import torch
from transformers import AutoTokenizer

# Load model (requires trust_remote_code=True)
from src.hf_model import RecurrentAdapterModel

model = RecurrentAdapterModel.from_pretrained(
    "hanseungwook/recurrent-adapter-metamath-cot-r16",
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-8B")

# Generate
prompt = "What is 25 * 37?"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_new_tokens=256,
        num_recurrence_steps=16,
        temperature=0.7,
    )

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Citation

If you use this model, please cite:

bibtex

@misc{recurrent-adapters-2025,
  author = {Han, Seungwook},
  title = {Recurrent Adapters for Mathematical Reasoning},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/hanseungwook/recurrent-adapter-metamath-cot-r16}}
}

License

Apache 2.0