Instructions to use AesSedai/MiniMax-M2.5-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AesSedai/MiniMax-M2.5-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="AesSedai/MiniMax-M2.5-GGUF",
	filename="IQ3_S/MiniMax-M2.5-IQ3_S-00001-of-00003.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 AesSedai/MiniMax-M2.5-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf AesSedai/MiniMax-M2.5-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf AesSedai/MiniMax-M2.5-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf AesSedai/MiniMax-M2.5-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf AesSedai/MiniMax-M2.5-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 AesSedai/MiniMax-M2.5-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf AesSedai/MiniMax-M2.5-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 AesSedai/MiniMax-M2.5-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf AesSedai/MiniMax-M2.5-GGUF:Q4_K_M

Use Docker

docker model run hf.co/AesSedai/MiniMax-M2.5-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use AesSedai/MiniMax-M2.5-GGUF with Ollama:
```
ollama run hf.co/AesSedai/MiniMax-M2.5-GGUF:Q4_K_M
```

Unsloth Studio new

How to use AesSedai/MiniMax-M2.5-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 AesSedai/MiniMax-M2.5-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 AesSedai/MiniMax-M2.5-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for AesSedai/MiniMax-M2.5-GGUF to start chatting

Pi new

How to use AesSedai/MiniMax-M2.5-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf AesSedai/MiniMax-M2.5-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": "MiniMax-M2.5-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use AesSedai/MiniMax-M2.5-GGUF with Docker Model Runner:
```
docker model run hf.co/AesSedai/MiniMax-M2.5-GGUF:Q4_K_M
```

Lemonade

How to use AesSedai/MiniMax-M2.5-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull AesSedai/MiniMax-M2.5-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.MiniMax-M2.5-GGUF-Q4_K_M

List all available models

lemonade list

This repo contains specialized MoE-quants for MiniMax-M2.5. The idea being that given the huge size of the FFN tensors compared to the rest of the tensors in the model, it should be possible to achieve a better quality while keeping the overall size of the entire model smaller compared to a similar naive quantization. To that end, the quantization type default is kept in high quality and the FFN UP + FFN GATE tensors are quanted down along with the FFN DOWN tensors.

Quant	Size	Mixture	PPL	1-(Mean PPL(Q)/PPL(base))	KLD
Q5_K_M	157.23 GiB (5.91 BPW)	Q8_0 / Q5_K / Q5_K / Q6_K	7.126261 ± 0.115850	+0.5877%	0.023465 ± 0.001079
Q4_K_M	130.52 GiB (4.90 BPW)	Q8_0 / Q4_K / Q4_K / Q5_K	7.173459 ± 0.116673	+1.2462%	0.041269 ± 0.001426
IQ4_XS	101.10 GiB (3.80 BPW)	Q8_0 / IQ3_S / IQ3_S / IQ4_XS	7.513587 ± 0.122746	+6.0549%	0.095077 ± 0.002168
IQ3_S	78.76 GiB (2.96 BPW)	Q8_0 / IQ2_S / IQ2_S / IQ3_S	8.284882 ± 0.135705	+16.9418%	0.244096 ± 0.004148

Provided here as well as a couple of graphs showing the Pareto frontier for KLD and PPL for my quants vs Unsloth.

Full graphs of all of the quants are available in the kld_data directory, as well as the raw data broken down per quant as well as a CSV with the collated data.

While the PPL between the quant methods is similar, I feel like the KLD of the quants provided here are slightly better and that these quants will offer better long context performance due to keeping the default type as Q8_0. This comes with a slight performance penalty in PP / TG due to the higher quality quantization but I think the tradeoff is worthwhile.