How much time facebook/wav2vec2-xls-r-300m model will take to train on 311919 size of dataset?

I want to train model for speech to text in tamil language where my dataset size is of around 311919 rows.

Split structure :
train - > 224581
test - > 31192
validation - > 56146

GPU details :
Name - Quadro P2000
size - 5 GB
Cuda cores - 1024

I am asking this question to schedule my training. thank you