Brief Details: TinyStories-33M is a GPT-Neo-based language model trained on simple stories dataset, optimized for generating straightforward narratives with 33M parameters.
Brief Details: A 2.8B parameter language model trained on RedPajama-1T dataset, offering base text generation capabilities with efficient GPU/CPU inference options
Brief Details: MPT-1b-RedPajama-200b is a 1.3B parameter decoder-only transformer trained on RedPajama dataset for 200B tokens, utilizing advanced features like FlashAttention and ALIBI.
Brief-details: A quantized 65B parameter Alpaca model available in multiple GGML formats (2-8 bit) for CPU+GPU inference, optimized for efficient local deployment using llama.cpp
Brief-details: Fine-tuned Stable Diffusion model trained on Fred Herzog's photography style, optimized for 768px resolution with specialized urban/street photography aesthetics.
Brief-details: AnimeScreencap is a Textual Inversion Embedding model for Stable Diffusion 2.x, specialized in warm, movie-stylized anime environments at 768x768 resolution
Brief-details: Quantized version of Meta's Llama-3 8B instruction model offering multiple compression formats (Q2-Q8) for different RAM/performance tradeoffs.
Brief Details: A fine-tuned Stable Diffusion model specialized in generating photorealistic images of Shah Rukh Khan, trained using DreamBooth technology with the instance prompt "a photo of srkay man"
Brief-details: A comprehensive collection of large language models (70B-120B parameters) quantized to 2-bit precision, optimizing storage while maintaining performance
Brief Details: Chinese CLIP model using ViT-B/16 image encoder and RoBERTa-wwm-base text encoder, trained on 200M Chinese image-text pairs for multimodal understanding and zero-shot classification.
Brief Details: A powerful 7B parameter code-specialized LLM supporting 92 programming languages with 64K context length, built on Qwen1.5 architecture for superior code generation and understanding.
Brief-details: IDEFICS2 8B chatty variant - Advanced multimodal model capable of processing interleaved image-text sequences for chat-like interactions, built by HuggingFace.
Brief-details: Microsoft's SpeechT5 voice conversion model - Transformer-based unified speech/text framework for converting speech between voices, built on CMU ARCTIC dataset.
Brief-details: SomethingV2 is an anime-focused text-to-image diffusion model optimized for vibrant but soft anime-style images, featuring built-in VAE and specific recommendations for high-quality outputs.
Brief-details: A specialized 11B parameter LLaMA-based model fine-tuned for creative text generation, optimized for long-form narrative content with enhanced vocabulary and genre diversity.
Brief Details: Qwen1.5-110B: Advanced 111B parameter language model with 32K context length, part of Qwen2 beta series. Supports multilingual tasks and improved chat capabilities.
BRIEF DETAILS: Medical-focused 8B parameter LLaMA3 model fine-tuned for healthcare Q&A. Optimized for clinical discussions with BF16 precision and medical domain expertise.
Brief-details: Mamba-2.8b-hf is a state-of-the-art 2.77B parameter language model using innovative Mamba architecture, optimized for efficient text generation and inference.
Brief-details: Bonito-v1 is a specialized text-to-text generation model based on Mistral-7B, designed for creating synthetic instruction tuning datasets from unannotated text.
Brief-details: MADLAD-400-10B-MT: A powerful 10.7B parameter multilingual translation model supporting 419 languages, based on T5 architecture with state-of-the-art performance.
Brief-details: Gorilla OpenFunctions v1 is an advanced LLM that converts natural language into executable API calls, supporting parallel functions and multiple function selection.
