VRAM
AI & ComputeVideo Random Access Memory. The dedicated memory on a GPU used to store model weights and intermediate computations. VRAM capacity determines which AI models a GPU can run.
Back to Glossary
Related terms
AI InferenceThe process of running a pre-trained AI model to generate predictions or outputs from new input data. Inference is the most common GPU workload type, used for chatbots, image generation, and more.AI TrainingThe process of developing an AI model from scratch or fine-tuning an existing one using large datasets. Training requires significant GPU resources and is a key use case for the OpenGPU Network.Fine-tuningAdapting a pre-trained AI model to perform better on specific tasks by training it on additional, domain-specific data. Fine-tuning requires less compute than training from scratch.LLMLarge Language Model. An AI model trained on vast amounts of text data that can understand and generate human language. Examples include GPT, LLaMA, and Mistral. LLM inference is a primary OpenGPU workload.RAGRetrieval-Augmented Generation. An AI architecture that combines a language model with external knowledge retrieval to provide more accurate and up-to-date responses. Used for knowledge-intensive applications.EmbeddingsVector representations of data (text, images, etc.) in high-dimensional space. Embeddings enable semantic search, recommendation systems, and similarity matching in AI applications.
