Instructions to use forge-community/gemma3-1b-functioncall-exp-mlx-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use forge-community/gemma3-1b-functioncall-exp-mlx-4bit 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("forge-community/gemma3-1b-functioncall-exp-mlx-4bit")

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 forge-community/gemma3-1b-functioncall-exp-mlx-4bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "forge-community/gemma3-1b-functioncall-exp-mlx-4bit"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "forge-community/gemma3-1b-functioncall-exp-mlx-4bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use forge-community/gemma3-1b-functioncall-exp-mlx-4bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "forge-community/gemma3-1b-functioncall-exp-mlx-4bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "forge-community/gemma3-1b-functioncall-exp-mlx-4bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "forge-community/gemma3-1b-functioncall-exp-mlx-4bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use forge-community/gemma3-1b-functioncall-exp-mlx-4bit 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 "forge-community/gemma3-1b-functioncall-exp-mlx-4bit"

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 forge-community/gemma3-1b-functioncall-exp-mlx-4bit

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use forge-community/gemma3-1b-functioncall-exp-mlx-4bit with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "forge-community/gemma3-1b-functioncall-exp-mlx-4bit"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "forge-community/gemma3-1b-functioncall-exp-mlx-4bit" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

gemma3-1b-functioncall-exp-mlx-4bit / README.md

adithyasean

Card, results and figures from the 2026 close-out evaluation

796bc12 verified 23 days ago

preview code

Raw History Blame Contribute Delete

2.11 kB

metadata

base_model: google/gemma-3-1b-it
library_name: mlx
license: gemma
language:
  - en
pipeline_tag: text-generation
tags:
  - mlx
  - function-calling
  - tool-use
  - forge
datasets:
  - Salesforce/xlam-function-calling-60k
  - Team-ACE/ToolACE

gemma3-1b-functioncall-exp-mlx-4bit

Gemma 3 1B (instruction-tuned) after one general function-calling fine-tuning stage, following the FunctionGemma recipe, quantized to 4-bit for MLX. It is the base of the four round-2 Forge specialists:

forge-community/gemma3-1b-smarthome-r2-it1 (smart-home)
forge-community/gemma3-1b-music-r2-it2 (youtube-music)
forge-community/gemma3-1b-gworkspace-r2-it2 (google-workspace)
forge-community/gemma3-1b-moviestore-r2-it3 (movie-store)

The full-precision weights are in forge-community/gemma3-1b-functioncall-exp.

Why this base

Chosen by a pre-registered rule (best smart-home dev probe, ties by suite) on the same data and gate: this base scored probe 32/37 and suite 57/68 against plain Gemma 3 1B's 31/37 and 54/68. The margin is within noise and is reported as such.

Training data

The function-calling stage used Salesforce/xlam-function-calling-60k (revision 26d14eb, SHA-256 checked against the official file) and Team-ACE/ToolACE. Their licences and terms apply in addition to the Gemma terms.

Citation

@thesis{ekanayaka2026forge,
  author = {Ekanayaka, Adithya},
  title  = {Agent-Operable Specialization of Small Tool-Calling Language Models at the Model Level},
  type   = {BSc (Hons) dissertation},
  school = {Coventry University and National Institute of Business Management},
  year   = {2026}
}