metadata

license: mit
tags:
  - generated_from_trainer
model-index:
  - name: predict-perception-bert-focus-victim
    results: []

predict-perception-bert-focus-victim

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: 0.2466
Rmse: 0.6201
Rmse Focus::a Sulla vittima: 0.6201
Mae: 0.4936
Mae Focus::a Sulla vittima: 0.4936
R2: 0.7293
R2 Focus::a Sulla vittima: 0.7293
Cos: 0.8261
Pair: 0.0
Rank: 0.5
Neighbors: 0.8155
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 Focus::a Sulla vittima	Mae	Mae Focus::a Sulla vittima	R2	R2 Focus::a Sulla vittima	Cos	Rank	Neighbors	Rsa
1.0247	1.0	15	1.0286	1.2665	1.2665	1.0280	1.0280	-0.1292	-0.1292	0.1304	0.5	0.3685	nan
0.9912	2.0	30	1.0039	1.2512	1.2512	1.0347	1.0347	-0.1020	-0.1020	0.0435	0.5	0.3333	nan
0.9147	3.0	45	0.9338	1.2067	1.2067	0.9770	0.9770	-0.0251	-0.0251	0.1304	0.5	0.3685	nan
0.8194	4.0	60	0.7641	1.0916	1.0916	0.8476	0.8476	0.1612	0.1612	0.4783	0.5	0.5284	nan
0.6636	5.0	75	0.6618	1.0159	1.0159	0.8012	0.8012	0.2735	0.2735	0.6522	0.5	0.4741	nan
0.523	6.0	90	0.5176	0.8984	0.8984	0.7044	0.7044	0.4318	0.4318	0.6522	0.5	0.4741	nan
0.402	7.0	105	0.3804	0.7702	0.7702	0.6042	0.6042	0.5824	0.5824	0.6522	0.5	0.5395	nan
0.3401	8.0	120	0.3594	0.7487	0.7487	0.5703	0.5703	0.6054	0.6054	0.7391	0.5	0.6920	nan
0.2615	9.0	135	0.3429	0.7312	0.7312	0.6049	0.6049	0.6236	0.6236	0.7391	0.5	0.6920	nan
0.1928	10.0	150	0.2889	0.6712	0.6712	0.5487	0.5487	0.6828	0.6828	0.7391	0.5	0.6920	nan
0.1703	11.0	165	0.2675	0.6458	0.6458	0.5188	0.5188	0.7064	0.7064	0.7391	0.5	0.6920	nan
0.1209	12.0	180	0.2826	0.6639	0.6639	0.5475	0.5475	0.6897	0.6897	0.7391	0.5	0.6920	nan
0.1428	13.0	195	0.2978	0.6815	0.6815	0.5777	0.5777	0.6731	0.6731	0.7391	0.5	0.6920	nan
0.1038	14.0	210	0.2924	0.6753	0.6753	0.5865	0.5865	0.6790	0.6790	0.6522	0.5	0.2760	nan
0.0951	15.0	225	0.2905	0.6731	0.6731	0.5750	0.5750	0.6811	0.6811	0.7391	0.5	0.6920	nan
0.0809	16.0	240	0.2676	0.6460	0.6460	0.5552	0.5552	0.7062	0.7062	0.7391	0.5	0.6920	nan
0.0811	17.0	255	0.2770	0.6572	0.6572	0.5543	0.5543	0.6959	0.6959	0.7391	0.5	0.6920	nan
0.0703	18.0	270	0.2634	0.6409	0.6409	0.5251	0.5251	0.7108	0.7108	0.8261	0.5	0.8155	nan
0.0595	19.0	285	0.2638	0.6413	0.6413	0.5196	0.5196	0.7104	0.7104	0.8261	0.5	0.8155	nan
0.0651	20.0	300	0.2520	0.6268	0.6268	0.4970	0.4970	0.7234	0.7234	0.8261	0.5	0.8155	nan
0.0637	21.0	315	0.2668	0.6451	0.6451	0.4965	0.4965	0.7071	0.7071	0.8261	0.5	0.8155	nan
0.0582	22.0	330	0.2455	0.6188	0.6188	0.4759	0.4759	0.7305	0.7305	0.8261	0.5	0.8155	nan
0.0616	23.0	345	0.2509	0.6255	0.6255	0.5084	0.5084	0.7246	0.7246	0.8261	0.5	0.8155	nan
0.0492	24.0	360	0.2510	0.6256	0.6256	0.4985	0.4985	0.7244	0.7244	0.8261	0.5	0.8155	nan
0.0504	25.0	375	0.2512	0.6259	0.6259	0.4849	0.4849	0.7242	0.7242	0.8261	0.5	0.8155	nan
0.0501	26.0	390	0.2585	0.6350	0.6350	0.5140	0.5140	0.7162	0.7162	0.8261	0.5	0.8155	nan
0.0411	27.0	405	0.2544	0.6299	0.6299	0.5148	0.5148	0.7207	0.7207	0.8261	0.5	0.8155	nan
0.044	28.0	420	0.2466	0.6201	0.6201	0.4964	0.4964	0.7293	0.7293	0.8261	0.5	0.8155	nan
0.042	29.0	435	0.2466	0.6201	0.6201	0.4836	0.4836	0.7293	0.7293	0.8261	0.5	0.8155	nan
0.0446	30.0	450	0.2466	0.6201	0.6201	0.4936	0.4936	0.7293	0.7293	0.8261	0.5	0.8155	nan

Framework versions

Transformers 4.16.2
Pytorch 1.10.2+cu113
Datasets 1.18.3
Tokenizers 0.11.0