# Low learning rate optimization

A diffusion model based on PyTorch and Diffusers library for unconditional generation of black and white handwritten digit images.

This is a version of Microsoft's Florence-2 model fine-tuned for 1 day using the Docmatix dataset (5% of the data) with a learning rate of 1e-6

An English-Spanish translation model fine-tuned from facebook/mbart-large-50, demonstrating excellent performance on evaluation sets.

An English-to-Spanish translation model fine-tuned based on facebook/mbart-large-50, achieving a BLEU score of 45.09 on the evaluation set.

This model is a fine-tuned English-to-Spanish translation model based on facebook/mbart-large-50, achieving a BLEU score of 45.03 on the evaluation set.

A Turkish text summarization model fine-tuned on the mlsum/tu dataset based on google/mt5-base