metadata
license: apache-2.0
base_model: Qwen/Qwen2.5-0.5B-Instruct
tags:
- moe
- custom-architecture
- nanosota
- reasoning
NanoSOTA-v2-Logic-Specialist
Description: Logic Specialist (Parallel vs Sequential).
How to Load (Required)
This model uses a custom 8-Expert MoE architecture. You must use the provided loader script.
from nanosota_moe import load_nanosota
model, tokenizer = load_nanosota("oscarz511/NanoSOTA-v2-Logic-Specialist")
prompt = "If it takes 3 hours to dry 3 shirts, how long for 30 shirts?"
inputs = tokenizer.apply_chat_template([
{"role": "system", "content": "You are NanoSOTA. Think step-by-step."},
{"role": "user", "content": prompt}
], return_tensors="pt", add_generation_prompt=True).to("cuda")
out = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(out[0]))