predict-perception-xlmr-cause-concept

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: 0.3933
• Rmse: 0.5992
• Rmse Cause::a Causata da un concetto astratto (es. gelosia): 0.5992
• Mae: 0.4566
• Mae Cause::a Causata da un concetto astratto (es. gelosia): 0.4566
• R2: 0.5588
• R2 Cause::a Causata da un concetto astratto (es. gelosia): 0.5588
• Cos: 0.3043
• Pair: 0.0
• Rank: 0.5
• Neighbors: 0.4340
• 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 Cause::a Causata da un concetto astratto (es. gelosia)	Mae	Mae Cause::a Causata da un concetto astratto (es. gelosia)	R2	R2 Cause::a Causata da un concetto astratto (es. gelosia)	Cos	Pair	Rank	Neighbors	Rsa
1.0114	1.0	15	0.9088	0.9109	0.9109	0.6455	0.6455	-0.0195	-0.0195	-0.0435	0.0	0.5	0.4027	nan
1.0	2.0	30	0.8833	0.8980	0.8980	0.6104	0.6104	0.0090	0.0090	0.2174	0.0	0.5	0.3681	nan
0.9533	3.0	45	0.8453	0.8785	0.8785	0.6072	0.6072	0.0517	0.0517	0.1304	0.0	0.5	0.3748	nan
0.9113	4.0	60	0.7797	0.8437	0.8437	0.6024	0.6024	0.1253	0.1253	0.0435	0.0	0.5	0.3028	nan
0.8312	5.0	75	0.5756	0.7249	0.7249	0.5128	0.5128	0.3542	0.3542	0.4783	0.0	0.5	0.4572	nan
0.7224	6.0	90	0.4977	0.6741	0.6741	0.5114	0.5114	0.4416	0.4416	0.2174	0.0	0.5	0.4009	nan
0.5789	7.0	105	0.6338	0.7607	0.7607	0.5059	0.5059	0.2889	0.2889	0.3043	0.0	0.5	0.4340	nan
0.4978	8.0	120	0.3342	0.5524	0.5524	0.4298	0.4298	0.6250	0.6250	0.2174	0.0	0.5	0.4274	nan
0.4572	9.0	135	0.3210	0.5413	0.5413	0.4343	0.4343	0.6399	0.6399	0.3043	0.0	0.5	0.4340	nan
0.3346	10.0	150	0.3456	0.5617	0.5617	0.4198	0.4198	0.6123	0.6123	0.3043	0.0	0.5	0.4340	nan
0.3046	11.0	165	0.3840	0.5921	0.5921	0.4312	0.4312	0.5692	0.5692	0.3043	0.0	0.5	0.4340	nan
0.3035	12.0	180	0.3929	0.5989	0.5989	0.4147	0.4147	0.5592	0.5592	0.3043	0.0	0.5	0.4340	nan
0.2199	13.0	195	0.3165	0.5376	0.5376	0.4065	0.4065	0.6449	0.6449	0.3043	0.0	0.5	0.4340	nan
0.2376	14.0	210	0.3108	0.5326	0.5326	0.3937	0.3937	0.6514	0.6514	0.3913	0.0	0.5	0.4286	nan
0.1639	15.0	225	0.3645	0.5769	0.5769	0.4094	0.4094	0.5911	0.5911	0.3913	0.0	0.5	0.4286	nan
0.1884	16.0	240	0.3762	0.5860	0.5860	0.4398	0.4398	0.5779	0.5779	0.3043	0.0	0.5	0.4340	nan
0.1767	17.0	255	0.3805	0.5894	0.5894	0.4540	0.4540	0.5732	0.5732	0.2174	0.0	0.5	0.4298	nan
0.1329	18.0	270	0.3555	0.5697	0.5697	0.4281	0.4281	0.6011	0.6011	0.2174	0.0	0.5	0.4298	nan
0.1834	19.0	285	0.4337	0.6292	0.6292	0.4402	0.4402	0.5135	0.5135	0.3913	0.0	0.5	0.4286	nan
0.1538	20.0	300	0.3554	0.5696	0.5696	0.4236	0.4236	0.6013	0.6013	0.3043	0.0	0.5	0.4340	nan
0.1459	21.0	315	0.3592	0.5726	0.5726	0.4348	0.4348	0.5971	0.5971	0.3043	0.0	0.5	0.4066	nan
0.1038	22.0	330	0.3732	0.5837	0.5837	0.4382	0.4382	0.5813	0.5813	0.3913	0.0	0.5	0.4664	nan
0.1432	23.0	345	0.3635	0.5760	0.5760	0.4394	0.4394	0.5922	0.5922	0.3913	0.0	0.5	0.4664	nan
0.1354	24.0	360	0.4359	0.6308	0.6308	0.4793	0.4793	0.5110	0.5110	0.3043	0.0	0.5	0.4340	nan
0.1404	25.0	375	0.3919	0.5982	0.5982	0.4650	0.4650	0.5603	0.5603	0.3913	0.0	0.5	0.4664	nan
0.103	26.0	390	0.4223	0.6209	0.6209	0.4691	0.4691	0.5263	0.5263	0.3043	0.0	0.5	0.4340	nan
0.1733	27.0	405	0.3972	0.6021	0.6021	0.4591	0.4591	0.5544	0.5544	0.3043	0.0	0.5	0.4340	nan
0.1019	28.0	420	0.3958	0.6011	0.6011	0.4593	0.4593	0.5559	0.5559	0.3043	0.0	0.5	0.4340	nan
0.1076	29.0	435	0.4015	0.6054	0.6054	0.4589	0.4589	0.5496	0.5496	0.3043	0.0	0.5	0.4340	nan
0.0999	30.0	450	0.3933	0.5992	0.5992	0.4566	0.4566	0.5588	0.5588	0.3043	0.0	0.5	0.4340	nan

Framework versions

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