Brief Details: Low-latency edge-device ASR model supporting English, French & German. Optimized for streaming, with commercial licensing planned. Built on k2/Icefall pipeline.
BRIEF-DETAILS: A 14.7B parameter LLM fine-tuned on GammaCorpus v2-500k dataset, built on Qwen 2.5. Features GQA attention with 40/8 head split and SwiGLU activation.
Brief Details: A 14.7B parameter LLM built on Qwen 2.5, fine-tuned on GammaCorpus v2-50k dataset with enhanced capabilities for structured dialogue generation
Brief-details: Qwen 2.5 14B-based model fine-tuned on GammaCorpus v2-10k dataset, featuring 14.7B parameters with advanced architecture elements like RoPE and SwiGLU.
Brief-details: A specialized 2.5D model merge focused on artistic image generation, combining multiple style approaches for enhanced visual output
Brief Details: A GGUF-formatted version of Rinna's Qwen 2.5 Bakeneko 32B instruction model, optimized for Japanese language tasks using TFMC/imatrix dataset
Brief-details: AI-infinity-V1-fp16 is a specialized model focused on realistic image generation with enhanced hand rendering capabilities, optimized in FP16 format for efficient deployment.
Brief Details: LLaMA 3 70B Instruct model optimized with AWQ quantization for efficient deployment while maintaining high performance in instruction-following tasks.
BRIEF-DETAILS: Optimized 4-bit quantized version of Qwen2.5-7B-Instruct using Unsloth's Dynamic Quantization, offering 60% memory reduction and 2x faster inference
BRIEF-DETAILS: A sophisticated AI image generation model focused on photorealistic outputs, known for high-quality facial details and natural lighting effects
BRIEF-DETAILS: DeepFloyd IF-II-L-v1.0: A non-commercial research-focused AI model for image processing with strict usage limitations and comprehensive legal safeguards.
BRIEF DETAILS: TrueSight is an advanced image generation model by Yntec, optimized for diverse scenarios from portraits to historical scenes, with particular strength in realistic human representation.
BRIEF-DETAILS: AI image generation model focused on realistic outputs, available on CivitAI and Hugging Face, developed by digiplay for photorealistic results.
Brief-details: IncredibleWorld is a creative AI model by Yntec specializing in generating festive and family-oriented scenes, with particular strength in brewing and holiday themes.
Brief Details: ViT5 model specialized in Sino-Vietnamese transliteration, trained on 450k MMORPG translations, featuring enhanced functionality from v1.0 with ONNX optimization
Brief-details: AnalogMadness v6 is a specialized AI model focused on creating realistic analog photography effects, developed by digiplay for artistic image generation
Brief Details: MGM - A specialized AI model focused on image generation tasks, available through HuggingFace's platform. Created by digiplay.
Brief-details: Runpod is a monitoring and statistics logging model developed by unslothai to track environment performance and detect potential issues in AI deployments.
Brief-details: SUPIR (SUPer Image Restoration) is an AI model focused on high-quality image restoration and enhancement, created by camenduru and hosted on Hugging Face.
Brief Details: CogVLM is a vision-language model developed by THUDM, designed for advanced visual understanding and multimodal interactions. Available on HuggingFace.
BRIEF-DETAILS: A 7B parameter medical variant of LLaMA 2 by llSourcell, fine-tuned for healthcare and medical domain tasks, hosted on HuggingFace.
