predict-perception-xlmr-focus-object

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

• Loss: 0.1927
• Rmse: 0.5495
• Rmse Focus::a Su un oggetto: 0.5495
• Mae: 0.4174
• Mae Focus::a Su un oggetto: 0.4174
• R2: 0.5721
• R2 Focus::a Su un oggetto: 0.5721
• Cos: 0.5652
• Pair: 0.0
• Rank: 0.5
• Neighbors: 0.5518
• 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 Su un oggetto	Mae	Mae Focus::a Su un oggetto	R2	R2 Focus::a Su un oggetto	Cos	Pair	Rank	Neighbors	Rsa
1.0316	1.0	15	0.6428	1.0035	1.0035	0.8806	0.8806	-0.4272	-0.4272	-0.4783	0.0	0.5	0.5302	nan
1.0005	2.0	30	0.4564	0.8456	0.8456	0.7078	0.7078	-0.0134	-0.0134	0.4783	0.0	0.5	0.4440	nan
0.9519	3.0	45	0.4151	0.8063	0.8063	0.6797	0.6797	0.0784	0.0784	0.1304	0.0	0.5	0.4888	nan
0.92	4.0	60	0.3982	0.7898	0.7898	0.6516	0.6516	0.1159	0.1159	0.2174	0.0	0.5	0.5036	nan
0.8454	5.0	75	0.2739	0.6550	0.6550	0.5292	0.5292	0.3919	0.3919	0.6522	0.0	0.5	0.4160	nan
0.7247	6.0	90	0.2413	0.6148	0.6148	0.5347	0.5347	0.4642	0.4642	0.4783	0.0	0.5	0.3453	nan
0.6055	7.0	105	0.3109	0.6978	0.6978	0.6115	0.6115	0.3098	0.3098	0.4783	0.0	0.5	0.4154	nan
0.5411	8.0	120	0.3932	0.7848	0.7848	0.6712	0.6712	0.1271	0.1271	0.4783	0.0	0.5	0.4154	nan
0.4784	9.0	135	0.1316	0.4540	0.4540	0.3750	0.3750	0.7079	0.7079	0.5652	0.0	0.5	0.6247	nan
0.4039	10.0	150	0.2219	0.5896	0.5896	0.4954	0.4954	0.5074	0.5074	0.5652	0.0	0.5	0.4838	nan
0.3415	11.0	165	0.1935	0.5505	0.5505	0.4443	0.4443	0.5704	0.5704	0.5652	0.0	0.5	0.6247	nan
0.3369	12.0	180	0.2118	0.5761	0.5761	0.4554	0.4554	0.5296	0.5296	0.5652	0.0	0.5	0.6247	nan
0.3083	13.0	195	0.1928	0.5496	0.5496	0.4368	0.4368	0.5718	0.5718	0.5652	0.0	0.5	0.6247	nan
0.2678	14.0	210	0.2205	0.5877	0.5877	0.4472	0.4472	0.5105	0.5105	0.5652	0.0	0.5	0.6247	nan
0.2199	15.0	225	0.2118	0.5760	0.5760	0.4689	0.4689	0.5297	0.5297	0.5652	0.0	0.5	0.6247	nan
0.2238	16.0	240	0.2461	0.6209	0.6209	0.5047	0.5047	0.4537	0.4537	0.5652	0.0	0.5	0.6247	nan
0.2233	17.0	255	0.2307	0.6011	0.6011	0.4618	0.4618	0.4879	0.4879	0.5652	0.0	0.5	0.6247	nan
0.1903	18.0	270	0.2207	0.5880	0.5880	0.4432	0.4432	0.5100	0.5100	0.6522	0.0	0.5	0.6622	nan
0.1714	19.0	285	0.2146	0.5798	0.5798	0.4368	0.4368	0.5236	0.5236	0.5652	0.0	0.5	0.6247	nan
0.1759	20.0	300	0.1745	0.5228	0.5228	0.4152	0.4152	0.6126	0.6126	0.5652	0.0	0.5	0.6247	nan
0.1505	21.0	315	0.1944	0.5519	0.5519	0.4170	0.4170	0.5684	0.5684	0.5652	0.0	0.5	0.6247	nan
0.1467	22.0	330	0.1802	0.5313	0.5313	0.3910	0.3910	0.5999	0.5999	0.6522	0.0	0.5	0.6622	nan
0.1441	23.0	345	0.2360	0.6081	0.6081	0.4755	0.4755	0.4760	0.4760	0.4783	0.0	0.5	0.4938	nan
0.1553	24.0	360	0.2129	0.5774	0.5774	0.4539	0.4539	0.5274	0.5274	0.5652	0.0	0.5	0.5518	nan
0.1163	25.0	375	0.1780	0.5281	0.5281	0.3952	0.3952	0.6048	0.6048	0.6522	0.0	0.5	0.6622	nan
0.1266	26.0	390	0.2163	0.5821	0.5821	0.4569	0.4569	0.5198	0.5198	0.5652	0.0	0.5	0.5518	nan
0.1416	27.0	405	0.1829	0.5352	0.5352	0.4082	0.4082	0.5939	0.5939	0.5652	0.0	0.5	0.5518	nan
0.1576	28.0	420	0.1930	0.5498	0.5498	0.4126	0.4126	0.5716	0.5716	0.6522	0.0	0.5	0.6622	nan
0.118	29.0	435	0.2070	0.5694	0.5694	0.4378	0.4378	0.5405	0.5405	0.5652	0.0	0.5	0.5518	nan
0.1179	30.0	450	0.1927	0.5495	0.5495	0.4174	0.4174	0.5721	0.5721	0.5652	0.0	0.5	0.5518	nan

Framework versions

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