# High accuracy

This model is a dialogue turn detection model fine-tuned based on Qwen3-0.6B. It uses the Unsloth and Huggingface TRL libraries to accelerate training, achieving an accuracy of 96.22% in the dialogue turn detection task.

This model is an English sentiment analysis model based on the RoBERTa architecture, capable of classifying text into positive, neutral, and negative categories.

A Russian emotion classification model fine-tuned based on lightweight RuBERT-tiny2, capable of recognizing five emotions

This is an efficient few-shot learning model based on SetFit, used for text classification tasks. It uses sentence-transformers/paraphrase-MiniLM-L6-v2 as the sentence embedding model and LogisticRegression for classification.

A Vietnamese text classification model fine-tuned from vinai/phobert-base-v2, achieving 89.62% accuracy on the validation set

PhoBERT is a Vietnamese pre-trained language model optimized based on the RoBERTa architecture, suitable for Vietnamese text classification tasks.

A sentiment analysis model fine-tuned from vinai/phobert-base-v2, achieving 93.6% accuracy on the evaluation set

This model is a fine-tuned version of FacebookAI's roberta-base on the Webi-CPC-11 dataset, primarily used for text classification tasks, achieving an accuracy of 84.32% on the evaluation set.

This model is an audio classification model based on the Audio Spectrogram Transformer (AST) architecture. After pre-training on the Audioset dataset, it was fine-tuned on the GTZAN music genre classification dataset.

This model is a speech classification model fine-tuned based on facebook/hubert-base-ls960, mainly used for binary classification tasks, and achieved an accuracy of 78.26% on the evaluation set.

This is a BERT model fine-tuned for spam detection with a validation accuracy of approximately 99%

A zero-shot classification model fine-tuned based on the Turkish BERT model, trained on the XNLI Turkish dataset, suitable for Turkish text classification tasks.

A sentence embedding model fine-tuned on the Korean triplet dataset based on the Alibaba-NLP/gte-multilingual-base model for semantic similarity calculation

This is a sentence-transformers model fine-tuned on a Korean triplet dataset based on Alibaba NLP/gte-multilingual-base, designed for semantic textual similarity tasks.

GPT-2 fine-tuned language identification model supporting text classification in 90 languages

A deepfake audio detection model fine-tuned based on facebook/wav2vec2-base, used to identify synthetic or tampered speech content

CodeAstra-7b is a fine-tuned language model specifically designed for vulnerability detection in multiple programming languages, capable of identifying potential security vulnerabilities.

A named entity recognition model specifically designed for Uzbek text, built on the XLM-RoBERTa large architecture, supporting multiple entity category recognition.

This model is trained on a large dataset of manually scored IELTS essays and can automatically score IELTS essays across multiple dimensions with an accuracy of 0.82.

A large language model fine-tuned via DPO based on Meta-Llama-3-70B-Instruct, demonstrating excellent performance in multiple benchmark tests

A lightweight Uzbek sentiment analysis model based on DistilBERT for text sentiment classification

A language identification model fine-tuned based on the Whisper Medium model, specifically designed for language classification tasks on the FLEURS dataset

Rhea-72b-v0.5 is a large language model fine-tuned based on Smaug-72B-v0.1, ranking first on the HuggingFace Open LLM Leaderboard.

A tabular classification model based on scikit-learn for predicting possible diseases based on symptoms.

A large GPT2 model trained on the Anthropic/hh - rlhf harmless dataset, specifically for harmful response detection or reinforcement learning from human feedback (RLHF).

A BERT-large-uncased fine-tuned adult text classification model for identifying adult content in text.

A sentiment analysis model for Amazon food reviews fine-tuned on distilbert-base-uncased, capable of classifying reviews into negative (-1), neutral (0), and positive (1) sentiments

Fine-tuned from a pre-trained Indonesian BERT model, used for sentiment analysis of Indonesian comments and reviews, and can classify text into three categories: negative, neutral, and positive.

A news classification model based on BERT-base-uncased, capable of accurately classifying various news categories.

Thai sentiment analysis model fine-tuned from WangchanBERTa, achieving 74.15% accuracy on the evaluation set

A speech accent multiclass classification model based on facebook/wav2vec2-base-960h, used to identify different speech accents.

A speech recognition model fine-tuned on the speech_commands dataset based on facebook/wav2vec2-base, achieving 97.35% accuracy

A binary audio classification model fine-tuned from facebook/wav2vec2-base, specifically designed to distinguish between the pronunciations of 'down' and 'on'

This is an audio classification model based on the AST (Audio Spectrogram Transformer) architecture, fine-tuned on the GTZAN music genre classification dataset with an accuracy rate of 92%.

A foundational speech recognition model based on the Wav2Vec2 architecture, trained on 960 hours of English speech data, suitable for speech classification tasks.

A zero-shot text classification model trained with Self-Supervised Tuning (SSTuning), based on ALBERT-xxlarge-v2 architecture, which can be directly applied to tasks like sentiment analysis and topic classification without fine-tuning.

A sequence classification model adapted from Flan-T5-xxl, specifically designed for natural language inference tasks with half the parameters and support for zero-shot classification

A voice processing model fine-tuned based on microsoft/wavlm-large, achieving 80% accuracy and 79.57% F1 score on the evaluation set

This model is a voice command recognition model fine-tuned on the speech_commands dataset based on facebook/wav2vec2-base, achieving an accuracy of 97.59%.

A speech recognition model fine-tuned on the speech_commands dataset based on facebook/wav2vec2-base, achieving 97.8% accuracy