--- language: - ru tags: - generation - question answering - instruction tuning license: cc-by-nc-4.0 --- ### Model Description This HF repository contains base LLMs instruction tuned (SFT) with full-parameter fine-tuning and then used to study whether monolingual or multilingual instruction tuning is more favourable. * [GitHub](https://github.com/hplt-project/monolingual-multilingual-instruction-tuning/tree/main) * [Paper](https://arxiv.org/abs/2309.08958) #### Instruction tuning details * Base model: [bloom-7b1](https://huggingface.co/bloom-7b1) * Instruction tuning language: Russian * Training method: full-parameter fine-tuning. * Best checkpoint: best cross-entropy on a validation set, trained for 3 epochs. * Dataset: machine-translated from [yahma/alpaca-cleaned](https://huggingface.co/datasets/yahma/alpaca-cleaned). You can download our data [HERE](https://github.com/hplt-project/monolingual-multilingual-instruction-tuning/tree/main/training-data). #### Usage The model checkpoint should be loaded using `transformers` library. Please refer to our Github repository [HERE](https://github.com/hplt-project/monolingual-multilingual-instruction-tuning/tree/main/fpft) for inference and training instructions. #### Citation ``` @inproceedings{chen-etal-2024-monolingual, title="Monolingual or multilingual instruction tuning: Which makes a better {Alpaca}", author="Pinzhen Chen and Shaoxiong Ji and Nikolay Bogoychev and Andrey Kutuzov and Barry Haddow and Kenneth Heafield", year="2024", booktitle = "Findings of the Association for Computational Linguistics: EACL 2024", } ```