TheFinAI/Fin-o1-14B

🦙 Fino1-8B

Fin-o1-8B is a fine-tuned version of Qwen3-14B, designed to improve performance on [financial reasoning tasks]. This model has been trained using SFT and RF on TheFinAI/Fino1_Reasoning_Path_FinQA, enhancing its capabilities in financial reasoning tasks. Check our paper arxiv.org/abs/2502.08127 for more details.

📌 Model Details

• —Model Name: Fin-o1-14B
• —Base Model: Qwen3-14B
• —Fine-Tuned On: TheFinAI/FinCoT Derived from FinQA, TATQA, DocMath-Eval, Econ-Logic, BizBench-QA, DocFinQA dataset.
• —Training Method: SFT and GRPO
• —Objective: [Enhance performance on specific tasks such as financial mathemtical reasoning]
• —Tokenizer: Inherited from Qwen3-8B

📊 Training Configuration

• —Training Hardware: GPU: [e.g., 8xA100]
• —Batch Size: [e.g., 16]
• —Learning Rate: [e.g., 2e-5]
• —Epochs: [e.g., 3]
• —Optimizer: [e.g., AdamW, LAMB]

🔧 Usage

To use Fin-o1-14B with Hugging Face's transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "TheFinAI/Fin-o1-14B"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

input_text = "What is the results of 3-5?"
inputs = tokenizer(input_text, return_tensors="pt")

output = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(output[0], skip_special_tokens=True))

💡 Citation

If you use this model in your research, please cite:

@article{qian2025fino1,
  title={Fino1: On the Transferability of Reasoning Enhanced LLMs to Finance},
  author={Qian, Lingfei and Zhou, Weipeng and Wang, Yan and Peng, Xueqing and Huang, Jimin and Xie, Qianqian},
  journal={arXiv preprint arXiv:2502.08127},
  year={2025}
}