Logical Form Extractor — argument structure extraction.
Built with Qwen.
Base model: unsloth/Qwen2.5-3B-Instruct
Modification: LoRA fine-tune (r=16, alpha=32, 6 epochs), merged into the base
weights. Adapter: models/lora_v7. Weights are modified from the original.
Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT,
Copyright (c) Alibaba Cloud. All Rights Reserved.
NON-COMMERCIAL USE ONLY. The Qwen Research License permits use, reproduction,
distribution and derivative works FOR NON-COMMERCIAL PURPOSES ONLY, defined as
research or evaluation. Full terms:
https://huggingface.co/unsloth/Qwen2.5-3B-Instruct/blob/main/LICENSE
Training data: the LOGIC dataset (Jin et al., Findings of EMNLP 2022), MIT
licensed, plus prose generated for this project.
MIT License, Copyright (c) 2022 Zhijing Jin, Abhinav Lalwani, Tejas Vaidhya,
Xiaoyu Shen, Yiwen Ding, Zhiheng Lyu, Mrinmaya Sachan, Rada Mihalcea,
Bernhard Schoelkopf. Permission is hereby granted, free of charge, to any person
obtaining a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including without
limitation the rights to use, copy, modify, merge, publish, distribute,
sublicense, and/or sell copies of the Software, and to permit persons to whom
the Software is furnished to do so, subject to the following conditions: The
above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS
IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED.
Citation: Jin et al. "Logical Fallacy Detection." Findings of EMNLP 2022,
7180-7198. https://aclanthology.org/2022.findings-emnlp.532