Brief Details: An 8.03B parameter LLaMA model optimized for mesh generation, converted to GGUF format with Q6_K quantization for efficient deployment via llama.cpp
Brief Details: A specialized 8B parameter LLaMA-based translator supporting mutual translation between Korean, Vietnamese, Indonesian, Cambodian, and Thai with strong BLEU/ROUGE scores.
Brief Details: Qwen2.5-Coder-14B-Instruct-F16-GGUF is a 14.8B parameter coding-specialized language model optimized for 16-bit inference using GGUF format, compatible with Llama.cpp
Brief Details: GENIE_zh_7b is a specialized 7.62B parameter Chinese language model for structuring electronic health records (EHR), built on Qwen2.5-7B-Instruct.
Brief-details: Advanced sentence embedding model based on NV-Embed-v2, fine-tuned for medical text similarity with 7.85B parameters and 4096-dim outputs
Brief-details: A quantized 7.8B parameter instruction-tuned language model supporting English and Korean, using AWQ 4-bit quantization for efficient deployment while maintaining performance.
Brief-details: 8B parameter storytelling-focused LLM merged using DARE-TIES method, combining Hathor, LexiMaid & Chara models. Optimized for RP/narrative generation.
BRIEF DETAILS: Lightweight 1.24B parameter GGUF-quantized Orca model optimized for agent-based instruction following, offering multiple quantization options from 0.7GB to 2.6GB with varying quality-size tradeoffs.
Brief-details: A massive 17B parameter self-merge of Flux.1-dev model, requiring 35-40GB VRAM for inference. Specialized in text-to-image generation with extensive resource demands.
Brief Details: A fine-tuned text-to-image model based on XenoGASM-MK2, specializing in artistic and versatile image generation with anime influences.
Brief-details: A specialized LoRA model for cartoon-style image generation, built on FLUX.1-dev. Features 64 network dimensions, trained on 22 images with constant LR scheduling and AdamW optimization.
Brief-details: A specialized LoRA model for high-quality poster generation built on FLUX.1-dev, featuring 64 network dimensions and optimized for 768x1024 resolution poster creation.
BRIEF DETAILS: A specialized LoRA model for generating motivational quote stickers, built on FLUX.1-dev. Features 64 network dimensions and optimized for 768x1024 resolution.
Brief-details: A 70B parameter LLaMA-based model optimized for roleplaying conversations, trained on 13,000 conversation pairs with rich character interactions and emotional expressions.
Brief-details: Qwen2.5 72B model with multiple GGUF quantizations (25GB-77GB), optimized for different hardware setups and RAM constraints. Features imatrix quantization for enhanced performance.
Brief Details: A fine-tuned RoBERTa-based model for off-topic classification, achieving 0.99 ROC-AUC with binary classification capabilities for enterprise LLM applications.
BRIEF DETAILS: Bulgarian-English language model based on Gemma-2-2b with 2.6B parameters, optimized for instruction-following and conversation. Enhanced Bulgarian capabilities while maintaining English performance.
BRIEF DETAILS: Fine-tuned Jina Embeddings model for off-topic classification with impressive 0.99 ROC-AUC score, supporting 1024 token context length for enterprise use.
Brief-details: Quantized INT4 version of Meta's Llama-3.2-3B-Instruct model optimized for NVIDIA GPUs, offering efficient inference with reduced memory footprint and ONNX runtime support.
Brief Details: ACE-0.6B-1024px is a unified visual generation model supporting multi-modal inputs and long-context processing for image editing and generation tasks, with 1024px resolution capabilities.
Brief-details: AstroSage-8B: Specialized 8B-parameter LLM for astronomy/astrophysics, outperforming GPT-4o on domain tasks. Built on Llama 3.1.
