Leveraging Pretrained Language Models for Enhanced Entity Matching: A Comprehensive Study of Fine-Tuning and Prompt Learning Paradigms

Table 7

The loss values of different methods at each training epoch on the structured iTunes-Amazon dataset.

Epochs	FT	HP	SP
1	0.663	0.563	0.624
2	0.367	0.366	0.404
3	0.123	0.156	0.225
4	0.0493	0.0683	0.0430
5	0.0464	0.0469	0.0301
6	0.0227	0.0244	7.97e − 4
7	0.0457	0.0346	5.12e − 5
8	0.0129	2.45e − 4	7.64e − 5
9	0.0106	1.01e − 4	6.42e − 5
10	1.35e − 3	2.34e − 4	6.27e − 5