LIVE-Layerwise-Image-Vectorization icon indicating copy to clipboard operation
LIVE-Layerwise-Image-Vectorization copied to clipboard

speed issue

Open gyula-coder opened this issue 2 years ago • 3 comments

when I run the command 'python main.py --config config/base.yaml --experiment experiment_5x1 --signature smile --target figures/smile.png --log_dir log/', It takes about 15 minutes to generate the result svg. It need more time when processing my picture, about 12 minutes per path with the config of experiment_5x1.

Is that normal? How do I speed up the process. Thanks.

gyula-coder avatar Nov 28 '23 13:11 gyula-coder

when I run the command 'python main.py --config config/base.yaml --experiment experiment_5x1 --signature smile --target figures/smile.png --log_dir log/', It takes about 15 minutes to generate the result svg. It need more time when processing my picture, about 12 minutes per path with the config of experiment_5x1.

Is that normal? How do I speed up the process. Thanks.

hello bro ,did you solve this problem thanks for your replying

ZZZZZZHE avatar Mar 28 '24 06:03 ZZZZZZHE

@stu-GYULA @ZZZZZZHE Thanks for your interest.

Probably you need resize your image to a low resolution, like 240* 240 (in our experiments) or 320*320, etc. Image resolution will largely impact the speed.

Also, you can turn off some save options in base.yaml to speed up, https://github.com/Picsart-AI-Research/LIVE-Layerwise-Image-Vectorization/blob/main/LIVE/config/base.yaml#L8C5-L11C16

ma-xu avatar Mar 28 '24 20:03 ma-xu

LIVE progressively generates a svg from the raster image with path setting. I hard to find a trade off from speed and the quality of svg output.

gyula-coder avatar Apr 01 '24 13:04 gyula-coder