Brief-details: A lightweight LLaVA variant with 1.05M parameters, designed for image-text-to-text tasks using Transformers architecture and Safetensors format.
Brief Details: 8B parameter code-focused LLM from IBM, fine-tuned for programming tasks across multiple languages with strong performance in Python/Java.
Brief Details: A 3.4B parameter transformer-based language model optimized for text generation, using BF16 precision and built on the LLaMA architecture.
Brief-details: Advanced text embedding model by Salesforce with 7.11B parameters, optimized for research tasks like retrieval and classification with high performance across MTEB benchmarks
Brief-details: 4-bit quantized version of Gemma 2 9B instruction model optimized for faster inference and lower memory usage with Unsloth's optimization techniques
Brief-details: A 13B parameter uncensored LLaMA2-based model optimized for reduced refusals and bias, available in multiple GGUF quantizations for efficient CPU/GPU inference
Brief Details: AWPortrait-FL: A specialized portrait generation model built on FLUX.1-dev, offering enhanced skin details and composition for high-quality fashion photography.
BRIEF DETAILS: Multilingual XLM-RoBERTa model trained on 198M tweets, optimized for Twitter text analysis and sentiment tasks across 30+ languages.
Brief Details: A 4-bit quantized version of Meta's Llama 3.1 70B model optimized for efficiency with Unsloth, offering 70% reduced memory usage and faster inference.
BRIEF DETAILS: 3B parameter multilingual LLM optimized for dialogue, supporting 8 languages. GGUF quantized versions available with RAM requirements from 4.66GB to 9.38GB.
Brief-details: A lightweight 15M parameter LLaMA2-based text generation model optimized for story creation, implemented in Transformers.js for browser-based applications
Brief-details: Anything-XL is a specialized text-to-image SDXL model focused on anime-style generation, offering high-quality outputs with structured prompting and flexible resolution support.
Brief-details: SSD-1B is a distilled version of SDXL, offering 60% faster performance while maintaining quality. Features 1.3B parameters and supports multiple resolutions.
BRIEF DETAILS: 4-bit quantized version of Mistral-Nemo-Instruct-2407 with 2.8B parameters, optimized for efficient deployment using GPTQ quantization method
Brief-details: A fine-tuned BERT model for Turkish Named Entity Recognition, achieving 96.17% F1-score. Handles person, organization, and location entities with high accuracy.
Brief-details: Large-scale Vision Transformer model pre-trained on ImageNet-21k (14M images) and fine-tuned on ImageNet-1K. Specializes in image classification using 16x16 patches.
BRIEF-DETAILS: A specialized LLaMA-based accelerator model (199M params) designed for speculative decoding, featuring multi-stage MLP architecture for faster inference
Brief-details: NLLB-200 is a powerful 3.3B parameter multilingual translation model supporting 200+ languages, developed by Facebook for research purposes.
Brief-details: A powerful multilingual translation model supporting 200 languages, distilled to 1.3B parameters with strong performance on low-resource languages.
Brief Details: LLaVA-NeXT (v1.6) - 13B parameter multimodal model combining vision and language capabilities with improved OCR and reasoning abilities
BRIEF DETAILS: FinancialBERT model specialized for sentiment analysis in financial texts. Achieves 98% accuracy with 3-way classification. Pre-trained on financial corpora.
