Instructions to use divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16 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("divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16")
config = load_config("divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16")

# 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 divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16"

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": "divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16 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 "divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16"

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 divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16 with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16"

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 "divinetribe/Nemotron-3-Nano-Omni-30B-Abliterated-MM-bf16" \
  --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"

Considerably better than the other abliterated model flying around

by wacomctl672 - opened Aug 24

Discussion

wacomctl672

Aug 24

•

edited Aug 26

Had quantized the model to gguf and used the audio vision mmproj (from ggml repository, I don't know whether this makes a difference but it worked) and it works well at q4km (maybe not as relevant since mlx focused). Thank you for the abliteration! edit: had to make an edit to llama.cpp\conversion\nemotron.py and in it def modify_tensors line 431, which i forgot to mention. i added this to the beginning of the function:

if "switch_mlp.fc1.weight" in name:
            yield f"blk.{bid}.ffn_up_exps.weight", data_torch
            return
elif "switch_mlp.fc2.weight" in name:
            yield f"blk.{bid}.ffn_down_exps.weight", data_torch
            return

divinetribe

Owner Aug 25

thanks finn, glad it's holding up for you. good to know the ggml audio vision mmproj works with the gguf, that's useful for anyone on llama.cpp instead of mlx. i'll add a note about it in the model card.

wacomctl672

Aug 25

uploaded it to save somebody the time if it ever gets needed. will take down if you dont like it. https://huggingface.co/wacomctl672/Nemotron-3-Nano-Omni-30B-Abliterated-MM-GGUF

divinetribe

Owner Aug 26

thanks for doing that, definitely keep it up. i only did the mlx side so a gguf fills a real gap for the llama.cpp folks, and good to know the ggml mmproj works with it. appreciate you sharing it back instead of sitting on it.

thanks
matt

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment