AION
Open-source releaseMLOps, Tooling & Infrastructure1 source · Aug 26, 2026

ollama/ollama v0.33.1

mlxrunner: avoid Metal GPU timeouts when loading models from slow storage

Qwenllama.cppOllamaMLX
← Back to brief

Key points

  • cmake: make external compat patches idempotent

Sources (1)

  • [1]ollama/ollama v0.33.1
    GitHub: ollama/ollama · Aug 26, 06:09 PM
    * mlxrunner: avoid Metal GPU timeouts when loading models from slow storage
    * cmake: make external compat patches idempotent

Extractive summary: sentences quoted from the sources.