BRIEF DETAILS: Anole-7b is a groundbreaking open-source multimodal model capable of interleaved image-text generation, building upon Chameleon with advanced capabilities.
BRIEF-DETAILS: SD3-Controlnet-Canny is a specialized controlnet model for SD3, optimized for 1024x1024 resolution with canny edge detection, featuring adjustable control weights for precise image generation.
Brief Details: Poro-34B is a 34B-parameter bilingual LLM trained on Finnish and English with code capabilities, using BLOOM architecture and 1T tokens.
Brief-details: Natural-SQL-7B is a 6.91B parameter LLM fine-tuned for Text-to-SQL conversion, offering superior performance in complex query generation and natural language understanding.
BRIEF-DETAILS: Multilingual translation model based on Llama-2-13B, supporting 53 languages with English as source. Features one-to-many translation capabilities and PyTorch integration.
Brief Details: StableLM-Tuned-Alpha 3B is an open-source chat-focused LLM with 4096 hidden size and 16 layers, fine-tuned on multiple instruction datasets for enhanced dialogue capabilities.
BRIEF-DETAILS: Camel-5b is a 5B parameter instruction-following LLM trained on 70k expert-curated prompts, offering strong performance for text generation and natural language tasks
Brief-details: YOLOv8-based model for real-time stock market trend prediction. Achieves 0.649 mAP@0.5 accuracy for detecting Up/Down trends in live trading data. Built for automated trading systems.
Brief Details: Instruction-tuned text embedding model that adapts to specific tasks via prompting. Achieves SOTA on 70+ embedding tasks, supports classification, retrieval & clustering.
Brief Details: BLOOM-zh is a 1.1B parameter Traditional Chinese-enhanced language model trained on 11.5B tokens, developed by MediaTek Research and partners for non-commercial research.
Brief Details: A powerful MoE-based chat model with 14.3B total params but only 2.7B active during runtime. Offers 1.74x faster inference vs Qwen1.5-7B.
Brief-details: A specialized repository containing LoRA networks and textual inversion embeddings for Stable Diffusion, focusing on character and outfit customization with 112 likes
Brief-details: LGM is a fast Text-to-3D and Image-to-3D model using Gaussian Splatting, generating high-resolution 3D content in 5 seconds with 415M parameters.
Brief-details: A powerful 25.5B parameter multimodal model combining InternViT-6B vision encoder and InternLM2-20B language model for advanced image/video understanding and generation.
Brief Details: Powerful 72.7B parameter chat model rivaling GPT-4, excelling in math, coding, and instruction-following. Built on Qwen2.5, optimized via RLHF.
Brief-details: Compact 1.1B parameter chat model based on Llama 2 architecture, trained on 3T tokens. Available in various GGUF quantizations for efficient CPU/GPU inference.
Brief-details: Image Mixer is an AI model for combining concepts and styles from multiple images, fine-tuned from Stable Diffusion, trained on LAION-5B aesthetics dataset at 640x640 resolution.
Brief-details: Long-context Llama 2 variant capable of processing 128k tokens, built by NousResearch using Flash Attention 2, optimized for extended context processing
Brief-details: A powerful 16.4B parameter MoE (Mixture of Experts) chat model by DeepSeek AI, featuring BF16 precision and specialized for conversational AI tasks.
Brief-details: 7B parameter instruction-tuned code generation model optimized for GGUF format, offering multiple quantization options and excellent performance for coding tasks
Brief Details: 13B parameter LLaMA-based model optimized for CPU/GPU inference with GGML quantization, offering various compression levels from 2-bit to 8-bit
