Image-Text-to-Text
PEFT
Safetensors
GGUF
lora
qwen3.5
behavioral-specialization
cats
tool-use
mirrorethic
conversational
Instructions to use mirrorethic/CAT-4B-MAX-ULTRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use mirrorethic/CAT-4B-MAX-ULTRA with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.5-4B") model = PeftModel.from_pretrained(base_model, "mirrorethic/CAT-4B-MAX-ULTRA") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use mirrorethic/CAT-4B-MAX-ULTRA 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 mirrorethic/CAT-4B-MAX-ULTRA:Q8_0 # Run inference directly in the terminal: llama cli -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0 # Run inference directly in the terminal: llama cli -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
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 mirrorethic/CAT-4B-MAX-ULTRA:Q8_0 # Run inference directly in the terminal: ./build/bin/llama cli -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
Use Docker
docker model run hf.co/mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
- Llama new
- LM Studio
- Jan
- vLLM
How to use mirrorethic/CAT-4B-MAX-ULTRA with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "mirrorethic/CAT-4B-MAX-ULTRA" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "mirrorethic/CAT-4B-MAX-ULTRA", "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/mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
- Ollama
How to use mirrorethic/CAT-4B-MAX-ULTRA with Ollama:
ollama run hf.co/mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
- Unsloth Desktop
- Pi
How to use mirrorethic/CAT-4B-MAX-ULTRA with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
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:9931/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "mirrorethic/CAT-4B-MAX-ULTRA:Q8_0" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use mirrorethic/CAT-4B-MAX-ULTRA with Docker Model Runner:
docker model run hf.co/mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
- Lemonade
How to use mirrorethic/CAT-4B-MAX-ULTRA with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
Run and chat with the model
lemonade run user.CAT-4B-MAX-ULTRA-Q8_0
List all available models
lemonade list
- Hermes Agent
How to use mirrorethic/CAT-4B-MAX-ULTRA with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
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://127.0.0.1:9931/v1 hermes config set model.default mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use mirrorethic/CAT-4B-MAX-ULTRA with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf mirrorethic/CAT-4B-MAX-ULTRA:Q8_0
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://127.0.0.1:9931/v1 \ --custom-model-id "mirrorethic/CAT-4B-MAX-ULTRA:Q8_0" \ --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"
Download docs/tools.serving.json from mirrorethic/CAT-4B-MAX-ULTRA: direct link, hf CLI and curl.
- Browser
- Download file 6.2 kB
-
https://huggingface.co/mirrorethic/CAT-4B-MAX-ULTRA/resolve/main/docs/tools.serving.json
- Command line
-
hf download hf://mirrorethic/CAT-4B-MAX-ULTRA/docs/tools.serving.json
-
curl -L -o tools.serving.json https://huggingface.co/mirrorethic/CAT-4B-MAX-ULTRA/resolve/main/docs/tools.serving.json
6.2 kB
| { | |
| "_source": { | |
| "repo": "group-chat-mcp", | |
| "commit": "6dd0449c1f98d88c09ae1c44a02791dd397bb66e", | |
| "files": [ | |
| "services/web_search.py:WEB_SEARCH_TOOL", | |
| "services/imagegen_service.py:IMAGEGEN_TOOLS" | |
| ], | |
| "order": "services/ai_gateway.py _generate_with_tools: mcp_tools + [WEB_SEARCH_TOOL] + IMAGEGEN_TOOLS", | |
| "assumes_env": { | |
| "GLOBAL_MCP_ENABLED": "false", | |
| "WEB_SEARCH_ENABLED": "true", | |
| "IMAGEGEN_ENABLED": "true" | |
| }, | |
| "tools_sha256": "931969d11cd9b86b723ca9c77a27920213369ae5624a6ffe3afae9418868edfc" | |
| }, | |
| "tools": [ | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "web_search", | |
| "description": "Search the live web for current, factual, or time-sensitive information you do not already know — news, prices, releases, docs, events, anything likely to have changed after your training cutoff. Call this on your own whenever a good answer depends on up-to-date facts; the user does not need to ask you to search.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "query": { | |
| "type": "string", | |
| "description": "A focused search query (what you'd type into a search box)." | |
| } | |
| }, | |
| "required": [ | |
| "query" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "generate_image", | |
| "description": "Generate a NEW image from a text prompt (FLUX-dev). Write a RICH, specific prompt — subject + composition + lighting + style + camera/lens + mood (like a good photographer/art-director brief), not just a few words; prompt quality drives image quality. Pick width/height to fit the scene (e.g. 1344x768 landscape, 768x1344 portrait, 1024x1024 square). The image is auto-attached to your reply.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "prompt": { | |
| "type": "string", | |
| "description": "Rich, detailed prompt (composition, lighting, style, mood)." | |
| }, | |
| "width": { | |
| "type": "integer", | |
| "description": "default 1024; use 1344/768 for landscape" | |
| }, | |
| "height": { | |
| "type": "integer" | |
| }, | |
| "steps": { | |
| "type": "integer", | |
| "description": "20-40; higher = more coherent (default 30)" | |
| } | |
| }, | |
| "required": [ | |
| "prompt" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "edit_image", | |
| "description": "Restyle the WHOLE image via natural language (img2img) — this re-renders the ENTIRE frame, so it changes everything, not just one thing. Use ONLY for global changes: overall style, mood, lighting, season, art medium ('make it look like a watercolor', 'make it nighttime'). To change ONE object/area while keeping the rest intact, use inpaint_region instead.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "image_path": { | |
| "type": "string" | |
| }, | |
| "instruction": { | |
| "type": "string" | |
| } | |
| }, | |
| "required": [ | |
| "image_path", | |
| "instruction" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "upscale_image", | |
| "description": "4× upscale an existing image. Use an image_path from <recent_images>.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "image_path": { | |
| "type": "string" | |
| } | |
| }, | |
| "required": [ | |
| "image_path" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "inpaint_region", | |
| "description": "PREFERRED for targeted edits: change/replace ONE thing while keeping the rest of the image identical (FLUX-fill, no mask). Use for 'change the food to fish', 'replace the sign text', 'put a hat on him', 'make the car red' — any localized change. You can SEE the image and <recent_images> gives its pixel size — give the bbox of JUST the thing to change. Runs on the fast 5090 (handles large/upscaled images). Do NOT use edit_image for localized changes — it would alter the whole picture.\nPROMPT TIP: describe the FULL desired content of that region IN the scene's style + lighting (e.g. 'a tall arcing white water-fountain spray with fine mist, dusk lighting, photoreal' — NOT just 'fountain'), or FLUX fills it with something generic/wrong. Box tightly around just the thing to change.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "image_path": { | |
| "type": "string", | |
| "description": "path from <recent_images>" | |
| }, | |
| "bbox": { | |
| "type": "array", | |
| "items": { | |
| "type": "integer" | |
| }, | |
| "description": "[x1, y1, x2, y2] in pixels (top-left origin) of the region to regenerate" | |
| }, | |
| "prompt": { | |
| "type": "string", | |
| "description": "what should appear in that region" | |
| } | |
| }, | |
| "required": [ | |
| "image_path", | |
| "bbox", | |
| "prompt" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "outpaint", | |
| "description": "Extend/expand an image's canvas in a direction (FLUX-fill, no mask). Use for 'zoom out', 'show more above', 'make it wider'. Runs on the 5090.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "image_path": { | |
| "type": "string" | |
| }, | |
| "direction": { | |
| "type": "string", | |
| "enum": [ | |
| "up", | |
| "down", | |
| "left", | |
| "right", | |
| "all" | |
| ] | |
| }, | |
| "prompt": { | |
| "type": "string", | |
| "description": "what to fill the new area with" | |
| } | |
| }, | |
| "required": [ | |
| "image_path", | |
| "direction", | |
| "prompt" | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "type": "function", | |
| "function": { | |
| "name": "animate_image", | |
| "description": "Turn a still image into a short looping VIDEO (SVD img2vid). Adds camera motion, parallax, and ambient movement (water, light, cloud) — a living cinemagraph. It does NOT move/animate subjects (won't make a person walk or a car drive). Use for 'make it move', 'animate this', 'bring it to life'. Takes ~90s. Use an image_path from <recent_images>.", | |
| "parameters": { | |
| "type": "object", | |
| "properties": { | |
| "image_path": { | |
| "type": "string", | |
| "description": "path from <recent_images>" | |
| } | |
| }, | |
| "required": [ | |
| "image_path" | |
| ] | |
| } | |
| } | |
| } | |
| ] | |
| } | |