Gitee AI provides a platform for AI model exploration, inference, training, deployment, and application development with serverless APIs and GPU compute services.
Gitee AI offers a comprehensive AI platform with multiple services. Its Serverless API provides on-demand, auto-scaling inference for large language models, supporting text, image, audio, and multimodal models with OpenAI SDK compatibility. The platform includes model fine-tuning services for customization with data preprocessing and training tracking. Reserved GPUs offer flexible GPU compute rental for training and inference, supporting NVIDIA and domestic hardware like Ascend and MetaX. An App Store enables developers to publish and monetize AI applications. Key strengths include support for domestic compute hardware, low-latency inference, and pay-as-you-go pricing. Typical use cases involve AI model deployment, custom model training, and application development.