InvokeAI-Benchmark is a simple tool to evaluate performance of InvokeAI.
pip install https://github.com/cloudmercato/invokeai-benchmark/archive/refs/heads/master.zip
$ invokeai-benchmark --api-url http://your-server:9090/api/v1/ -p "Large landscape of Dijon" batch_id : 809daf7d-d0a3-4066-a1df-74c2e530580f item_id : 151 session_id : 5dfa456e-7f58-4c94-8fd9-2321c400dd31 image_path : http://your-server:9090/api/v1/images/i/c6968dbe-f34a-4991-a07c-eeaf2157d9cb.png/full prompt : Large landscape of Dijon width : 512 height : 512 seed : 0 cfg_scale : 7.5 scheduler : euler steps : 50 fp32 : True priority : 0 error : None created_at : 2023-11-28 05:53:32.795 started_at : 2023-11-28 05:53:32.804 completed_at : 2023-11-28 05:53:42.510 queue_elapsed : 0.009 task_elapsed : 9.706 total_elapsed : 9.715 pixel_rate : 27008.448 version : 3.4.0post2
This project is created with ❤️ for free by Cloud Mercato under BSD License. Feel free to contribute by submitting a pull request or an issue.