Instructions to use Sweaterdog/Andy-3.5-beta with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Sweaterdog/Andy-3.5-beta with Transformers:

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Sweaterdog/Andy-3.5-beta", dtype="auto")

llama-cpp-python

How to use Sweaterdog/Andy-3.5-beta with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Sweaterdog/Andy-3.5-beta",
	filename="Andy-3.5-beta-10.F16.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use Sweaterdog/Andy-3.5-beta with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Sweaterdog/Andy-3.5-beta:F16
# Run inference directly in the terminal:
llama-cli -hf Sweaterdog/Andy-3.5-beta:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Sweaterdog/Andy-3.5-beta:F16
# Run inference directly in the terminal:
llama-cli -hf Sweaterdog/Andy-3.5-beta:F16

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 Sweaterdog/Andy-3.5-beta:F16
# Run inference directly in the terminal:
./llama-cli -hf Sweaterdog/Andy-3.5-beta:F16

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 Sweaterdog/Andy-3.5-beta:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Sweaterdog/Andy-3.5-beta:F16

Use Docker

docker model run hf.co/Sweaterdog/Andy-3.5-beta:F16

LM Studio
Jan
Ollama
How to use Sweaterdog/Andy-3.5-beta with Ollama:
```
ollama run hf.co/Sweaterdog/Andy-3.5-beta:F16
```

Unsloth Studio new

How to use Sweaterdog/Andy-3.5-beta 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 Sweaterdog/Andy-3.5-beta 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 Sweaterdog/Andy-3.5-beta to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Sweaterdog/Andy-3.5-beta to start chatting

Docker Model Runner
How to use Sweaterdog/Andy-3.5-beta with Docker Model Runner:
```
docker model run hf.co/Sweaterdog/Andy-3.5-beta:F16
```

Lemonade

How to use Sweaterdog/Andy-3.5-beta with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Sweaterdog/Andy-3.5-beta:F16

Run and chat with the model

lemonade run user.Andy-3.5-beta-F16

List all available models

lemonade list

Uploaded models

Developed by: Sweaterdog
License: apache-2.0
Finetuned from model : unsloth/Qwen2.5-7B-bnb-4bit

DO NOT DOWNLOAD THESE MODELS, THEY ARE DEPRECATED AND WILL EVENTUALLY BE DELETED.

INSTEAD, DOWNLOAD THE LATEST ANDY-3.5 MODELS HERE

The MindCraft LLM tuning CSV file can be found here, this can be tweaked as needed. MindCraft-LLM

This is a very very early access Beta Model

This model is NOT a final version, but instead is a test to see how well models can be with a small dataset. This dataset is also a test of how smaller models can be improved from extremely high quality, and as close to real-world scenarios as possible.

This model listed here (Andy-3.5-beta-10) is NOT the final model, but instead a preview for the new training method, this model performs well at playing Minecraft and can even play with no instructions other than history. That all being said, this model was trained on a small dataset, meaning it doesn't have every single example it may need, the final version will have a much larger dataset. Also, if you want to use this model, you can test Modelfile, or Modelfile 2, I haven't had a chance to dive deep into which performs better, but the model is alright, it isn't the best, but better than a non-tuned model.

Where data came from

The storing memory parts are real examples from in-game interactions

The coding is artifical and was generated by GPT-o1, with the instruction to include reasoning and thinking in the comments of the code

The playing is artificial and was generated by me, a human, and used prompts focusing on points where some models fail, such as mining.

This model should not be a reflection on how smaller models play Minecraft, if it performs well, and better than Andy-v2-qwen, then Yay! If not, I wasn't expecting it to be better, (And neither should you!)

You are totally allowed to test the beta model.

I hope this model performs well for you!

How you can help make Andy-3.5

I am one person, who is paying and working for this model, I want it to be amazing for all people to play with it for free. However, that is not feasible finacially for one person like me. IF you would like to help out, you can find me on discord, via the MindCraft server., my username on there is "Sweaterdog"

ALSO

The models are going to change, I am changing hyperparameters on tuning to (hopefully) increase performance and decrease hallucinations

BTW, if you want to download this model, I suggest using llama.cpp to make a quantization of it, I would have done it during tuning but I ran out of GPU time on google colab