Illustrate performance gain of each pretrained models & how each models got lighter

[snoop2head]

For each pretrained models, display

• wandb log
• table which consists of maximum f1 score in 100 epochs and FLOPS(either inference time)
• percentage gain in both metrics of f1 score and FLOPS(either inference time)

[lkm2835]

It seems important which device(CPU or GPU, v100 or smaller) will measure the inference time.

[snoop2head]

@lkm2835
I am thinking of Colab basic as a barometer for computing power.
According to stackoverflow post, Colab's computing power is

• GPU: Teska K80
• CPU: 1xsingle core hyper threaded Xeon Processors @2.3Ghz i.e(1 core, 2 threads)

But torch's version compatibility might cause a problem.
Affixing torch and torchvision version to colab's version is one of the candidate I am considering on.