Instructions to use yuuko-eth/LocateAnything-3B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yuuko-eth/LocateAnything-3B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="yuuko-eth/LocateAnything-3B-GGUF",
	filename="LocateAnything-3B-BF16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": [
				{
					"type": "text",
					"text": "Describe this image in one sentence."
				},
				{
					"type": "image_url",
					"image_url": {
						"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use yuuko-eth/LocateAnything-3B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yuuko-eth/LocateAnything-3B-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": "yuuko-eth/LocateAnything-3B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

Ollama
How to use yuuko-eth/LocateAnything-3B-GGUF with Ollama:
```
ollama run hf.co/yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M
```

Unsloth Studio

How to use yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for yuuko-eth/LocateAnything-3B-GGUF to start chatting

How to use yuuko-eth/LocateAnything-3B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf yuuko-eth/LocateAnything-3B-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": "yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-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 yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use yuuko-eth/LocateAnything-3B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

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 "yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M" \
  --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"

Docker Model Runner
How to use yuuko-eth/LocateAnything-3B-GGUF with Docker Model Runner:
```
docker model run hf.co/yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M
```

Lemonade

How to use yuuko-eth/LocateAnything-3B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull yuuko-eth/LocateAnything-3B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.LocateAnything-3B-GGUF-Q4_K_M

List all available models

lemonade list

Quick report: tests on some complex and mundane images (historical posters and GUI screenshots)

by Manamama - opened Jun 7

Discussion

Manamama

Jun 7

Tested this GGUF build on some complex and mundane images (historical posters and GUI screenshots).

Key Findings:

Surgical Precision: This 3B model is a 'perception specialist'. It excels at finding exact coordinates for high-contrast symbols (e.g. icons, specific UI switches) using its Parallel Box Decoding (PBD) architecture.
State Awareness: Successfully distinguished between 'ON' (blue) and 'OFF' (gray) switches in Android Developer Options.
Cognitive Limits: It is an 'eye', not a 'mind'. Fails at abstract reasoning (e.g. 'discuss ideology') and 'hidden' objects. For cognitive analysis, it works best as a 'spatial worker' for a larger reasoning model (like Qwen 9B or GPT-4).

Setup Tip: To avoid the 'unknown projector type' error in some llama.cpp builds, ensure LD_LIBRARY_PATH points to the folder containing the forked libmtmd.so.

Plus:

2. Head-to-Head: The specialist vs. The Generalist

We contrasted LocateAnything-3B (The Eye) with Qwen3.5-9B-Neo-heretic (The Brain).

Feature	LocateAnything-3B (The Specialist)	Qwen3.5-9B (The Generalist)
Speed	🚀 Blazing Fast (PBD Architecture)	🐢 Slower (Autoregressive)
Precision	🎯 Pinpoint Bounding Boxes	🔍 Approximate "Areas"
Reasoning	❌ None (Keyword-driven)	✅ Deep (Philosophical/Logical)
UI Testing	Best for Click-point detection	Best for Behavioral analysis
Poetry/Humor	🤖 "I am a scanner."	🎭 "Let me tell you about Lenin..."

✅ Successes (The "Cerebellum" at work)

Iconography Recognition: Flawlessly identifies high-contrast symbols (e.g., the Dove of Peace).
Noun Grounding: High precision finding specific body parts or objects (e.g., worker's clenched fist, gray cap).
Spatial Awareness: Understands relative positioning commands (e.g., "Locate the most prominent object in the bottom right").
UI State Awareness: Distinguishes "blue on" vs "gray off" switches with 100% factual accuracy.

❌ Failures (Where the "Frontal Lobe" is missing)

Conceptual Reasoning: It cannot "Discuss Ideology." It treats abstract concepts as search queries for objects.
Occlusion/Inference: It fails at "hidden" objects. It lacks the "imagination" to infer what it cannot physically see (unlike Qwen3.5, which catches the "hidden guy").
Linguistic Noise: Highly sensitive to prompt clutter. It performs best with Atomic Nouns.

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