Repository navigation
Project Maya v1.0.6
A standard speed test: ./maya.sh --bench (Windows: START-MAYA.bat --bench) - decode and prefill on your machine in a few minutes, comparable everywhere. Update with git pull.
What's new
--benchmeasures the installed model on your machine in a few minutes, with Maya stopped: decode on the same
three questions everywhere (after a warm-up answer) and prefill at 2k and 8k tokens, through the engine the way the
dashboard starts it. It writesmaya-bench.txtwith the results and the engine's per-token breakdown;--report
includes it - so speed reports from different machines can be compared.- README: single-GPU speeds (1x Tesla V100 32 GB, 64 GB RAM: decode up to 19 tokens/s, prefill up to 370 tokens/s).