Brief-details: A specialized Textual Inversion Embedding model for Stable Diffusion 2.0, trained on 44 V-Ray style images to create high-quality rendered looks with soft outputs.
Brief-details: PyTorch implementation of text-to-3D DreamFusion powered by Stable Diffusion, enabling 3D model generation from text prompts with efficient rendering capabilities.
Brief-details: T5-base model fine-tuned on WikiSQL dataset for English-to-SQL translation tasks, featuring 56K+ training samples and Apache 2.0 license.
Brief Details: A T5-based question generation model that creates natural reading comprehension questions from given answers and context, trained on SQuAD, CoQA, and MSMARCO datasets.
Brief-details: DALL·E mini is an open-source text-to-image generation model built with JAX/Flax, offering a simplified version of OpenAI's DALL·E that runs on less demanding hardware.
Brief Details: Advanced image-to-video generation model capable of creating 16fps videos up to 10 seconds long, supporting high resolutions up to 1360x768 with BF16 precision.
Brief Details: 4-bit quantized Llama 3.1 (8B) model with optimized performance (99.3% of original), reduced VRAM usage (6.1GB), and faster decoding
Brief-details: A 72B parameter reward model for mathematical reasoning, designed to enhance Qwen2.5-Math training by providing detailed feedback on reasoning steps and solution quality.
BRIEF DETAILS: Extended context Llama-3 70B model (262k tokens) optimized for instruction-following tasks. Built by Gradient AI with NTK-aware RoPE scaling.
Brief Details: A specialized text-to-image model based on Stable Diffusion, featuring rich coloring and anime-style artwork with multiple versions optimized for different artistic styles.
BRIEF-DETAILS: 1.8B parameter chat model by H2O.ai, based on Llama 2 architecture with Mistral sliding window attention, optimized for conversation and knowledge tasks.
Brief-details: A 7B parameter Mistral-based GPTQ-quantized model fine-tuned with DPO, featuring improved multi-turn capabilities and strong performance on benchmarks like MTBench
Brief-details: A comprehensive PDF processing toolkit combining layout analysis, form detection, recognition, and table reconstruction capabilities using advanced ML models
Brief-details: Florence-2-DocVQA is an 823M parameter image-text-to-text model fine-tuned from Microsoft's Florence-2, specialized for document visual question answering tasks.
Brief Details: Powerful 46.7B parameter Mixtral-based LLM fine-tuned on 1M+ GPT-4 generated entries, offering strong performance across various benchmarks with ChatML support.
Brief-details: A 7B parameter GPTQ-quantized chatbot model fine-tuned on UltraChat and UltraFeedback datasets, optimized for helpful assistant-style interactions
Brief-details: 7B parameter GGUF model optimized for text generation, featuring multi-language support (EN/ZH), high MMLU scores (63.82%), and various quantization options for efficient deployment.
Brief Details: A 6.74B parameter LLM specialized in legal domain knowledge, built on LLaMA-1-7B architecture with continuous pre-training and reading comprehension optimization.
Brief-details: Extended context (32K) instruction-tuned Llama2 7B model in GGUF format, optimized for efficient inference with multiple quantization options
Brief-details: Facebook's SeamlessM4T Large - A unified multilingual model supporting 101 languages for speech input, 96 for text, and 35 for speech output. Enables S2ST, S2TT, T2ST, T2TT, and ASR tasks.
Brief Details: High-performance video-to-video generation model capable of 720P output. Specializes in spatiotemporal continuity and resolution enhancement with OpenCLIP integration.
