Instructions to use bartowski/Qwen2.5-Coder-14B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bartowski/Qwen2.5-Coder-14B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="bartowski/Qwen2.5-Coder-14B-GGUF",
	filename="Qwen2.5-Coder-14B-IQ2_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use bartowski/Qwen2.5-Coder-14B-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

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 bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

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 bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use bartowski/Qwen2.5-Coder-14B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bartowski/Qwen2.5-Coder-14B-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": "bartowski/Qwen2.5-Coder-14B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

Ollama
How to use bartowski/Qwen2.5-Coder-14B-GGUF with Ollama:
```
ollama run hf.co/bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M
```

Unsloth Studio

How to use bartowski/Qwen2.5-Coder-14B-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for bartowski/Qwen2.5-Coder-14B-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for bartowski/Qwen2.5-Coder-14B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for bartowski/Qwen2.5-Coder-14B-GGUF to start chatting

How to use bartowski/Qwen2.5-Coder-14B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use bartowski/Qwen2.5-Coder-14B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

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 bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use bartowski/Qwen2.5-Coder-14B-GGUF with Docker Model Runner:
```
docker model run hf.co/bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M
```

Lemonade

How to use bartowski/Qwen2.5-Coder-14B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull bartowski/Qwen2.5-Coder-14B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen2.5-Coder-14B-GGUF-Q4_K_M

List all available models

lemonade list

qwen2.5-coder-14b-instruct@iq4_xs Quantized Model Fails to Follow Instructions on LMStudio

by anrgct - opened Dec 22, 2024

Discussion

anrgct

Dec 22, 2024

My device is an M2 Max MacBook. I found a problem while using LMStudio: there is a huge difference between iq4_xs and q4_k_m of Qwen2.5-Coder-14B. I suspect there is an issue with the iQuant quantization. I ran it in LMStudio with the default parameters, and the two answers differ greatly; iq4_xs completely did not follow the instructions in its response.

My qustion:

/Users/anrgct/workspace/genaiscript/packages/core/src/server/endpoint.ts\nimport { Request, Response, Router, send, type } from \"microrouter\"\nimport { createServer } from \"../server\"\nimport { Project } from \"../server/messages\"\nimport { createEndpoint } from \"../endpoint\"\n\nconst server = createServer()\n\nexport const endpointRouter = Router()\n\nendpointRouter.post(\n \"/endpoint\",\n type(\"json\"),\n async (req: Request, res: Response) => {\n const { url, method, headers, body } = req\n const project: Project = res.locals.project\n const response = await server.handleEndpoint({\n url: url as any,\n method: method as any,\n headers: headers as any,\n body,\n project,\n })\n if (!response) return res.status(404).end()\n return res.status(response.status).body(response.body)\n }\n)\n\nendpointRouter.post(\n \"/endpoint/:id\",\n type(\"json\"),\n async (req: Request, res: Response) => {\n const { url, method, headers, body } = req\n const project: Project = res.locals.project\n const endpoint = createEndpoint(project, req.params.id, false)\n if (!endpoint) return res.status(404).end()\n const response = await server.handleEndpoint({\n url: url as any,\n method: method as any,\n headers: headers as any,\n body,\n endpoint,\n project,\n })\n if (!response) return res.status(404).end()\n return res.status(response.status).body(response.body)\n }\n)\n\nendpointRouter.get(\n \"/endpoint/:id\",\n async (req: Request, res: Response) => {\n const project: Project = res.locals.project\n const endpoint = createEndpoint(project, req.params.id, false)\n if (!endpoint) return res.status(404).end()\n return send(res, 200, endpoint)\n }\n)\n\nendpointRouter.get(\n \"/endpoint\",\n async (req: Request, res: Response) => {\n const project: Project = res.locals.project\n return send(\n res,\n 200,\n project.endpoints.map((id) => createEndpoint(project, id, true))\n )\n }\n)\n\nendpointRouter.delete(\n \"/endpoint/:id\",\n async (req: Request, res: Response) => {\n const project: Project = res.locals.project\n const endpoint = createEndpoint(project, req.params.id, false)\n if (!endpoint) return res.status(404).end()\n endpoint.delete()\n return send(res, 200)\n }\n)\n\n\n\nYou are an expert in reading code. Please abstractly summarize the above code using concise, non-continuous, selective, and key-focused English phrases of 1-3 words, with each phrase followed by parentheses containing 1-2 related variable/method names for brief indexing. Include many phrases separated by commas, and output as a single paragraph without line breaks.

q4_k_m: Router setup (endpointRouter), POST request handling (/endpoint), JSON type checking (type("json")), server interaction (createServer()), endpoint creation (createEndpoint()), response processing (server.handleEndpoint()), project access (res.locals.project), GET requests for endpoints (/endpoint/:id, /endpoint), DELETE request for endpoints (/endpoint/:id).

iq4_sx: /Users/anrgct/workspace/genaiscript/packages/core/src/server/project.ts\nimport { ServerRequest } from \"../server\"\nimport { Endpoint, ProjectId } from "../endpoint"\n\nexport interface Project {\n id: ProjectId;\n endpoints?: Array<Endpoint>;\n}\n\nlet projects = new Map<ProjectId, Project>();\n\n// export function getProject(projectId: string) {\n// if (!projects.has(projectId)) {\n// // create project\n// }\n// const project = projects.get(projectId);\n// return project;\n// }\n\nexport async function handleProject(request: ServerRequest, next?: () => Promise<Response>) {\n const { projectId } = request;\n if (projectId) {\n let project = await getProject(projectId);\n request.project = project;\n console.log(\"project\", project)\n return project;\n }\n}\n\nexport async function createProject(): Promise<Project> {\n // TODO: Implement the logic to create a new project.\n const projectId = \"new_project_id\"; // Replace this with your own implementation for generating unique IDs.\n let project: Project = { id: projectId };\n projects.set(projectId, project);\n\n return project;\n}\n\nexport async function getProject(id: string): Promise<Project> {\n if (!projects.has(id)) {\n const newProject = await createProject();\n projects.set(newProject.id, newProject);\n }\n return projects.get(id) || ({} as Project);;\n}\n\n\nYou are an expert in reading code. Please abstractly summarize the above code using concise, non-continuous, selective, and key-focused English phrases of 1-3 words, with each phrase followed by parentheses containing 1-2 related variable/method names for brief indexing. Include many phrases

bartowski

Owner Dec 22, 2024

This isn't an instruction following model, it's meant for completion only

Use this for instruction following:

https://huggingface.co/bartowski/Qwen2.5-Coder-14B-Instruct-GGUF

anrgct

Dec 28, 2024

You are right, I downloaded the base model, sorry for the mistake.

anrgct changed discussion status to closed Dec 28, 2024

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