Brief-details: An image-to-video model focused on anime-style transformations, featuring GGUF optimization and specialized workflows for quality adjustments (q4-q8 variants).
Brief-details: A super-realistic image generation model that requires "SuperRealism2" trigger word, available in Safetensors format for enhanced photorealistic outputs.
Brief-details: DINOv2-based model specialized for X-ray image analysis, developed by StanfordAIMI. Released in February 2023, optimized for 224x224 medical imaging tasks.
Brief-details: CogVLM - State-of-the-art visual language model with 17B parameters (10B vision, 7B language). Excels in cross-modal tasks and visual understanding.
Brief-details: Autoformer is a state-of-the-art time series forecasting model using decomposition transformers and auto-correlation, specifically designed for long-term predictions in tourism data.
Brief-details: Norwegian ASR model (39M params) fine-tuned for verbatim transcription, outputs lowercase text without punctuation, ideal for linguistic analysis
Brief Details: Stable Diffusion 1.5 (SD1.5) - A powerful text-to-image generative AI model, building on the success of the SD series with improved image quality and coherence.
Brief-details: RegNetY-3.2GF model for image classification with 19.4M params, trained on ImageNet-1k. Features enhanced implementation with stochastic depth and gradient checkpointing.
Brief-details: A specialized document layout analysis model based on DIT architecture, trained on DocLayNet for 4 epochs to segment 11 different document element types
Brief-details: A 4B-parameter multimodal LLM combining InternViT vision encoder with Qwen2.5-3B-Instruct, featuring Mixed Preference Optimization for enhanced reasoning across vision-language tasks.
Brief Details: ONNX-optimized DistilBERT model fine-tuned on SQuAD v1.1, achieving 87.1 F1 score. Efficient question-answering model with reduced parameters.
Brief-details: Swin2SR transformer-based model for 4x image super-resolution, optimized for real-world applications with BSRGAN architecture and PSNR metric
Brief Details: IndoBERT model fine-tuned for Indonesian sentiment analysis, achieving 94.52% accuracy and 94.51% F1 score on evaluation. Based on indobert-base-p1.
Brief-details: Anime-style merged model balancing artistic and realistic elements, optimized for danbooru prompts with emphasis on simple, quality-focused inputs.
Brief-details: A comprehensive archive of AI model checkpoints maintained by LMFResearchSociety, featuring over 9TB of backed-up models facing potential HuggingFace storage limitations.
Brief-details: 12B parameter GGUF quantized model with multiple compression variants (Q2-Q8), offering flexible performance-size tradeoffs. Features recommended Q4_K variants for balanced efficiency.
BRIEF: A 100B parameter Japanese language model optimized for instruction-following, available in GGUF format for efficient deployment using llama.cpp with CUDA support.
BRIEF-DETAILS: Quantized version of Stable Diffusion XL in GGUF format with multiple compression levels (Q4_K_S, Q5_K_S, Q8) for efficient deployment across different hardware configurations.
Brief-details: A quantized 12B parameter language model available in various GGUF formats, optimized for different size/performance tradeoffs, ranging from 3.1GB to 10.2GB.
BRIEF-DETAILS: 12B parameter merged LLM combining Magnum-Picaro v2, Violet-Lotus, Rocinante & Wayfarer models. Optimized for natural dialogue & adventure scenarios.
Brief-details: MN-Sappho-n-12B-GGUF is a quantized version of the MN-Sappho model, offering various compression formats from 4.9GB to 13.1GB with different quality-performance tradeoffs.
