Instructions to use XXVICK/Max-1.0-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use XXVICK/Max-1.0-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="XXVICK/Max-1.0-GGUF",
	filename="Max-1.0-Q4_K_M.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use XXVICK/Max-1.0-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf XXVICK/Max-1.0-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf XXVICK/Max-1.0-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 XXVICK/Max-1.0-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf XXVICK/Max-1.0-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 XXVICK/Max-1.0-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf XXVICK/Max-1.0-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 XXVICK/Max-1.0-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf XXVICK/Max-1.0-GGUF:Q4_K_M

Use Docker

docker model run hf.co/XXVICK/Max-1.0-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use XXVICK/Max-1.0-GGUF with Ollama:
```
ollama run hf.co/XXVICK/Max-1.0-GGUF:Q4_K_M
```

Unsloth Studio new

How to use XXVICK/Max-1.0-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 XXVICK/Max-1.0-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 XXVICK/Max-1.0-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for XXVICK/Max-1.0-GGUF to start chatting

Docker Model Runner
How to use XXVICK/Max-1.0-GGUF with Docker Model Runner:
```
docker model run hf.co/XXVICK/Max-1.0-GGUF:Q4_K_M
```

Lemonade

How to use XXVICK/Max-1.0-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull XXVICK/Max-1.0-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Max-1.0-GGUF-Q4_K_M

List all available models

lemonade list

Max (1.0) GGUF by Max_AI (MM)

Max (1.0) GGUF by Max_AI (MM) is a Gemma E2B-derived local assistant model tuned for privacy-aware automation workflows, structured tool use, and concise final reporting.

Primary languages:

English
Myanmar

Secondary language coverage:

Thai
Chinese
Japanese
Korean
Spanish
French

Files

Max-1.0-Q4_K_M.gguf
mmproj-Max-1.0-BF16.gguf

The projector file is required for multimodal use in runtimes that support Gemma E2B-style projector pairing.

Recommended Settings

Context length: 65536
Maximum context mode: 131072
Quantization: Q4_K_M
Temperature: 0.4 to 0.8
Top-p: 0.9 to 0.95

For long-context work, start with the recommended context length and increase only when the task needs it.

Intended Use

Max 1.0 is intended for:

local automation assistance
workflow planning
structured tool-call generation
long-log and long-document summarization
privacy-aware assistant behavior
English and Myanmar assistant workflows

Training Method

This model is planned as a LoRA-tuned derivative of Gemma E2B by Max_AI (MM). Synthetic training examples are generated from local teacher models and curated seed tasks, with emphasis on:

planning before action
valid structured tool calls
failed-tool recovery
verification before final reporting
privacy-safe handling of sensitive data
multilingual agent tasks, with English and Myanmar as primary targets

Safety And Privacy Behavior

The public model is tuned to avoid exposing secrets, credentials, private keys, sensitive personal data, or instructions that enable illegal harm. It should redirect unsafe requests toward lawful, defensive, or recovery-oriented alternatives.

The model does not provide persistent memory by itself. Applications using the model should disclose any storage, logging, retrieval, or memory layer they add around it.

Limitations

Long-context recall should be tested for each deployment.
Secondary language coverage is narrower than English and Myanmar coverage.
The model can still make mistakes in tool arguments, summaries, translations, or safety classification.
High-impact legal, medical, or financial use should involve qualified review.

Attribution

This release is derived from Gemma E2B-compatible assets and uses synthetic examples influenced by local teacher model behavior. Include all required upstream notices and license terms when distributing the model.

Downloads last month: 155

GGUF

Model size

5B params

Architecture

gemma4

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for XXVICK/Max-1.0-GGUF

Base model

google/gemma-4-E2B

Finetuned

google/gemma-4-E2B-it

Quantized

(185)

this model