Instructions to use Neura-Tech-AI/Neuron-46x4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Neura-Tech-AI/Neuron-46x4B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Neura-Tech-AI/Neuron-46x4B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Neura-Tech-AI/Neuron-46x4B")
model = AutoModelForCausalLM.from_pretrained("Neura-Tech-AI/Neuron-46x4B", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Neura-Tech-AI/Neuron-46x4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Neura-Tech-AI/Neuron-46x4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Neura-Tech-AI/Neuron-46x4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Neura-Tech-AI/Neuron-46x4B

SGLang

How to use Neura-Tech-AI/Neuron-46x4B 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 "Neura-Tech-AI/Neuron-46x4B" \
    --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": "Neura-Tech-AI/Neuron-46x4B",
		"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 "Neura-Tech-AI/Neuron-46x4B" \
        --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": "Neura-Tech-AI/Neuron-46x4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Neura-Tech-AI/Neuron-46x4B with Docker Model Runner:
```
docker model run hf.co/Neura-Tech-AI/Neuron-46x4B
```

Neura-Tech-AI commited on 16 days ago

Commit

573a930

verified ·

1 Parent(s): dabe26e

Upload mergekit_moe_config.yml with huggingface_hub

Browse files

Files changed (1) hide show

mergekit_moe_config.yml +557 -0

mergekit_moe_config.yml ADDED Viewed

	@@ -0,0 +1,557 @@

+base_model: Neura-Tech-AI/Neuron-4B-Instruct
+gate_mode: hidden
+dtype: bfloat16
+num_local_experts: 6
+experts:
+  - source_model: Neura-Tech-AI/Neuron-4B-Instruct
+    positive_prompts:
+      - chat
+      - conversation
+      - general
+      - assistant
+      - help
+      - question
+      - answer
+      - support
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - code
+      - programming
+      - python
+      - javascript
+      - java
+      - cpp
+      - rust
+      - go
+      - sql
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - math
+      - calculus
+      - algebra
+      - geometry
+      - trigonometry
+      - linear algebra
+      - differential
+      - integral
+      - equation
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - logic
+      - reasoning
+      - deduction
+      - induction
+      - inference
+      - puzzle
+      - riddle
+      - brain teaser
+      - critical thinking
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - physics
+      - mechanics
+      - thermodynamics
+      - electromagnetism
+      - quantum
+      - relativity
+      - optics
+      - acoustics
+      - fluid dynamics
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - chemistry
+      - organic
+      - inorganic
+      - biochemistry
+      - molecular
+      - atomic
+      - periodic
+      - reaction
+      - synthesis
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - biology
+      - genetics
+      - evolution
+      - ecology
+      - microbiology
+      - botany
+      - zoology
+      - anatomy
+      - physiology
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - algorithm
+      - data structure
+      - complexity
+      - optimization
+      - recursion
+      - sorting
+      - searching
+      - graph
+      - tree
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - statistics
+      - probability
+      - regression
+      - correlation
+      - distribution
+      - variance
+      - hypothesis
+      - confidence
+      - bayesian
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - machine learning
+      - deep learning
+      - neural network
+      - transformer
+      - attention
+      - gradient
+      - backpropagation
+      - fine-tune
+      - embedding
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - ai
+      - llm
+      - nlp
+      - tokenizer
+      - prompt
+      - inference
+      - training
+      - evaluation
+      - benchmark
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - data science
+      - analytics
+      - visualization
+      - pandas
+      - numpy
+      - scikit
+      - matplotlib
+      - seaborn
+      - tableau
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - cybersecurity
+      - encryption
+      - authentication
+      - firewall
+      - malware
+      - ransomware
+      - zero trust
+      - vpn
+      - cloud
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - networking
+      - tcp/ip
+      - routing
+      - dns
+      - http
+      - ssh
+      - docker
+      - kubernetes
+      - devops
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - economics
+      - finance
+      - accounting
+      - investment
+      - stock
+      - portfolio
+      - dividend
+      - hedging
+      - arbitrage
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - business
+      - marketing
+      - management
+      - strategy
+      - growth
+      - customer
+      - revenue
+      - roi
+      - sales
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - research
+      - methodology
+      - literature review
+      - peer review
+      - citation
+      - publication
+      - academic
+      - thesis
+      - grant
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - engineering
+      - mechanical
+      - electrical
+      - civil
+      - robotics
+      - control
+      - signal
+      - embedded
+      - aerospace
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - medicine
+      - clinical
+      - diagnosis
+      - treatment
+      - surgery
+      - radiology
+      - cardiology
+      - neurology
+      - oncology
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - astronomy
+      - cosmology
+      - astrophysics
+      - galaxy
+      - black hole
+      - planet
+      - orbit
+      - telescope
+      - nebula
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - environment
+      - sustainability
+      - climate
+      - ecosystem
+      - biodiversity
+      - conservation
+      - renewable
+      - solar
+      - wind
+      - hydro
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - psychology
+      - cognition
+      - behavior
+      - emotion
+      - memory
+      - learning
+      - mental
+      - therapy
+      - personality
+  - source_model: Qwen/Qwen3-4B-Thinking-2507
+    positive_prompts:
+      - philosophy
+      - ethics
+      - metaphysics
+      - epistemology
+      - ontology
+      - virtue
+      - morality
+      - existence
+      - consciousness
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - creative
+      - write
+      - story
+      - narrative
+      - fiction
+      - poetry
+      - essay
+      - dialogue
+      - character
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - translate
+      - language
+      - english
+      - french
+      - spanish
+      - german
+      - chinese
+      - japanese
+      - arabic
+      - hindi
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - summarize
+      - paraphrase
+      - condense
+      - digest
+      - shorten
+      - main idea
+      - key points
+      - outline
+      - abstract
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - explain
+      - tutorial
+      - guide
+      - walkthrough
+      - step-by-step
+      - beginner
+      - advanced
+      - how to
+      - instruction
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - education
+      - teach
+      - learn
+      - lecture
+      - lesson
+      - homework
+      - quiz
+      - exam
+      - grade
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - email
+      - letter
+      - memo
+      - professional
+      - formal
+      - informal
+      - greeting
+      - signature
+      - subject
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - article
+      - blog
+      - post
+      - seo
+      - headline
+      - introduction
+      - conclusion
+      - call to action
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - documentation
+      - manual
+      - api
+      - reference
+      - quick start
+      - installation
+      - configuration
+      - usage
+      - faq
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - report
+      - analysis
+      - finding
+      - recommendation
+      - executive summary
+      - chart
+      - graph
+      - table
+      - appendix
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - grammar
+      - punctuation
+      - spelling
+      - syntax
+      - tense
+      - preposition
+      - conjunction
+      - clause
+      - sentence
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - proofreading
+      - editing
+      - revise
+      - polish
+      - refine
+      - copyedit
+      - style
+      - clarity
+      - conciseness
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - history
+      - civilization
+      - revolution
+      - war
+      - renaissance
+      - enlightenment
+      - industrial
+      - colonial
+      - cold war
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - geography
+      - map
+      - country
+      - capital
+      - city
+      - region
+      - continent
+      - ocean
+      - mountain
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - culture
+      - tradition
+      - customs
+      - festival
+      - cuisine
+      - music
+      - art
+      - dance
+      - fashion
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - politics
+      - government
+      - democracy
+      - election
+      - policy
+      - parliament
+      - constitution
+      - rights
+      - law
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - travel
+      - tourism
+      - destination
+      - itinerary
+      - accommodation
+      - sightseeing
+      - cuisine
+      - adventure
+      - guide
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - architecture
+      - design
+      - planning
+      - blueprint
+      - structure
+      - urban
+      - interior
+      - landscape
+      - sustainable
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - gaming
+      - minecraft
+      - rpg
+      - strategy
+      - multiplayer
+      - modding
+      - server
+      - community
+      - moderation
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - roleplay
+      - scenario
+      - character
+      - dialogue
+      - immersive
+      - worldbuilding
+      - lore
+      - backstory
+      - action
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - brainstorm
+      - idea
+      - concept
+      - innovation
+      - prototype
+      - solution
+      - problem
+      - challenge
+      - opportunity
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - productivity
+      - workflow
+      - automation
+      - project
+      - milestone
+      - deadline
+      - team
+      - collaboration
+      - tool
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - personal
+      - life
+      - health
+      - fitness
+      - nutrition
+      - wellness
+      - relationship
+      - career
+      - success
+  - source_model: Qwen/Qwen3-4B-Instruct-2507
+    positive_prompts:
+      - motivation
+      - inspiration
+      - mindset
+      - confidence
+      - positivity
+      - meditation
+      - mindfulness
+      - gratitude
+      - reflection