Deep LearningNatural Language ProcessingModel Optimization
Evaluating RAM Requirements for Ollama LLMs
4 min read · 667 wordsBy Orandi Felix
** Ollama-XL exceeded the 61 GB available, even at batch size 1. This highlights the need for careful infrastructure planning with the largest models.
** When deploying Ollama models, always leave at least 20% of your system's RAM free for other processes to avoid memory swaps, which drastically impact performance.