Instructions to use deepsweet/Ornith-1.0-35B-MLX-VL-oQ5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use deepsweet/Ornith-1.0-35B-MLX-VL-oQ5 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("deepsweet/Ornith-1.0-35B-MLX-VL-oQ5")

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 deepsweet/Ornith-1.0-35B-MLX-VL-oQ5 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5"

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": "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use deepsweet/Ornith-1.0-35B-MLX-VL-oQ5 with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use deepsweet/Ornith-1.0-35B-MLX-VL-oQ5 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 "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5"

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 deepsweet/Ornith-1.0-35B-MLX-VL-oQ5

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use deepsweet/Ornith-1.0-35B-MLX-VL-oQ5 with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5"

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 "deepsweet/Ornith-1.0-35B-MLX-VL-oQ5" \
  --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"

Ornith-1.0-35B-MLX-VL-oQ5

File size: 1,113 Bytes

01b5998

---
language: en
license: mit
pipeline_tag: text-generation
tags:
- mlx
library_name: mlx
base_model: deepreinforce-ai/Ornith-1.0-35B
---

This model was converted to MLX format and quantized from [Ornith-1.0-35B](https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B) using [oMLX](https://github.com/jundot/omlx).

>[!CAUTION]
>The conversion required to stack expert MLP weights into fused per-layer tensors. Treat this as an experiment and act accordingly.

## What is "oQ"?

See ["oQ: oMLX Universal Dynamic Quantization"](https://github.com/jundot/omlx/blob/main/docs/oQ_Quantization.md) for details.

## What is "VL"?

"VL" is Vision-Language, meaning quantization preserves the original model's multimodality.

No "VL" means quantization is Text-Only.

## What is "FP16"?

"FP16" is an M1/M2 Apple Silicon tweak that delivers a very noticeable prompt processing boost, because older M-series lack native BF16 hardware support. See ["Metal FP32 Vs BF16 Vs FP16 benchmark"](https://github.com/deepsweet/metal-fp32-bf16-fp16) for details.

No "FP16" means quantization is better suited for M3+ Apple Silicon.