Image-Morpher/logs/71seconds.log

2025-03-09 17:59:17,957 - ERROR - Model loading failed: [Errno 2] No such file or directory: '/mnt/slurm_home/nalin/.cache/huggingface/hub/models--DarkFlameUniverse--Stable-Diffusion-2-1-Base-8bit/refs/main', switching to standard 2-1 base.
2025-03-09 18:00:29,214 - INFO - Total execution time: 71.52 seconds
2025-03-09 18:00:29,215 - INFO - Model Path: DarkFlameUniverse/Stable-Diffusion-2-1-Base-8bit
2025-03-09 18:00:29,215 - INFO - Image Path 0: ./assets/Trump.jpg
2025-03-09 18:00:29,215 - INFO - Image Path 1: ./assets/Biden.jpg
2025-03-09 18:00:29,215 - INFO - Use LCM: True
2025-03-09 18:00:29,215 - INFO - Number of inference steps: 8
2025-03-09 18:00:29,215 - INFO - Guidance scale: 1

71 seconds with standard 2-1.. and the vae attention slicing ?! and cudnn benchmark.

lcm-lora works with the stabilityai/sd2-1 model card? 

i think for this run the lora was already partially trained so it picked it up and trained the rest, so it was quicker