned do have the answer, it's just they spend the other 97% of the thought cap spiraling lol.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment