Model Training Calculator

Define your transformer model, its training data, and enter your preferred training parameters. The calculator will help you understand if your model will fit in memory and support you while you tweak the training parameters to make it fit. Finally the calculator will provide an estimate of how long training will take.


¡GPUs do not have enough memory to train this model!

¡Dataset to parameter ratio may not be compute optimised!

Time to train: 2.08×1032.08 \times 10^{3} hours

Memory Required per GPU: 25925.29%


Model Details

Number of Model Parameters

Compute Details


Training Details


Explanation