Instructions to use aixonlab/Aether-12b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use aixonlab/Aether-12b with Transformers:

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

pipe = pipeline("text-generation", model="aixonlab/Aether-12b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("aixonlab/Aether-12b")
model = AutoModelForCausalLM.from_pretrained("aixonlab/Aether-12b")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use aixonlab/Aether-12b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "aixonlab/Aether-12b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "aixonlab/Aether-12b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/aixonlab/Aether-12b

SGLang

How to use aixonlab/Aether-12b 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 "aixonlab/Aether-12b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "aixonlab/Aether-12b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "aixonlab/Aether-12b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "aixonlab/Aether-12b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Unsloth Studio new

How to use aixonlab/Aether-12b with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for aixonlab/Aether-12b to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for aixonlab/Aether-12b to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for aixonlab/Aether-12b to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="aixonlab/Aether-12b",
    max_seq_length=2048,
)

Docker Model Runner
How to use aixonlab/Aether-12b with Docker Model Runner:
```
docker model run hf.co/aixonlab/Aether-12b
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Aether-12b

Aether-12b is a fine-tuned large language model based on Arcanum-12b, further trained on the CleverBoi-Data-20k dataset.

Model Details 📊

Developed by: AIXON Lab
Model type: Causal Language Model
Language(s): English (primarily), may support other languages
License: apache-2.0
Repository: https://huggingface.co/aixonlab/Aether-12b

Model Architecture 🏗️

Base model: Arcanum-12b
Parameter count: ~12 billion
Architecture specifics: Transformer-based language model

Open LLM Leaderboard Evaluation Results

Coming Soon !

Training & Fine-tuning 🔄

Aether-12b was fine-tuned on the following dataset:

Dataset: theprint/CleverBoi-Data-20k
Fine-tuning method: TRL SFTTrainer with AdamW optimizer, cosine decay LR scheduler, bfloat16 precision.

The CleverBoi-Data-20k dataset improved the model in the following ways:

Enhanced reasoning and problem-solving capabilities
Broader knowledge across various topics
Improved performance on specific tasks like writing, analysis, and problem-solving
Better contextual understanding and response generation

Intended Use 🎯

As an assistant or specific role bot.

Ethical Considerations 🤔

As a fine-tuned model based on Arcanum-12b, this model may inherit biases and limitations from its parent model and the fine-tuning dataset. Users should be aware of potential biases in generated content and use the model responsibly.