predict-perception-bert-blame-assassin

This model is a fine-tuned version of dbmdz/bert-base-italian-xxl-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5128
• Rmse: 1.0287
• Rmse Blame::a L'assassino: 1.0287
• Mae: 0.8883
• Mae Blame::a L'assassino: 0.8883
• R2: 0.5883
• R2 Blame::a L'assassino: 0.5883
• Cos: 0.6522
• Pair: 0.0
• Rank: 0.5
• Neighbors: 0.5795
• Rsa: nan

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 20
• eval_batch_size: 8
• seed: 1996
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Rmse	Rmse Blame::a L'assassino	Mae	Mae Blame::a L'assassino	R2	R2 Blame::a L'assassino	Cos	Pair	Rank	Neighbors	Rsa
1.0184	1.0	15	1.2219	1.5879	1.5879	1.4308	1.4308	0.0191	0.0191	0.3913	0.0	0.5	0.3781	nan
0.9214	2.0	30	1.0927	1.5017	1.5017	1.3634	1.3634	0.1227	0.1227	0.5652	0.0	0.5	0.4512	nan
0.7809	3.0	45	0.8206	1.3013	1.3013	1.1808	1.1808	0.3412	0.3412	0.4783	0.0	0.5	0.3819	nan
0.6593	4.0	60	0.5894	1.1029	1.1029	1.0145	1.0145	0.5268	0.5268	0.7391	0.0	0.5	0.6408	nan
0.4672	5.0	75	0.4759	0.9910	0.9910	0.8868	0.8868	0.6180	0.6180	0.7391	0.0	0.5	0.4884	nan
0.3356	6.0	90	0.4220	0.9332	0.9332	0.8083	0.8083	0.6612	0.6612	0.6522	0.0	0.5	0.4249	nan
0.2782	7.0	105	0.4477	0.9612	0.9612	0.8046	0.8046	0.6406	0.6406	0.6522	0.0	0.5	0.6101	nan
0.2075	8.0	120	0.4389	0.9518	0.9518	0.8050	0.8050	0.6476	0.6476	0.6522	0.0	0.5	0.5795	nan
0.1725	9.0	135	0.4832	0.9985	0.9985	0.8356	0.8356	0.6121	0.6121	0.7391	0.0	0.5	0.6616	nan
0.1642	10.0	150	0.4368	0.9494	0.9494	0.8060	0.8060	0.6493	0.6493	0.6522	0.0	0.5	0.5795	nan
0.1172	11.0	165	0.4538	0.9677	0.9677	0.8174	0.8174	0.6357	0.6357	0.7391	0.0	0.5	0.4884	nan
0.104	12.0	180	0.4672	0.9819	0.9819	0.8384	0.8384	0.6249	0.6249	0.7391	0.0	0.5	0.4884	nan
0.0822	13.0	195	0.4401	0.9530	0.9530	0.8107	0.8107	0.6467	0.6467	0.7391	0.0	0.5	0.4884	nan
0.0755	14.0	210	0.4464	0.9598	0.9598	0.8251	0.8251	0.6416	0.6416	0.7391	0.0	0.5	0.4884	nan
0.0801	15.0	225	0.4834	0.9988	0.9988	0.8604	0.8604	0.6119	0.6119	0.7391	0.0	0.5	0.4884	nan
0.053	16.0	240	0.4846	1.0001	1.0001	0.8651	0.8651	0.6109	0.6109	0.7391	0.0	0.5	0.4884	nan
0.0573	17.0	255	0.4970	1.0128	1.0128	0.8743	0.8743	0.6010	0.6010	0.7391	0.0	0.5	0.4884	nan
0.0571	18.0	270	0.4803	0.9956	0.9956	0.8503	0.8503	0.6144	0.6144	0.6522	0.0	0.5	0.5795	nan
0.0483	19.0	285	0.4936	1.0093	1.0093	0.8740	0.8740	0.6037	0.6037	0.6522	0.0	0.5	0.5795	nan
0.0414	20.0	300	0.5138	1.0297	1.0297	0.8943	0.8943	0.5875	0.5875	0.6522	0.0	0.5	0.5795	nan
0.0513	21.0	315	0.5240	1.0399	1.0399	0.9050	0.9050	0.5793	0.5793	0.7391	0.0	0.5	0.4884	nan
0.0499	22.0	330	0.5275	1.0434	1.0434	0.9048	0.9048	0.5765	0.5765	0.7391	0.0	0.5	0.4884	nan
0.0423	23.0	345	0.5350	1.0508	1.0508	0.8872	0.8872	0.5705	0.5705	0.6522	0.0	0.5	0.5795	nan
0.0447	24.0	360	0.4963	1.0120	1.0120	0.8754	0.8754	0.6016	0.6016	0.7391	0.0	0.5	0.4884	nan
0.0364	25.0	375	0.5009	1.0167	1.0167	0.8809	0.8809	0.5979	0.5979	0.6522	0.0	0.5	0.5795	nan
0.0412	26.0	390	0.5060	1.0219	1.0219	0.8781	0.8781	0.5938	0.5938	0.6522	0.0	0.5	0.5795	nan
0.0297	27.0	405	0.5027	1.0185	1.0185	0.8838	0.8838	0.5964	0.5964	0.7391	0.0	0.5	0.4884	nan
0.0416	28.0	420	0.5071	1.0230	1.0230	0.8867	0.8867	0.5929	0.5929	0.7391	0.0	0.5	0.4884	nan
0.0327	29.0	435	0.5124	1.0283	1.0283	0.8883	0.8883	0.5887	0.5887	0.6522	0.0	0.5	0.5795	nan
0.0383	30.0	450	0.5128	1.0287	1.0287	0.8883	0.8883	0.5883	0.5883	0.6522	0.0	0.5	0.5795	nan

Framework versions

• Transformers 4.16.2
• Pytorch 1.10.2+cu113
• Datasets 1.18.3
• Tokenizers 0.11.0