Ollama pulls local models on a gpu. ollama library, cli run, privacy, open weights, modelfile, gpu layers, metal cuda, quantize, prompt window, keep alive, manifest, digest, numa, rocblas.
ollama ollama-ui ai-memory ollama-tools docker-ollama ollama-android ollama-cli local-ai-agents ollama-library ollama-agent ollama-cloud ollama-models ollama-code-agent ollama-proxy ollama-desktop-agent ollama-stack local-llm-ollama ollama-windows ollama-gpu ollama-setup
-
Updated
Nov 21, 2026 - C++