contract-ner-model-da

This model is a fine-tuned version of xlm-roberta-base on a custom contracts dataset. It achieves the following results on the evaluation set:

Training procedure

The following hyperparameters were used during training:

Training Loss	Epoch	Step	Validation Loss	Micro F1
0.8971	0.24	200	0.0205	0.0
0.0173	0.48	400	0.0100	0.2921
0.0092	0.73	600	0.0065	0.7147
0.0063	0.97	800	0.0046	0.8332
0.0047	1.21	1000	0.0047	0.8459
0.0042	1.45	1200	0.0039	0.8694
0.0037	1.69	1400	0.0035	0.8888
0.0032	1.93	1600	0.0035	0.8840
0.0025	2.18	1800	0.0029	0.8943
0.0023	2.42	2000	0.0024	0.9104
0.0023	2.66	2200	0.0032	0.8808
0.0021	2.9	2400	0.0022	0.9338
0.0018	3.14	2600	0.0020	0.9315
0.0015	3.39	2800	0.0026	0.9297