# Low-resource fine-tuning

A code reasoning model fine-tuned based on Qwen2.5-Coder-1.5B-Instruct, which enhances programming problem-solving and logical reasoning abilities

This model is a fine-tuned version of google/gemma-2-9b-it for function calling tasks, with training data entirely manually annotated using the Russian version of the DiTy/function-calling dataset.

A fine-tuned model specializing in translating Taiwan-style Traditional Chinese to English

An instruction-following model fine-tuned based on Mistral-7B-V0.3, optimized for role-playing and story creation

BART-large model fine-tuned on the XSum dataset for generating English abstractive summaries

An English-Arabic translation model fine-tuned based on Helsinki-NLP/opus-mt-en-ar, achieving a BLEU score of 29.991 on the evaluation set

RigoChat-7b-v2 is a Spanish-optimized model based on Qwen-2.5, with performance enhanced through DPO fine-tuning, focusing on accurate responses to Spanish queries.

An English-Arabic translation model fine-tuned from Helsinki-NLP/opus-mt-en-ar, specifically optimized for Egyptian dialect

A Turkish text-to-speech model fine-tuned based on Microsoft's SpeechT5 TTS, focusing on high-quality Turkish speech synthesis

A Bengali large language model continuously pre-trained based on the Llama 3.2 architecture, fine-tuned on a large Bengali dataset to improve the ability of Bengali text generation and understanding.

A Japanese-optimized large language model fine-tuned based on Meta-Llama-3-8B-Instruct, excelling in multiple Japanese benchmark tests

An enhanced version based on Meta-Llama-3-8B-Instruct, fine-tuned and optimized through the Cabra 30k dataset, with special optimization for Portuguese understanding and response.

This is a mental health counseling dialogue model fine-tuned based on Mistral-7B-Instruct-v0.2, specifically designed for generating mental health-related dialogue responses.

A large language model fine-tuned with QLoRA on the Open-Platypus-Chat dataset based on Jamba-v0.1, supporting conversational tasks

A Portuguese-optimized model fine-tuned from Qwen1.5-7B-Chat, excelling in multiple Brazilian benchmark tests

Urdu speech synthesis model fine-tuned on the fleurs dataset based on microsoft/speecht5_tts

A text summarization model fine-tuned on Norwegian news summarization datasets using RLHF strategy based on NorGPT-3B model

Einstein-v4-7B is a large language model fully fine-tuned on diverse scientific datasets based on Mistral-7B-v0.1, specializing in STEM field tasks

A summarization model fine-tuned from google/mt5-small on the CNN/DailyMail dataset, suitable for English text summarization tasks.

This is a fine-tuned adapter for the Meta Llama 2 7B model, specifically designed for translating Czech text into English.

A Chinese dialogue model fine-tuned based on Qwen-7B, integrating MOSS dataset and school math problem data

MERAK-7B-V2 is an Indonesian large language model fine-tuned based on Meta Llama-2-7B-Chat-HF, supporting Indonesian and English.

LINCE-ZERO is a 7-billion-parameter Spanish instruction-tuned large language model based on Falcon-7B architecture, fine-tuned using 80,000 proprietary instruction data.

Aira-2 is the second version of the Aira instruction tuning series, a Portuguese instruction-tuned model based on GPT-2, suitable for text generation and dialogue tasks.

Historical multilingual and monolingual ByT5 base model, currently supporting English (British Library corpus - books).

A BART-large fine-tuned model for generating ChatGPT role settings based on character descriptions

A male single-speaker VITS model specifically designed for Persian text-to-speech, fine-tuned on the persian-tts-dataset-male based on the persian-tts-female-vits model

A vision-language model based on ViT and GPT2 architectures, specifically fine-tuned for Pokémon description generation tasks

A fine-tuned model based on the mbart-large-50 architecture, suitable for multilingual summarization tasks

Portuguese automatic speech recognition model fine-tuned based on tonyalves/output

Wav2Vec2 is a powerful speech recognition model that extracts features from raw audio through self-supervised learning and achieves high-performance speech recognition with limited labeled data.

A fine-tuned HuBERT extra-large speech recognition model based on 960 hours of Librispeech data, achieving a WER of only 1.8 on the LibriSpeech test set

Urdu automatic speech recognition model based on wav2vec2 architecture, fine-tuned on Common Voice dataset

A lightweight question-answering model based on MobileBERT, fine-tuned on the SQuAD v1.1 dataset, suitable for English Q&A tasks.

A fine-tuned abstract generation model based on PegasusXSUM for German datasets