Update README.md

README.md

@@ -1,3 +1,30 @@
----
-license: apache-2.0
----
+---
+license: apache-2.0
+datasets:
+- databricks/databricks-dolly-15k
+language:
+- en
+metrics:
+- rouge
+base_model:
+- openai-community/gpt2-xl
+pipeline_tag: text-generation
+---
+# teacher-gpt2-1.5B
+
+[paper](https://arxiv.org/abs/2306.08543) | [code](https://github.com/microsoft/LMOps/tree/main/minillm)
+
+**teacher-gpt2-1.5B** is a gpt2-xlarge (1.5B) model supervised fine-tuned on [databricks-dolly-15k](https://huggingface.co/datasets/aisquared/databricks-dolly-15k).
+
+It is used as the teacher model for MiniLLM series.
+
+
+## Citation
+```
+@inproceedings{minillm,
+  title={MiniLLM: Knowledge Distillation of Large Language Models},
+  author={Gu, Yuxian and Dong, Li and Wei, Furu and Huang, Minlie},
+  booktitle={Proceedings of ICLR},
+  year={2024}
+}
+```