# Mixed Precision Training

Kanana is a bilingual (Korean/English) language model series developed by Kakao, excelling in Korean tasks while remaining competitive in English tasks, significantly reducing computational costs compared to models of similar scale.

This is a T5 text summarization model fine-tuned on the SAMSum dataset, suitable for dialogue and conversation summarization tasks.

ModernBertMultilingual is a multilingual model trained from scratch, supporting Chinese, English, Japanese, and Korean, with excellent performance in mixed East Asian language text tasks.

This model is a fine-tuned speech synthesis (TTS) model based on microsoft/speecht5_tts, primarily used for text-to-speech conversion tasks.

A text-to-image generation model fine-tuned with Keras Dreambooth, capable of producing images in Bengali clay art style

This model is a fine-tuned version of bert-base-uncased on an unspecified dataset, suitable for natural language processing tasks.

This model is a fine-tuned version of Google's ViT model on an unknown dataset, primarily used for image classification tasks.

This is an automatic speech recognition (ASR) model trained on the LibriSpeech dataset, capable of converting English speech into text.

A wav2vec2-large-lv60 speech recognition model fine-tuned on the LibriSpeech dataset, supporting English speech-to-text tasks

A speech recognition model fine-tuned based on facebook/wav2vec2-base, with a word error rate (WER) of 1.0

A speech recognition model fine-tuned on the LibriSpeech ASR clean dataset, based on wav2vec2-base and bart-base

This model is a question-answering model fine-tuned on the SQuAD dataset based on distilbert-base-cased, specifically trained for the Keras.io question-answering tutorial.

ResNet50 v1.5 is an improved version of the original ResNet50 v1 model, achieving approximately 0.5% higher top1 accuracy by adjusting convolution strides.