Closed caniko closed 2 years ago
Are the results frequency? So Hz? Could you also include average time with SD between each inference?
Results are in Hz +/- SD already. More info here: https://github.com/DeepLabCut/DLC-inferencespeed-benchmark
Also see Table 1 of the paper: https://elifesciences.org/articles/61909
Thanks!
Does the published DLC-live benchmarks help? https://deeplabcut.github.io/DLC-inferencespeed-benchmark/
Not sure if all models have listed VRAM in their specs, but you could probably infer it by looking up the model number.