aether-mini-v1

Aether Mini V1

Aether Mini V1 is a fast, efficient 0.5B parameter language model optimized for general tasks, code generation, and instruction following. Based on Qwen2.5 architecture, fine-tuned for enterprise use cases.

Features

• Fast Inference: Sub-second response times on consumer hardware
• Code Generation: Optimized for Python, JavaScript, TypeScript, and more
• Instruction Following: Fine-tuned for precise instruction adherence
• Low Resource: Runs on CPU with minimal memory requirements

Usage

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("RoeAcquisitions/aether-mini-v1")
tokenizer = AutoTokenizer.from_pretrained("RoeAcquisitions/aether-mini-v1")

messages = [
    {"role": "user", "content": "Write a Python function to calculate fibonacci numbers"}
]

text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer([text], return_tensors="pt")

generated_ids = model.generate(**inputs, max_new_tokens=512)
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)

API

Access via Aether Tech AI API:

curl https://aether-models.nebulahq.work/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "aether-mini-v1",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Pricing

• Input: $0.10 per 1M tokens
• Output: $0.30 per 1M tokens

License

Apache 2.0

Contact

• Website: https://nebulahq.work
• Email: [email protected]