Brief-details: A specialized LoRA model for generating Bondrewd character from Made in Abyss, featuring two distinct tokens (bdsks/bfsks) for normal and transformed appearances.
Brief Details: CySecBERT: A specialized 110M parameter BERT model adapted for cybersecurity tasks, trained on 4.3M entries of security-related content from multiple sources.
Brief-details: A text-to-image diffusion model combining 7th Heaven, Izumi, AbyssOrangeMix, and Anything3.0, optimized for anime-style character generation with detailed control over artistic elements.
BRIEF DETAILS: Multilingual Longformer model supporting 94 languages with 16384 token context window, based on XLM-RoBERTa. Features 768 hidden size and 12 layers.
Brief Details: SpeechBrain's Whisper Large-v2 model fine-tuned for Mongolian ASR, achieving 64.92% WER on CommonVoice test set, with frozen encoder and fine-tuned decoder architecture.
Brief Details: AudioLDM-S-Full is a text-to-audio generation model using latent diffusion, capable of creating diverse audio content from textual descriptions.
Brief-details: YOLOv8n model for hard hat detection with 83.6% mAP@0.5. Supports detection of hardhats vs no-hardhats. Built on Ultralytics framework.
BRIEF-DETAILS: YOLOv8m model specialized in aircraft detection with 99.5% mAP@0.5, supporting single-class detection of planes. Built on Ultralytics framework with 4.8K+ downloads.
Brief-details: YOLOv8m model specialized for table detection in documents, supporting bordered and borderless tables with 95.19% mAP@0.5 accuracy. Built on Ultralytics framework.
Brief Details: YOLOv8n model specialized for table detection in documents, achieving 96.7% mAP@0.5 accuracy. Supports bordered and borderless table detection.
Brief-details: YOLOv8-based segmentation model for PCB defect detection, achieving 56.8% mAP@0.5 for box detection and 55.7% for mask segmentation. Specializes in identifying 4 defect types.
Brief-details: YOLOv8-based Pokemon classification model supporting 150+ Pokemon species. Features 3.2% top-1 accuracy and 9.7% top-5 accuracy. Built on ultralytics framework.
BRIEF DETAILS: A comprehensive StableDiffusion-based painting model focused on illustration styles, featuring custom hypernetworks and embeddings for enhanced artistic generation.
Brief-details: Doc-UFCN model for historical document text line detection, trained on 10 datasets. Achieves up to 98% AP@.5 on ScribbleLens dataset. MIT licensed.
Brief Details: Text2LIVE is an innovative AI model for zero-shot text-driven image and video editing, enabling texture manipulation and visual effects without pre-training.
Brief Details: YOLOv8-based pothole segmentation model achieving 89.5% mAP@0.5 for mask detection. Built for road maintenance and infrastructure monitoring applications.
Brief-details: A versatile LoRA model collection featuring various artistic styles including alley backgrounds, impasto, photo portraits, and specialized styles like Fumo, NOZ watches, and ink painting
Brief-details: A speech processing model using wav2vec2 architecture with clustering and cross-contrastive loss, achieving 15.6% WER improvement on LibriSpeech data.
Brief-details: German T5 model (142M params) trained on GC4 corpus for text generation tasks, optimized for German language processing with MIT license
Brief-details: A Dreambooth-trained model specialized in generating D&D-style cover art images, built on Stable Diffusion v1-5 base model with CreativeML OpenRAIL-M license.
BRIEF-DETAILS: Specialized coloring book image generation model trained on 100+ line art images. Creates black & white illustrations perfect for coloring activities.
