KaniTTS is a high-speed, high-fidelity text-to-speech model optimized for real-time conversational AI applications. It adopts a two-stage pipeline architecture, combining a large language model and an efficient audio codec, achieving excellent speed and audio quality. It supports multiple languages and is suitable for edge/server deployment.

Audio ProcessingTransformersMultiple Languages

Zen Nano 0.6B is the smallest model in the Zen model family, optimized for edge deployment and capable of running efficiently on edge devices, providing strong support for related applications.

Natural Language ProcessingSafetensorsEnglish

This is a statically quantized version based on the kshitijthakkar/loggenix-moe-0.3B-A0.1B model, specifically optimized for low-resource environments and edge AI applications, offering multiple quantization levels.

Natural Language ProcessingTransformersEnglish

AFM-4.5B is a 4.5-billion parameter instruction-tuned model developed by Arcee.ai, designed specifically for enterprise-level applications. It supports multilingual processing and demonstrates excellent performance in various deployment environments from the cloud to the edge.

Natural Language ProcessingTransformersMultiple Languages

LogicLlama is a fine-tuned language model developed by Goppa AI, based on the 1-billion-parameter version of LLaMA, enhanced with knowledge injection and logical reasoning capabilities, suitable for edge device applications.

Natural Language ProcessingTransformers

MobileLLM is an optimized Transformer architecture language model developed by Meta, specifically designed for resource-constrained edge-side applications, with parameter scales ranging from 125M to 1.5B.

Natural Language ProcessingTransformersEnglish

GLM-Edge-V-2B is a vision-language model that supports image-text to text conversion. It is designed for edge computing scenarios, has a scale of 5 billion parameters, and can handle tasks such as image description. It is suitable for the integrated applications of natural language processing and computer vision.

MultimodalSafetensors

Xinyuan-VL-2B is a high-performance multimodal large model for edge-side applications launched by Cylingo Group, fine-tuned based on Qwen/Qwen2-VL-2B-Instruct, utilizing over 5 million multimodal data points and a small amount of pure text data.

MultimodalTransformersMultiple Languages

This is an ultra-compact, high-performance zero-shot classification model designed specifically for edge devices and browser applications.

Natural Language ProcessingTransformersEnglish