Instructions to use cudo528/Darwin-36B-Opus-mlx-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use cudo528/Darwin-36B-Opus-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("cudo528/Darwin-36B-Opus-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
- Pi
How to use cudo528/Darwin-36B-Opus-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 "cudo528/Darwin-36B-Opus-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": "cudo528/Darwin-36B-Opus-mlx-4bit" } ] } } }Run Pi
# Start Pi in your project directory: pi
- MLX LM
How to use cudo528/Darwin-36B-Opus-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 "cudo528/Darwin-36B-Opus-mlx-4bit"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "cudo528/Darwin-36B-Opus-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": "cudo528/Darwin-36B-Opus-mlx-4bit", "messages": [ {"role": "user", "content": "Hello"} ] }' - Hermes Agent
How to use cudo528/Darwin-36B-Opus-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 "cudo528/Darwin-36B-Opus-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 cudo528/Darwin-36B-Opus-mlx-4bit
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use cudo528/Darwin-36B-Opus-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 "cudo528/Darwin-36B-Opus-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 "cudo528/Darwin-36B-Opus-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"
Darwin-36B-Opus (MLX Optimized)
This repository provides an MLX-optimized version of FINAL-Bench/Darwin-36B-Opus, a state-of-the-art Mixture-of-Experts (MoE) model engineered for advanced reasoning, chain-of-thought (CoT) processes, and complex problem-solving.
Apple Silicon(Mac) ํ๊ฒฝ์ MLX ํ๋ ์์ํฌ์์ ์ต์ ์ ์๋์ ํจ์จ์ ์ธ ๋ฉ๋ชจ๋ฆฌ ๋์ญํญ์ผ๋ก ๊ตฌ๋๋๋๋ก ํฌ๋งท ๋ณํ ๋ฐ ์ต์ ํ๊ฐ ์๋ฃ๋ ๋ฒ์ ์ ๋๋ค.
๐ Model Highlights
- Evolutionary Merge & MoE: Qwen ๊ธฐ๋ฐ ์ํคํ ์ฒ๋ฅผ ์งํ์ ๋จธ์ง(Evolutionary Merge) ๊ธฐ๋ฒ๊ณผ Mixture-of-Experts(MoE) ๊ตฌ์กฐ๋ก ๊ฒฐํฉํ์ฌ ์ฒด๊ธ์ ๋ฐ์ด๋๋ ์ต๊ณ ์ ํจ์จ์ฑ์ ์๋ํฉ๋๋ค.
- Advanced Reasoning: ๊น์ ๋ค๋จ๊ณ ์ฌ๊ณ ๊ณผ์ (
<thought>)์ ๊ฑฐ์ณ ๋ณต์กํ ์ฝ๋ฉ, ์ํ, ๊ณผํ์ ๋์ ๋ฅผ ํด๊ฒฐํ๋ ์ฌ์ธต ์ถ๋ก (Chain-of-Thought) ๋ฅ๋ ฅ์ด ํ์ฌ๋์ด ์์ต๋๋ค. - Top-Tier Benchmark Performance: ๋ฐ์ฌ๊ธ ๋์ด๋์ ํ๋์ฝ์ด ์ถ๋ก ํ ์คํธ์ธ GPQA Diamond ๋ฒค์น๋งํฌ์์ 88.4%์ ๊ณ ๋์ ์ ๊ธฐ๋กํ์ต๋๋ค.
- Multilingual Support: ํ๊ตญ์ด, ์์ด, ์ค๊ตญ์ด, ์ผ๋ณธ์ด๋ฅผ ํฌํจํ 10๊ฐ ์ด์์ ๋ค๊ตญ์ด ํ๊ฒฝ์ ์๋ฒฝํ๊ฒ ์ง์ํฉ๋๋ค.
๐ ๏ธ Requirements & Installation
์ด ๋ชจ๋ธ์ ๊ตฌ๋ํ๋ ค๋ฉด ์ต์ ๋ฒ์ ์ mlx-lm ๋ผ์ด๋ธ๋ฌ๋ฆฌ๊ฐ ํ์ํฉ๋๋ค.
pip install -U mlx-lm
๐ป How to Use
1. Python API๋ฅผ ์ด์ฉํ ์คํ
from mlx_lm import load, generate
model, tokenizer = load("cudo528/Darwin-36B-Opus-MLX-4bit")
prompt = "Quantum computing๊ณผ ๊ธฐ์กด ์ปดํจํฐ์ ๊ทผ๋ณธ์ ์ธ ์ถ๋ก ๋ฐฉ์ ์ฐจ์ด๋ฅผ ํ๊ตญ์ด๋ก ๋ช
ํํ๊ฒ ์ค๋ช
ํด์ค."
response = generate(
model,
tokenizer,
prompt=prompt,
max_tokens=2048,
verbose=True # ์ฌ์ธต ์ถ๋ก ๊ณผ์ (<thought>)์ ํ์ธํ๋ ค๋ฉด True๋ก ์ค์
)
print(response)
2. ํฐ๋ฏธ๋(CLI) ํ๊ฒฝ์์ ์ฆ์ ์คํ
mlx_lm.generate --model cudo528/Darwin-36B-Opus-MLX-4bit --prompt "Explain the concept of Hybrid Vigor in AI." --max-tokens 1024
๐ Quantization & Lineage
- Original Base Model: FINAL-Bench/Darwin-36B-Opus
- Format: MLX Formatted (
.safetensors) - License: Apache-2.0
๋ณธ ๋ณํ ๋ฒ์ ์ ์๋ณธ ๋ชจ๋ธ์ ๊ฐ์ค์น ์ฑ์ง๊ณผ ๋ผ์ด์ ์ค๋ฅผ ์ ์ ์ผ๋ก ์ค์ํฉ๋๋ค. ๋ก์ปฌ ํ๋์จ์ด ์ฌ์์ ๋ง์ถฐ ์์ํ ๋นํธ ์(์: 8-bit, 4-bit)๋ฅผ ์กฐ์ ํ ๊ฒฝ์ฐ, ํ์ผ ๋ชฉ๋ก์ ์์ ์ ํ์ธํด ์ฃผ์๊ธฐ ๋ฐ๋๋๋ค.
- Downloads last month
- 29
4-bit
Model tree for cudo528/Darwin-36B-Opus-mlx-4bit
Base model
FINAL-Bench/Darwin-36B-OpusEvaluation results
- Accuracy on GPQA Diamondself-reported88.400