pgajo/mbert_EW-TT-PE_U0_S1_DROP1_mbert_E10_DEV72.0

Model description:

Model: bert-base-multilingual-cased

Dataset: TASTEset

Unshuffled ratio: ['0']

Shuffled ratio: ['1']

Best exact match epoch: 10

Best exact match: 72.1

Best epoch: 10

Drop duplicates: ['1']

Max epochs = 10

Optimizer lr = 3e-05

Optimizer eps = 1e-08

Batch size = 32

Dataset path = pgajo/EW-TT-PE_U0_S1_DROP1_mbert

Results

epoch	train_loss	train_f1	train_exact	dev_loss	dev_f1	dev_exact	test_loss	test_f1	test_exact
1	3.35	8.36	0.76	2.78	13.81	3.04	0	0	0
2	2.25	27.1	16.59	1.57	55.74	46.96	0	0	0
3	1.14	64.77	55.56	1.35	66.21	59.12	0	0	0
4	0.62	79.58	73.88	1.19	68.45	62.15	0	0	0
5	0.36	87.7	84.73	1.4	72.12	66.3	0	0	0
6	0.23	92.1	88.94	1.2	74.8	70.17	0	0	0
7	0.16	94.35	92.47	1.3	74.04	67.13	0	0	0
8	0.12	95.37	94.75	1.34	76.44	69.61	0	0	0
9	0.08	96.92	95.99	1.33	77.65	71.55	0	0	0
10	0.08	97.02	96.13	1.57	77.43	72.1	0	0	0