Instructions to use Ruiruiz30/Jev-Omni-MLX-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ruiruiz30/Jev-Omni-MLX-4bit with MLX:

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

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("Ruiruiz30/Jev-Omni-MLX-4bit")
config = load_config("Ruiruiz30/Jev-Omni-MLX-4bit")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use Ruiruiz30/Jev-Omni-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 "Ruiruiz30/Jev-Omni-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": "Ruiruiz30/Jev-Omni-MLX-4bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use Ruiruiz30/Jev-Omni-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 "Ruiruiz30/Jev-Omni-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 Ruiruiz30/Jev-Omni-MLX-4bit

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Ruiruiz30/Jev-Omni-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 "Ruiruiz30/Jev-Omni-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 "Ruiruiz30/Jev-Omni-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"

Jev-Omni-MLX-4bit / assets /jevbench-accuracy.svg

Ruiruiz30

correct JevBench group-macro reporting

3ec2559 verified 15 days ago

preview code

Raw History Blame Contribute Delete

3.18 kB

	<svg xmlns="http://www.w3.org/2000/svg" width="920" height="470" viewBox="0 0 920 470" preserveAspectRatio="xMinYMin meet" overflow="hidden">
	<title>JevBench public accuracy for Jev-Omni-MLX-4bit</title>
	<rect width="100%" height="100%" rx="18" fill="#0b1220"/>
	<style>
	text { font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif; fill: #e5e7eb; }
	.muted { fill: #9ca3af; } .grid { stroke: #263449; stroke-width: 1; }
	.axis { stroke: #64748b; stroke-width: 1.2; } .value { font-weight: 700; }
	</style>
	<text x="54" y="45" font-size="24" font-weight="700">JevBench public accuracy</text>
	<text x="54" y="70" font-size="14" class="muted">Ruiruiz30/Jev-Omni-MLX-4bit · 231 decisions · exact argmax</text>
	<line x1="80" y1="375.0" x2="870" y2="375.0" class="grid"/>
	<text x="66" y="380.0" text-anchor="end" font-size="12" class="muted">0%</text>
	<line x1="80" y1="321.0" x2="870" y2="321.0" class="grid"/>
	<text x="66" y="326.0" text-anchor="end" font-size="12" class="muted">20%</text>
	<line x1="80" y1="267.0" x2="870" y2="267.0" class="grid"/>
	<text x="66" y="272.0" text-anchor="end" font-size="12" class="muted">40%</text>
	<line x1="80" y1="213.0" x2="870" y2="213.0" class="grid"/>
	<text x="66" y="218.0" text-anchor="end" font-size="12" class="muted">60%</text>
	<line x1="80" y1="159.0" x2="870" y2="159.0" class="grid"/>
	<text x="66" y="164.0" text-anchor="end" font-size="12" class="muted">80%</text>
	<line x1="80" y1="105.0" x2="870" y2="105.0" class="grid"/>
	<text x="66" y="110.0" text-anchor="end" font-size="12" class="muted">100%</text>
	<line x1="80" y1="105" x2="80" y2="375" class="axis"/>
	<line x1="80" y1="375" x2="870" y2="375" class="axis"/>
	<rect x="122.8" y="105.0" width="112" height="270.0" rx="8" fill="#4ade80" opacity="0.92"/>
	<text x="178.8" y="95.0" text-anchor="middle" font-size="16" class="value">100.00%</text>
	<text x="178.8" y="403" text-anchor="middle" font-size="14">Easy</text>
	<text x="178.8" y="423" text-anchor="middle" font-size="12" class="muted">n=48</text>
	<rect x="320.2" y="108.8" width="112" height="266.2" rx="8" fill="#60a5fa" opacity="0.92"/>
	<text x="376.2" y="98.8" text-anchor="middle" font-size="16" class="value">98.61%</text>
	<text x="376.2" y="403" text-anchor="middle" font-size="14">Original</text>
	<text x="376.2" y="423" text-anchor="middle" font-size="12" class="muted">n=72</text>
	<rect x="517.8" y="170.7" width="112" height="204.3" rx="8" fill="#fbbf24" opacity="0.92"/>
	<text x="573.8" y="160.7" text-anchor="middle" font-size="16" class="value">75.68%</text>
	<text x="573.8" y="403" text-anchor="middle" font-size="14">Hard</text>
	<text x="573.8" y="423" text-anchor="middle" font-size="12" class="muted">n=111</text>
	<rect x="715.2" y="137.7" width="112" height="237.3" rx="8" fill="#c084fc" opacity="0.92"/>
	<text x="771.2" y="127.7" text-anchor="middle" font-size="16" class="value">87.88%</text>
	<text x="771.2" y="403" text-anchor="middle" font-size="14">All</text>
	<text x="771.2" y="423" text-anchor="middle" font-size="12" class="muted">n=231</text>
	<text x="54" y="430" font-size="12" class="muted">Bars are question accuracy; group-macro accuracy is 85.90% across 195 groups.</text>
	</svg>