Instructions to use Trilogix1/Hugston-Macaron-V1-Tall with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Trilogix1/Hugston-Macaron-V1-Tall 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 Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Trilogix1/Hugston-Macaron-V1-Tall: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 Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Trilogix1/Hugston-Macaron-V1-Tall: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 Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M

Use Docker

docker model run hf.co/Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M

LM Studio
Jan
Ollama
How to use Trilogix1/Hugston-Macaron-V1-Tall with Ollama:
```
ollama run hf.co/Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M
```

Unsloth Studio

How to use Trilogix1/Hugston-Macaron-V1-Tall 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 Trilogix1/Hugston-Macaron-V1-Tall 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 Trilogix1/Hugston-Macaron-V1-Tall to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Trilogix1/Hugston-Macaron-V1-Tall to start chatting

How to use Trilogix1/Hugston-Macaron-V1-Tall with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Trilogix1/Hugston-Macaron-V1-Tall: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": "Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Trilogix1/Hugston-Macaron-V1-Tall with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Trilogix1/Hugston-Macaron-V1-Tall: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 Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use Trilogix1/Hugston-Macaron-V1-Tall with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Trilogix1/Hugston-Macaron-V1-Tall: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 "Trilogix1/Hugston-Macaron-V1-Tall: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 Trilogix1/Hugston-Macaron-V1-Tall with Docker Model Runner:
```
docker model run hf.co/Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M
```

Lemonade

How to use Trilogix1/Hugston-Macaron-V1-Tall with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Trilogix1/Hugston-Macaron-V1-Tall:Q4_K_M

Run and chat with the model

lemonade run user.Hugston-Macaron-V1-Tall-Q4_K_M

List all available models

lemonade list

the loars files

by suncityldp - opened 6 days ago

Discussion

suncityldp

6 days ago

mindlab-research/Macaron-V1-Tall,that loars folder：four LoRA specialists for chat, personal-agent tasks, coding, and GenUI,I think you haven't quantized them.

Trilogix1

Owner 6 days ago

Can you clarify your statement, is difficult to understand what do you mean exactly. This particular model was quantized by Hugston (you notice easy as it was done with an old build that didnt include mtp).
I hope this answer your question.

suncityldp

6 days ago

Macaron-V1 Tall：A Qwen3.6-35B-A3B-based Macaron-V1 model with the full Macaron specialist system.
Mixture of LoRA (MoL) architecture with four specialists across Chat, Agent, Coding, and GenUI.
BF16 checkpoint with four LoRA adapters stored under loras/L0 through loras/L3.，each of them are 7.55GB，so it should be Qwen3.6-35B-A3B + 4*7.55GB，I'm not sure if your version includes these things.
Hugston-Macaron-V1-Tall

Trilogix1

Owner 6 days ago

•

edited 6 days ago

The model should be a composed macaron method already, the loras should be optional. Well if you merge only one or many of them, do not change the original model size :) but you are right on point, something better may come out. However we choose coding loras usually, hopefully you will see an update soon.

Here the coding lora merged (not what I expected) : https://hugston.com/models/macaron-v1-tall-coding-loras-merged-hf-20260726-164843q4-k-m It will be there for some days in case of public interest. Would upload it in HF but is not worth it better the original Macaron with no loras that is already uploaded here.

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