Table 1.

DistilRoBERTa fine-tuning training metrics. The model obtained optimal fine-tuning after 2 training epochs.

Step	Epoch	Training loss	Validation loss	Precision	Accuracy	F1-score
500	0.4	0.5903	0.4695	0.7342	0.7728	0.7890
1000	0.8	0.3986	0.3469	0.8144	0.8596	0.8684
1500	1.2	0.2366	0.1939	0.9313	0.9260	0.9253
2000	1.6	0.1476	0.1560	0.9207	0.9452	0.9465
2500	2.0	0.1284	0.1167	0.9561	0.9592	0.9592