# Multi-task Instruction Fine-tuning

This is a language model merged using the mergekit tool through the SCE method, combining multiple 7B-parameter scale models

This is an 8B-parameter model based on the Llama-3.1 architecture, supporting instruction understanding and text generation tasks in Uzbek and English.

Based on the Qwen2.5-7B-Instruct foundation model, it integrates the advantages of multiple 7B-scale models through model ensemble techniques.

Multilingual E5 large instruction model supporting text embedding and classification tasks for over 100 languages

Dolphin3.0-R1-Mistral-24B is a 24B-parameter large language model based on the Mistral architecture, trained by Eric Hartford, focusing on reasoning and first-principles analysis.

KafkaLM 8b is a German large language model fine-tuned from Llama3 8b, specializing in German business scenarios

Meta's Llama 3 series of large language models, including 8B and 70B scale pre-trained and instruction-tuned generative text models, optimized for dialogue scenarios, with excellent performance in industry benchmark tests.

This is a 34B-parameter large language model merged using the DARE Ties method via mergekit, based on the Yi-34B-200K architecture, integrating the capabilities of three models: Nous-Capybara-34B, Tess-M-v1.3, and airoboros-3_1-yi-34b-200k.

AgentLM-7B is an agent-enhanced language model obtained by mixed training on the AgentInstruct dataset and ShareGPT dataset based on the Llama-2-chat model.

A 6.9 billion parameter dialogue-specific language model developed by Together in collaboration with multiple AI research institutions, trained on the RedPajama-Data-1T dataset and enhanced with dialogue capabilities through fine-tuning with OASST1 and Dolly2 data

An encoder-decoder model based on the T5 architecture, optimized through Flan prompt tuning, supporting multilingual task processing

FLAN-T5 is an instruction-fine-tuned language model based on T5, achieving superior performance through fine-tuning on over 1,000 multilingual tasks with the same parameter count

FLAN-T5 is an instruction-fine-tuned language model based on T5, supporting 60+ languages, achieving stronger performance through fine-tuning on 1000+ tasks with the same parameter count