# High Precision F1

A binary toxicity classifier fine-tuned on xlm-roberta-large, supporting toxicity detection in 15 languages

A fasttext-based text classifier for categorizing text into mathematical or non-mathematical classes, suitable for LLM pre-training data organization

Image orientation correction model based on SwinV2 architecture for automatic detection and correction of image orientation

A text classifier fine-tuned based on RoBERTa-large, specifically designed to determine whether news articles involve government regulation topics.

A fine-tuned version of the microsoft/layoutlmv3-base model for document image classification tasks, achieving an F1 score of 0.8177 on the evaluation set

A medical symptom diagnosis classification model fine-tuned based on bert-base-cased, used to predict 22 possible diagnoses from symptom descriptions

A fine-tuned xlm-roberta-large model based on multilingual training data for zero-shot text classification, using the Manifesto Project coding scheme.

Gender recognition model fine-tuned on Librispeech-clean-100 dataset, achieving an F1 score of 0.9993 on the test set

Wojood is a corpus for Arabic nested named entity recognition (NER), containing 550,000 word instances (Modern Standard Arabic and dialects).

A software requirement classification model fine-tuned on sentence-transformers/all-MiniLM-L6-v2, designed to distinguish between functional (F) and non-functional (NF) requirements

A clinical text processing model fine-tuned based on Bio_ClinicalBERT, performing well in F1 score

A Chinese spelling correction model based on the T5 architecture, excelling on the SIGHAN2015 test set, supporting automatic correction of Chinese text.

A biomedical literature section classification model fine-tuned on PubMedBERT for identifying text section types

A model fine-tuned on the recipe ingredient NER dataset based on xlm-roberta-base, used to identify various labels in recipe ingredients.

A text sentiment classification model fine-tuned based on microsoft/deberta-v3-base, trained on an unknown dataset with an evaluated F1 score of 0.4468

SpanBERT is a pre-trained language model developed by Facebook Research, specifically fine-tuned on the TACRED dataset for relation extraction tasks.

HebEMO is a tool designed to detect sentiment polarity and extract emotions from modern Hebrew user-generated content (UGC), trained on a unique COVID-19 related dataset.

A Russian sentiment analysis model fine-tuned on the RuSentiment dataset based on XLM-RoBERTa-Large

A Transformer-based Persian language understanding model, reconstructed vocabulary and fine-tuned on new corpora, expanding multi-domain application capabilities

The first multilingual Twitter sentiment classification model for four Nigerian languages (Hausa, Igbo, Nigerian Pidgin, and Yoruba)

A Transformer-based Persian language understanding model, reconstructed vocabulary and fine-tuned on new Persian corpora to enhance multi-domain application capabilities