Unsloth MCP Server

Unsloth MCP Server

Provides tools for optimizing, fine-tuning, and deploying large language models with Unsloth, enabling 2x faster training with 80% less memory through model loading, fine-tuning, text generation, and model export capabilities.

OtotaO

Research & Data
Visit Server

README

Unsloth MCP Server

An MCP server for Unsloth - a library that makes LLM fine-tuning 2x faster with 80% less memory.

What is Unsloth?

Unsloth is a library that dramatically improves the efficiency of fine-tuning large language models:

  • Speed: 2x faster fine-tuning compared to standard methods
  • Memory: 80% less VRAM usage, allowing fine-tuning of larger models on consumer GPUs
  • Context Length: Up to 13x longer context lengths (e.g., 89K tokens for Llama 3.3 on 80GB GPUs)
  • Accuracy: No loss in model quality or performance

Unsloth achieves these improvements through custom CUDA kernels written in OpenAI's Triton language, optimized backpropagation, and dynamic 4-bit quantization.

Features

  • Optimize fine-tuning for Llama, Mistral, Phi, Gemma, and other models
  • 4-bit quantization for efficient training
  • Extended context length support
  • Simple API for model loading, fine-tuning, and inference
  • Export to various formats (GGUF, Hugging Face, etc.)

Quick Start

  1. Install Unsloth: pip install unsloth
  2. Install and build the server:
    cd unsloth-server
    npm install
    npm run build
    
  3. Add to MCP settings:
    {
      "mcpServers": {
        "unsloth-server": {
          "command": "node",
          "args": ["/path/to/unsloth-server/build/index.js"],
          "env": {
            "HUGGINGFACE_TOKEN": "your_token_here" // Optional
          },
          "disabled": false,
          "autoApprove": []
        }
      }
    }
    

Available Tools

check_installation

Verify if Unsloth is properly installed on your system.

Parameters: None

Example:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "check_installation",
  arguments: {}
});

list_supported_models

Get a list of all models supported by Unsloth, including Llama, Mistral, Phi, and Gemma variants.

Parameters: None

Example:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "list_supported_models",
  arguments: {}
});

load_model

Load a pretrained model with Unsloth optimizations for faster inference and fine-tuning.

Parameters:

  • model_name (required): Name of the model to load (e.g., "unsloth/Llama-3.2-1B")
  • max_seq_length (optional): Maximum sequence length for the model (default: 2048)
  • load_in_4bit (optional): Whether to load the model in 4-bit quantization (default: true)
  • use_gradient_checkpointing (optional): Whether to use gradient checkpointing to save memory (default: true)

Example:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "load_model",
  arguments: {
    model_name: "unsloth/Llama-3.2-1B",
    max_seq_length: 4096,
    load_in_4bit: true
  }
});

finetune_model

Fine-tune a model with Unsloth optimizations using LoRA/QLoRA techniques.

Parameters:

  • model_name (required): Name of the model to fine-tune
  • dataset_name (required): Name of the dataset to use for fine-tuning
  • output_dir (required): Directory to save the fine-tuned model
  • max_seq_length (optional): Maximum sequence length for training (default: 2048)
  • lora_rank (optional): Rank for LoRA fine-tuning (default: 16)
  • lora_alpha (optional): Alpha for LoRA fine-tuning (default: 16)
  • batch_size (optional): Batch size for training (default: 2)
  • gradient_accumulation_steps (optional): Number of gradient accumulation steps (default: 4)
  • learning_rate (optional): Learning rate for training (default: 2e-4)
  • max_steps (optional): Maximum number of training steps (default: 100)
  • dataset_text_field (optional): Field in the dataset containing the text (default: 'text')
  • load_in_4bit (optional): Whether to use 4-bit quantization (default: true)

Example:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "finetune_model",
  arguments: {
    model_name: "unsloth/Llama-3.2-1B",
    dataset_name: "tatsu-lab/alpaca",
    output_dir: "./fine-tuned-model",
    max_steps: 100,
    batch_size: 2,
    learning_rate: 2e-4
  }
});

generate_text

Generate text using a fine-tuned Unsloth model.

Parameters:

  • model_path (required): Path to the fine-tuned model
  • prompt (required): Prompt for text generation
  • max_new_tokens (optional): Maximum number of tokens to generate (default: 256)
  • temperature (optional): Temperature for text generation (default: 0.7)
  • top_p (optional): Top-p for text generation (default: 0.9)

Example:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "generate_text",
  arguments: {
    model_path: "./fine-tuned-model",
    prompt: "Write a short story about a robot learning to paint:",
    max_new_tokens: 512,
    temperature: 0.8
  }
});

export_model

Export a fine-tuned Unsloth model to various formats for deployment.

Parameters:

  • model_path (required): Path to the fine-tuned model
  • export_format (required): Format to export to (gguf, ollama, vllm, huggingface)
  • output_path (required): Path to save the exported model
  • quantization_bits (optional): Bits for quantization (for GGUF export) (default: 4)

Example:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "export_model",
  arguments: {
    model_path: "./fine-tuned-model",
    export_format: "gguf",
    output_path: "./exported-model.gguf",
    quantization_bits: 4
  }
});

Advanced Usage

Custom Datasets

You can use custom datasets by formatting them properly and hosting them on Hugging Face or providing a local path:

const result = await use_mcp_tool({
  server_name: "unsloth-server",
  tool_name: "finetune_model",
  arguments: {
    model_name: "unsloth/Llama-3.2-1B",
    dataset_name: "json",
    data_files: {"train": "path/to/your/data.json"},
    output_dir: "./fine-tuned-model"
  }
});

Memory Optimization

For large models on limited hardware:

  • Reduce batch size and increase gradient accumulation steps
  • Use 4-bit quantization
  • Enable gradient checkpointing
  • Reduce sequence length if possible

Troubleshooting

Common Issues

  1. CUDA Out of Memory: Reduce batch size, use 4-bit quantization, or try a smaller model
  2. Import Errors: Ensure you have the correct versions of torch, transformers, and unsloth installed
  3. Model Not Found: Check that you're using a supported model name or have access to private models

Version Compatibility

  • Python: 3.10, 3.11, or 3.12 (not 3.13)
  • CUDA: 11.8 or 12.1+ recommended
  • PyTorch: 2.0+ recommended

Performance Benchmarks

Model VRAM Unsloth Speed VRAM Reduction Context Length
Llama 3.3 (70B) 80GB 2x faster >75% 13x longer
Llama 3.1 (8B) 80GB 2x faster >70% 12x longer
Mistral v0.3 (7B) 80GB 2.2x faster 75% less -

Requirements

  • Python 3.10-3.12
  • NVIDIA GPU with CUDA support (recommended)
  • Node.js and npm

License

Apache-2.0

Recommended Servers

Crypto Price & Market Analysis MCP Server

Crypto Price & Market Analysis MCP Server

A Model Context Protocol (MCP) server that provides comprehensive cryptocurrency analysis using the CoinCap API. This server offers real-time price data, market analysis, and historical trends through an easy-to-use interface.

Featured
TypeScript
MCP PubMed Search

MCP PubMed Search

Server to search PubMed (PubMed is a free, online database that allows users to search for biomedical and life sciences literature). I have created on a day MCP came out but was on vacation, I saw someone post similar server in your DB, but figured to post mine.

Featured
Python
dbt Semantic Layer MCP Server

dbt Semantic Layer MCP Server

A server that enables querying the dbt Semantic Layer through natural language conversations with Claude Desktop and other AI assistants, allowing users to discover metrics, create queries, analyze data, and visualize results.

Featured
TypeScript
mixpanel

mixpanel

Connect to your Mixpanel data. Query events, retention, and funnel data from Mixpanel analytics.

Featured
TypeScript
Sequential Thinking MCP Server

Sequential Thinking MCP Server

This server facilitates structured problem-solving by breaking down complex issues into sequential steps, supporting revisions, and enabling multiple solution paths through full MCP integration.

Featured
Python
Nefino MCP Server

Nefino MCP Server

Provides large language models with access to news and information about renewable energy projects in Germany, allowing filtering by location, topic (solar, wind, hydrogen), and date range.

Official
Python
Vectorize

Vectorize

Vectorize MCP server for advanced retrieval, Private Deep Research, Anything-to-Markdown file extraction and text chunking.

Official
JavaScript
Mathematica Documentation MCP server

Mathematica Documentation MCP server

A server that provides access to Mathematica documentation through FastMCP, enabling users to retrieve function documentation and list package symbols from Wolfram Mathematica.

Local
Python
kb-mcp-server

kb-mcp-server

An MCP server aimed to be portable, local, easy and convenient to support semantic/graph based retrieval of txtai "all in one" embeddings database. Any txtai embeddings db in tar.gz form can be loaded

Local
Python
Research MCP Server

Research MCP Server

The server functions as an MCP server to interact with Notion for retrieving and creating survey data, integrating with the Claude Desktop Client for conducting and reviewing surveys.

Local
Python