Builds mymodel from Modelfile while quantizing the weights to q4_K_M, a 4-bit format that shrinks memory use and speeds up CPU inference at a small accuracy cost. The -q flag only works when the Modelfile imports a source model, not for imported GGUF files. q4_K_M is a solid default; q8_0 trades size for fidelity.
Looking for more? Search all 7,657 commands — works offline, in English or Spanish, and fixes typos.