Add model card for LatentChem

README.md

@@ -0,0 +1,41 @@
+---
+license: mit
+library_name: transformers
+pipeline_tag: text-generation
+tags:
+- chemistry
+- reasoning
+- latent-thinking
+---
+
+# LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoning
+
+LatentChem is a latent reasoning interface that decouples chemical computation from textual generation. It enables models to perform multi-step reasoning directly in continuous latent space, overcoming the representation mismatch between discrete language tokens and continuous chemical logic.
+
+## Summary
+Chemical large language models traditionally rely on explicit Chain-of-Thought (CoT) in natural language. LatentChem introduces a shift toward implicit latent computation. When optimized for task success, the model progressively internalizes reasoning, abandoning verbose textual derivations in favor of more efficient continuous latent dynamics. This approach delivers significant performance gains and a substantial inference speedup (averaging 10.84x) across chemical reasoning benchmarks.
+
+## Resources
+- **Paper:** [LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoning](https://huggingface.co/papers/2602.07075)
+- **GitHub Repository:** [https://github.com/xinwuye/LatentChem](https://github.com/xinwuye/LatentChem)
+
+## Installation
+
+To set up the environment for LatentChem, follow the instructions from the official repository:
+
+```bash
+git clone https://github.com/xinwuye/LatentChem.git
+cd LatentChem
+conda env create -f env.yml
+conda activate latentchem_dev
+```
+
+## Citation
+```bibtex
+@article{ye2026latentchem,
+  title={LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoning},
+  author={Ye, Xinwu and Mao, Yicheng and Zhang, Jia and Liu, Yimeng Houses and Hao, Li and Wu, Fang and Li, Zhiwei and Liao, Yuxuan and Wang, Zehong and Liu, Zhiyuan and others},
+  journal={arXiv preprint arXiv:2602.07075},
+  year={2026}
+}
+```