Brief-details: Unstuffer v0.2 is an AI model by jondurbin focused on text preprocessing and cleaning, designed to improve text quality for language models and analysis tasks.
Brief-details: T5-based model for extracting atomic claims from summaries, achieving 73.4 F1 score. Developed by Babelscape for factuality evaluation in NLP.
Brief-details: 12B parameter hybrid SSM-Transformer model with 256K context length. Excels in long-form tasks, supports multiple languages, and offers efficient inference through various quantization options.
Brief-details: GPTQ-quantized version of Guanaco 65B with multiple quantization options (3-bit & 4-bit). Optimized for efficient GPU inference with various group sizes and configurations.
Brief Details: A specialized AI model focused on improving hand generation in images, likely a LoRA or fine-tuning targeting hand-related artifacts and deformities.
Brief-details: Gemma 1.1 7B Instruction-Tuned (IT) is Google's advanced language model offering balanced performance and efficiency, requiring explicit license agreement for access.
BRIEF-DETAILS: Meditron-7B is a medical-focused LLM with 7B parameters, fine-tuned from Llama-2-7B using medical literature and guidelines, designed for healthcare applications
Brief-details: OpenChat 3.5-1210 is a state-of-the-art 7B parameter open-source LLM that outperforms ChatGPT (March) and Grok-1, with enhanced coding and mathematical capabilities
Brief Details: Aya-23-35B is a large language model by CohereForAI with 35B parameters, focusing on research and commercial applications through Cohere's platform.
Brief Details: 13B parameter LLaMA-based multilingual model with enhanced Chinese capabilities, trained on 125B tokens. Supports translation, coding, Q&A, and more. Available as delta weights.
Brief-details: Advanced PDF document analysis model that segments and classifies page elements (text, titles, tables, etc.) with dual model approach - visual (VGT) and non-visual (LightGBM)
Brief Details: C4AI Command-R-Plus: A Cohere-developed model focused on advanced command processing and response generation, part of the C4AI suite.
Brief Details: UNI - A non-commercial academic AI model by MahmoodLab, restricted under CC-BY-NC-ND 4.0 license, requiring institutional verification for access
Brief Details: sv3d by StabilityAI - A specialized AI model focused on 3D generation and processing, requiring license agreement acceptance for usage
BRIEF DETAILS: Dual-expert MoE Mistral model specialized in roleplay & storytelling, combining RP models in Expert 1 and story models in Expert 2. 7B params.
Brief-details: A specialized ControlNet model optimized for refining and correcting malformed hands in AI-generated images using diffusion-based conditional inpainting techniques.
Brief Details: Mixtral-8x7B is a powerful language model by MistralAI featuring a mixture-of-experts architecture with 8 expert groups and 7B parameters
Brief-details: FinGPT's DOW30 stock forecasting model built on Llama-2-7b with LoRA fine-tuning, specialized for financial market prediction and analysis
BRIEF DETAILS: PepMLM-650M is a specialized AI model for generating peptide binders, leveraging masked language modeling to design sequences that can bind to target proteins without structural data.
Brief-details: A compact variant of Stable Diffusion XL, created by echarlaix, optimized for experimental testing and development with reduced computational requirements.
BRIEF DETAILS: LLaMA 3.2-based Korean-English vision-language model (5B params). Unique dual functionality - works as both vision-language and pure language model. Optimized for Korean OCR and selective knowledge reasoning.
