# Multi-round fine-tuning optimization

A Spanish-to-Spanish machine translation model fine-tuned based on Helsinki-NLP/opus-mt-es-es

A Kurdish Sorani dialect speech recognition model fine-tuned based on openai/whisper-small

A speech recognition model fine-tuned on Dutch datasets based on OpenAI Whisper Large V2, with a word error rate of 10.14%

This model is a fine-tuned Arabic-to-English translation model based on facebook/nllb-200-distilled-600M, achieving a BLEU score of 63.3172 on the evaluation set.

An automatic paraphrase model based on BART-large architecture, trained on the ParaBank 2 dataset, capable of generating high-quality English sentence paraphrases

A Tamil speech recognition model fine-tuned on the CommonVoice dataset, based on Harveenchadha/vakyansh-wav2vec2-tamil-tam-250

This is a Bengali speech recognition model fine-tuned on a specific dataset, primarily used for speech-to-text tasks.

Persian automatic speech recognition model fine-tuned on the COMMON_VOICE - FA dataset

Natural language inference model fine-tuned on MNLI task based on Catalan RoBERTa model

Text summarization model fine-tuned on the cnn_dailymail dataset, trained using the DistilBERT architecture