🧠 Andy‑4 ⛏️

Andy‑4 is an 8 billion‑parameter specialist model tuned for Minecraft gameplay via the Mindcraft framework. Trained on a single RTX 3090 over three weeks, Andy‑4 delivers advanced reasoning, multi‑step planning, and robust in‑game decision‑making.

⚠️ Certification:
Andy‑4 is not yet certified by the Mindcraft developers. Use in production at your own discretion.

🔍 Model Specifications

Parameters: 8 B
Training Hardware: 1 × NVIDIA RTX 3090
Duration: ~3 weeks total
Data Volumes:
- Messages: 179,384
- Tokens: 425,535,198
- Conversations: 62,149
Base Architecture: Llama 3.1 8B
License: Andy 1.1 License
Repository: https://huggingface.co/Sweaterdog/Andy‑4

📊 Training Regimen

Andy‑4‑base‑1 dataset
- Epochs: 2
- Learning Rate: 7e-5
- Dataset Size: 47.4k
Andy‑4‑base‑2 dataset
- Epochs: 4
- Learning Rate: 3e-7
- Dataset Size: 48.9k
Fine‑tune (FT) dataset
- Epochs: 2.5
- Learning Rate: 2e-5
- Dataset Size: 4.12k

Optimizer: AdamW_8bit with cosine decay
Quantization: 4‑bit (bnb-4bit) for inference
Warm Up Steps: 0.1% of each dataset

🚀 Installation

First, you need to choose your quantization, this chart is with the base of 8192 set as the context window

Quantization	VRAM Required
F16	16 GB+
Q5_K_M	8 GB+
Q4_K_M	6–8 GB
Q3_K_M	6 GB (low)
Q2_K	4–6 GB (ultra)

1. Installation directly on Ollama

Visit Andy-4 on Ollama
Copy the command after choosing model type / quantization
Run the command in the terminal
Set the profile's model to be what you installed, such as ollama/sweaterdog/andy-4:latest

2. Manual Download & Modelfile

Download
- From the HF Files tab, grab your chosen .GGUF quant weights (e.g. Andy-4.Q4_K_M.gguf).
- Download the provided Modelfile.

Edit

Change

FROM YOUR/PATH/HERE

FROM /path/to/Andy-4.Q4_K_M.gguf

Optional: Increase the parameter num_ctx to a higher value for longer conversations if you:

A. Have extra VRAM

B. Quantized the context window

C. Can use a smaller model

Create
```
ollama create andy-4 -f Modelfile
```

This registers the Andy‑4 model locally.

If you lack a GPU, check the Mindcraft Discord guide for free cloud setups.

🔧 Context‑Window Quantization

To lower VRAM use for context windows:

Windows

Close Ollama.

In System Properties → Environment Variables, add:

OLLAMA_FLASH_ATTENTION=1  
OLLAMA_KV_CACHE_TYPE=q8_0   # or q4_0 for extra savings, but far more unstable

Restart Ollama.

Linux/macOS

export OLLAMA_FLASH_ATTENTION=1
export OLLAMA_KV_CACHE_TYPE="q8_0"   # or "q4_0", but far more unstable
ollama serve

📌 Acknowledgments

Click to expand

Data & Models by: @Sweaterdog
Framework: Mindcraft (https://github.com/kolbytn/mindcraft)
LoRA Weights: https://huggingface.co/Sweaterdog/Andy-4-LoRA

⚖️ License

See Andy 1.1 License.

This work uses data and models created by @Sweaterdog.

Sweaterdog
/

Andy-4