Instructions to use aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA

Run Hermes

hermes

MLX LM

How to use aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

AICoven Llama 3.2 3B — MCP Tool Calling

A LoRA fine-tuned version of Llama 3.2 3B Instruct (4-bit MLX) optimized for MCP (Model Context Protocol) tool calling in the AICoven app.

This model runs 100% on-device on Apple Silicon via MLX.

Model Details


Base Model	`mlx-community/Llama-3.2-3B-Instruct-4bit`
Architecture	LlamaForCausalLM
Quantization	4-bit (group size 64)
Fine-tuning	LoRA (rank 16, alpha 32, 16 layers)
Framework	MLX / mlx-lm
Size	~1.8 GB

Training

Fine-tuned on Apple M4 (24GB) using mlx-lm LoRA with gradient checkpointing.

Dataset: 177 synthetic examples in Chat-ML format
- 114 single tool calls
- 32 multi-turn (tool → result → response chains)
- 31 no-tool / conversational (negative examples to reduce over-triggering)
Tool Coverage: 30 MCP tools across GitHub, Google Workspace (Gmail, Calendar, Drive, Docs, Sheets, Slides), Slack, Notion, Trello, TickTick, GA4, and more
Hyperparameters: lr=1e-4, batch_size=1, 150 iterations, max_seq_length=3072
Training Loss: Converged from 1.004 → 0.010 (val: 0.013)
Peak Memory: 6.4 GB

Evaluation

Tested on a 50-example novel test set (prompts never seen during training):

Metric	Result
Accuracy	86% (43/50)
Refusals	0%
Format Errors	2% (1/50)
Unambiguous Tool Selection	94%+

The model outputs strict JSON tool calls without markdown code fences or conversational fluff.

Intended Use

This model is designed for the AICoven macOS/iOS app to provide local, private AI agent capabilities. It selects and invokes MCP tools based on user requests, supporting:

Single tool calls (e.g., "What time is it in Tokyo?")
Multi-step reasoning chains (e.g., "Find Python files in Documents and count them")
Graceful no-tool responses for conversational queries

How to Use

from mlx_lm import load, generate

model, tokenizer = load("aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA")
response = generate(model, tokenizer, prompt="What's the weather like?", max_tokens=256)

Limitations

Optimized specifically for AICoven's tool schema; may not generalize to arbitrary tool-calling formats
3B parameter model — best for well-defined tool selection, not open-ended reasoning
Requires Apple Silicon (M1+) for MLX inference

License

This model inherits the Llama 3.2 Community License from Meta.

Built with Llama.

Downloads last month: 20

Safetensors

Model size

0.5B params

Tensor type

F16

U32

MLX

Hardware compatibility

4-bit

Model tree for aicoven/Llama-3.2-3B-Instruct-4bit-MCP-LoRA

Base model

meta-llama/Llama-3.2-3B-Instruct

Finetuned

mlx-community/Llama-3.2-3B-Instruct

Quantized

mlx-community/Llama-3.2-3B-Instruct-4bit

Adapter

(5)

this model