Brief Details: A powerful 176B parameter MoE model quantized to 4-bit precision, supporting 5 languages with 65K context window and sparse mixture of 8 experts.
Brief Details: MixTAO-7Bx2-MoE-v8.1 is a 12.9B parameter Mixture of Experts model with impressive benchmark scores (77.50 avg), featuring BF16 precision and Apache 2.0 license.
Brief-details: GGUF-compatible text embedding model for sentence similarity, optimized for RAG applications. 137M params, multiple quantization options from 48MB-262MB.
Brief-details: Powerful 7B parameter Mistral-based model with 16k context, optimized for code and general tasks, using ChatML format. Strong benchmark performance and efficient GGUF quantization.
BRIEF-DETAILS: A 34B parameter code-focused model merging Phind-CodeLlama and WizardCoder, optimized for Python/JavaScript programming with GGUF quantization variants
BRIEF DETAILS: A 7B parameter code assistant model fine-tuned on CodeLlama-7b, trained on 140k programming problems, achieving 63.1% pass@1 on HumanEval
Brief-details: GGML variant of OpenOrca-Platypus2-13B, a merged model combining Platypus2-13B and OpenOrcaxOpenChat for enhanced STEM and reasoning capabilities.
Brief-details: CodeLlama-7B-Python is a specialized 7B parameter LLM optimized for Python coding, available in GGUF format for efficient CPU/GPU inference
Brief-details: A powerful 13B parameter code generation model optimized for Python, achieving 64% pass@1 on HumanEval. Available in multiple GGUF quantizations for efficient deployment.
Brief-details: A 13B parameter LLM based on Vicuna, quantized to 4-bit precision. Trained on high-quality GPT-4 generated instructions with reduced censorship. Uses Alpaca prompt format.
Brief-details: A 13B parameter LLaMA-2 fine-tune trained on 242k GPT-4 generated entries, offering strong performance on reasoning and general instruction tasks.
BRIEF-DETAILS: Stanford Alpaca-7B weight diff model for reconstructing LLaMA-based language model. Features text generation capabilities with PyTorch integration.
Brief-details: OpenLLaMA 13B is an open-source reproduction of Meta's LLaMA model, trained on RedPajama dataset with Apache 2.0 license. Shows comparable performance to original LLaMA.
Brief-details: RWKV 7B model with groundbreaking 128k context window, optimized for novel writing and multi-language processing. Features 1:1 word-to-token ratio and trains on diverse datasets including web novels and wuxia content.
Brief-details: Mistral-7B-v0.1-GGUF is a powerful 7B parameter LLM optimized for efficient deployment, featuring sliding window attention and multiple quantization options for various performance needs
BRIEF DETAILS: Text-to-image LoRA model based on FLUX.1-dev, specialized in anatomical detail generation. 4 likes, 2.3k+ downloads, CC BY-NC 4.0 license.
Brief Details: NSFW-GEN-ANIME is a specialized anime text-to-image generator with 3.47B parameters, focusing on explicit content generation using the StableDiffusionXL pipeline.
Brief Details: Stable Diffusion-based image generation model optimized for anime-style artwork, featuring enhanced capabilities for character rendering and scene composition
Brief Details: BART-large fine-tuned model for generating ChatGPT prompts. Achieves 2.50 validation loss after 4 epochs. Apache 2.0 licensed.
Brief-details: A specialized embedding model for Stable Diffusion that improves image quality by identifying and avoiding "boring" characteristics in generated images. Created by FoodDesert with Apache 2.0 license.
Brief Details: A 350M parameter dialogue-optimized model fine-tuned from Facebook's OPT-350M, notable for efficient training using just 6GB VRAM in under an hour
