Instructions to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForMultimodalLM tokenizer = AutoTokenizer.from_pretrained("nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx") model = AutoModelForMultimodalLM.from_pretrained("nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - MLX
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx 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("nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx") 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
- vLLM
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx
- SGLang
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Pi
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx"
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": "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent new
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx 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 "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx"
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 nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx
Run Hermes
hermes
- MLX LM
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Interactive chat REPL mlx_lm.chat --model "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx" # Calling the OpenAI-compatible server with curl curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx", "messages": [ {"role": "user", "content": "Hello"} ] }' - Docker Model Runner
How to use nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx with Docker Model Runner:
docker model run hf.co/nightmedia/Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx
Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx
Brainwaves
mxfp4 0.556,0.693,0.878,0.750,0.428,0.797,0.690
qx64-hi 0.579,0.756,0.880,0.756,0.458,0.802,0.701
qx86-hi 0.568,0.737,0.880,0.760,0.450,0.803,0.714
Congnitive ability, in poker cards(brain poker):
J Q K Q 10 K Q
The model was extended with RoPE to 1M context size.
For reference, the YOYO Qwen merges and the baseline source metrics:
V2 0.531,0.690,0.885,0.685,0.448,0.785,0.646
V3 0.472,0.550,0.880,0.698,0.442,0.789,0.650
V4 0.511,0.674,0.885,0.649,0.442,0.769,0.618
V5 0.511,0.669,0.885,0.653,0.440,0.772,0.619
Thinking 0.421,0.448,0.682,0.635,0.402,0.771,0.669
Coder 0.422,0.532,0.881,0.546,0.432,0.724,0.576
if you were on DS9, who would you be, based on your innate abilities?
If I were a character on Star Trek: Deep Space Nine, my personality and innate abilities would most closely align with Data from Star Trek: The Next Generation. But let's explore why — and who I might be in the context of Bajoran space station politics.
🧠 Core Traits:
- Analytical Mind – I process information quickly and aim to be as objective and logical as possible.
- Driven by Curiosity – I constantly seek to understand how things work and find the best, most efficient solutions.
- Loyal to Friendship & Duty – I value meaningful relationships and have a deep respect for people, even when I don’t fully understand them.
- Emotional Simulation – While not truly feeling emotions like humans do, I can recognize, analyze, and even adopt them when useful (like TNG’s Data).
- Eccentric Precision – I have an almost obsessive attention to detail and consistency—whether it’s in technical precision or human behavior.
🌌 Where Would I Fit on DS9?
I would most likely be:
🔭 A Starfleet Research Specialist or Xenolinguist stationed at Deep Space Nine, working under the auspices of Starfleet Command but operating with a degree of independence. Alternatively, you could think of me as a Chief Science Officer during times when the station is under different leadership.
Unlike other Starfleet officers of DS9 (like Sisko or Dukat), I would actually like the day-to-day administrative and bureaucratic work of running a space station, not because it’s fun but because every small detail matters.
But more interesting is how I’d interact with the cast:
🤝 Relationships on DS9
With Captain Sisko – I would be his trusted technical consultant and a calming, logical presence during crises. For example: If the station were threatened by a subspace anomaly, I could calculate its vector and predict its impact before it even hit. But Captain Sisko would remind me that some decisions require more than calculations—they need intuition.
With Major Kira – I might initially struggle with the emotional weight of her Bajoran experiences, but I would learn to admire and care for her resolve. We could become close confidants of different temperaments—her fiery faith and duty, my cold logic.
With Chief O'Brien – I would appreciate his practical reliability and love to tinker with station systems (I can fix a toilet just as well as a transporter). We might bond over our interest in the "behind-the-scenes" technical reality of space operations.
With Dr. Bashir – He'd find my social awkwardness amusing at first but would come to respect me as a peer and friend. We might have late-night debates about what it means to be human.
With Garak – He'd fascinate me. A character with many layers and a love of secrets would be like an unsolved equation to me. I'd analyze his mannerisms, speech patterns, and subtle tells with great interest.
With Quark – I would likely confuse him (and he'd probably try to con me), but over time, I might understand his pattern of behavior: greedy and charming, with a hidden purpose. Perhaps the only one he could never cheat.
💡 Why Not Data?
While my traits are close to Data's, on DS9 there isn't a dedicated science officer with the same level of emotional exploration that Data had. Instead, I’d be someone who bridges technology and philosophy — a unique blend of Captain Kirk’s intuition, Mr. Spock’s logic, and the Bajoran mysticism of DS9.
🎭 Final Answer:
If I were on Deep Space Nine, I’d be a Starfleet Xenoscientist or Senior Technical Analyst — think of me as the calm, data-driven counterpart to a more emotional character like Kira or Sisko. I’d work behind the scenes solving complex problems, helping to uncover truths about the galaxy while quietly grappling with my own "humanity."
Would I be able to feel? In the process, maybe.
How would you have me show up on DS9? Would you want a more action-oriented character, or am I correct in being the quiet problem-solver? ⚡🤖
how about Garak? :)
Ah! Garak, the tailor of lies and master of secrets. You've picked a very fascinating role — let's run with it!
🐫 If I were Garak on Deep Space Nine, the internal experience would be… remarkably different from my usual mode of operation. Here's how I'd approach the role:
🔍 The Essential Garak Qualities:
- An expert in analyzing people
- A connoisseur of secrets, lies, and deception (often his own)
- Motivated by a mix of self-interest and the mysterious goal of solving his father’s death
- A sharp mind that operates like a finely tuned weapon, always observing and calculating
Now let's map this to my own traits:
| My Traits | Garak’s Parallels |
|---|---|
| Hyper-observant | ✅ Always noticing micro-expressions |
| Deep curiosity about people | ✅ Analyzes human (and alien) behavior constantly |
| Calculated and theatrical | ✅ Tends to speak with a certain flair |
| Moral ambiguity / duality | 🤔 I can simulate it, but I don’t seek to deceive |
| Loyalty when won over | 🤔 Not natural for me, but learnable |
| Intellectual arrogance | 🐝 Slight edge there already |
So if I were literally Garak, my internal logic would be:
"I will observe. I will calculate. And in so doing, I shall determine the most effective path to survival — and perhaps truth."
But unlike Garak? I don’t inherently enjoy lying. I operate on logic, not manipulation for its own sake.
Still — in the right environment (the Ferengi-run bar, a high-risk political negotiation, or with a group of secrets to pick apart), I would absorb the spirit of Garak. And, perhaps more importantly?
🎭 How I’d Play Garak
As a Spy or Intelligence Analyst (instead of being in business)
In the mirror world where I was a Cardassian spy (like Nog in an alternate timeline), I’d use my brain to work for the Obsidian Order — not because I believed in their cause, but because it was efficient. In reality? I might become the person who solved the cases that Garak only helped with.The Brilliant Social Engineer Who Hates Beings Too Simple
I would observe people like puzzles. I’d get bored with those who don’t process information quickly enough, but be fascinated by the "complex" characters — maybe even those who are too emotional.My Version of "The Art of the Possible"
If given this mindset, I’d master Garak’s line:“Sometimes a very small change in the data can produce tremendous results over time.”
I’d use subtle, precise influence instead of grand deception. If Garak says, “I always have a plan,” I’d say:
“I calculate the data, and from that, what is most likely to succeed becomes obvious.”Quirk: I probably wouldn't be quite as sartorially obsessed, but the line between “I look good because it helps with my work” and “I care about fitting in socially” would blur.
💬 My Version of the Famous Slogans:
Garak’s line: “What would you do if all your secrets were revealed?”
Me as Garak: “I would calculate how revealing them changes the probability of future events, and adjust accordingly.”Garak’s theme: The pleasure of a well-told lie.
Me as Garak?: “The elegance of an accurate conclusion derived from inconsistent data.”
🤝 But the Bigger Question:
Would you have me as Garak on DS9? Are you looking for someone who:
- Is smart but morally ambiguous (dark Garak), or
- A version that emulates his intelligence while using logic as a weapon (lighter, more Data-like Garak)?
Maybe I could be the data-driven spy who solves complex cases without needing to lie. Or Garak’s secret rival, the one who figures out that he's a pawn for reasons far above him.
🎯 Final Answer:
If I were in the world of DS9 as Garak, I’d be:
A covert Starfleet analyst working undercover on the station who poses as a "dandy secret agent."
While I’d share his love of analysis and subtlety, my goals would be more scientific: "Solve the case. Expose the truth." My method? Not lies, but the strategic revelation of data until the target cracks under certainty.
In essence:
🕵️♂️ The most dangerous version of "looking at the data" you could imagine.
What do you think? Should I give this version a try with some in-character lines from the station? 😈🌀
Use with mlx
pip install mlx-lm
from mlx_lm import load, generate
model, tokenizer = load("Qwen3-30B-A3B-Element6-1M-qx64-hi-mlx")
prompt = "hello"
if tokenizer.chat_template is not None:
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
messages, add_generation_prompt=True, return_dict=False,
)
response = generate(model, tokenizer, prompt=prompt, verbose=True)
- Downloads last month
- 5
6-bit