# Multilingual Instruction Fine-tuning

IBM-Granite's 2-billion parameter instruction model supporting multilingual and long-context tasks with structured reasoning capabilities.

Mistral-Small-3.1-24B-Instruct-2503 is a 24B-parameter multilingual instruction fine-tuned model that supports multiple languages and tasks, suitable for text generation and dialogue scenarios.

Granite-3.0-8B-Instruct is an 8-billion-parameter model fine-tuned from Granite-3.0-8B-Base, trained using a combination of various open-source instruction datasets and internally synthesized datasets.

Gemma 2 JPN is a Japanese-text fine-tuned version of the Gemma 2 2B model, excelling in Japanese language processing and suitable for various text generation tasks.

LLaMAntino-3-ANITA is a multilingual (English + Italian) large language model based on Meta Llama 3, specifically optimized for Italian NLP tasks.

Mixtral-8x22B-Instruct-v0.1 is a large language model fine-tuned for instructions based on Mixtral-8x22B-v0.1, supporting multiple languages and function calling capabilities.

Calme-7B is a 7-billion-parameter language model fine-tuned based on Mistral-7B, excelling in generating clear, calm, and coherent text.

A multilingual instruction-following model fine-tuned based on Mistral-7B-Instruct-v0.2, supporting 101 languages

Aya 101 is a large-scale multilingual generative language model supporting instructions in 101 languages, outperforming similar models in various evaluations.

Mixtral-8x7B is a pre-trained generative sparse mixture of experts model that outperforms Llama 2 70B on most benchmarks.

FLAN-T5 XL is an instruction-finetuned language model based on the T5 architecture, with significantly improved multilingual and few-shot performance after fine-tuning on 1000+ tasks.

FLAN-T5 is a language model optimized through instruction fine-tuning based on the T5 model, supporting multilingual task processing and outperforming the original T5 model with the same parameter count.