Brief Details: YOLOv5m-based smoke detection model achieving 99.47% mAP@0.5, optimized for real-time smoke detection with PyTorch implementation
Brief-details: Privacy policy intent classifier achieving 88% F1-score, fine-tuned on PolicyIE dataset with 5 distinct privacy-related categories
BRIEF DETAILS: Stable Diffusion v2 embedding model for creating striking silhouetted landscapes with colorful backgrounds, trained on Inkpunk and Dreamlike Diffusion merger
Brief Details: GIT-large-vatex is a generative image-to-text Transformer model fine-tuned on VATEX, specializing in video captioning and visual tasks.
BRIEF DETAILS: GIT-base-vatex is a 177M parameter vision-language model fine-tuned on VATEX, specialized in video captioning and visual question answering using CLIP image tokens.
Brief Details: Financial sentiment analysis model fine-tuned on Twitter data. 110M params, achieves 88.4% accuracy for classifying financial tweets as bullish/bearish/neutral.
Brief Details: YOLOv5m model specialized for license plate detection with 98.8% mAP@0.5 accuracy. Built on PyTorch, supports real-time inference and custom training.
Brief Details: YOLOv5m-based object detection model specialized for forklift detection, achieving 85.15% mAP@0.5 on validation data. Popular with 1000+ downloads.
BRIEF DETAILS: Multilingual text-to-image model fine-tuned on Stable Diffusion, featuring unique namespace control system and support for 4 languages. Optimized for diverse image generation.
Brief-details: YOLOv5m model fine-tuned for construction safety object detection, achieving 0.37 mAP@0.5 on validation. Supports real-time safety monitoring.
Brief Details: YOLOv5n model specialized for construction safety object detection, achieving 0.37 mAP@0.5 on validation. Popular with 1100+ downloads.
Brief Details: EkmanClassifier is a BERT-based emotion classification model that identifies six universal emotions (happiness, sadness, anger, fear, disgust, surprise) in text.
Brief-details: A DistilBERT-based model fine-tuned for logical fallacy classification across 14 categories, achieving high accuracy in identifying reasoning flaws.
BRIEF-DETAILS: Japanese implementation of Facebook's Generative Spoken Language Model (GSLM) for textless NLP, featuring speech-to-unit and unit-to-speech conversion capabilities.
BRIEF DETAILS: Specialized cinematic image generation model based on SD 1.5, optimized for 16:9 ratio, requires "syberart" keyword trigger. Best for movie-style scenes & portraits.
Brief-details: Anime-focused text-to-image model with 123M parameters, optimized for high-quality anime art generation using Stable Diffusion. Supports danbooru tags and detailed prompting.
Brief Details: Arabic text summarization model built on BERT architecture, optimized for MSA content with 23.9g CO2 emissions and Rouge-L score of 1.137
Brief-details: A Vision Transformer (ViT) model with 22.9M parameters for image classification, pre-trained on ImageNet-21k and fine-tuned on ImageNet-1k with augmentation.
Brief-details: Vision Transformer (ViT) model with 22.1M params, trained on ImageNet-1k. Features 224x224 input, 16x16 patches, augmentation-enhanced training.
Brief-details: Vision Transformer (ViT) model trained on ImageNet-21k, featuring 104M params, patch size 32, and advanced augmentation techniques for superior image classification.
Brief-details: A fine-tuned Whisper ASR model specialized for Tamil language, achieving 6.5% WER on Common Voice test set, trained on multiple Tamil ASR corpuses.
