xlnet-base-cased

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

• Loss: 1.6908
• Accuracy: 0.8273
• Precision: 0.8307
• Recall: 0.8273
• Precision Macro: 0.7836
• Recall Macro: 0.7606
• Macro Fpr: 0.0159
• Weighted Fpr: 0.0152
• Weighted Specificity: 0.9756
• Macro Specificity: 0.9865
• Weighted Sensitivity: 0.8218
• Macro Sensitivity: 0.7606
• F1 Micro: 0.8218
• F1 Macro: 0.7664
• F1 Weighted: 0.8189

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	Precision Macro	Recall Macro	Macro Fpr	Weighted Fpr	Weighted Specificity	Macro Specificity	Weighted Sensitivity	Macro Sensitivity	F1 Micro	F1 Macro	F1 Weighted
1.2613	1.0	643	0.7758	0.7676	0.7673	0.7676	0.5269	0.5129	0.0220	0.0212	0.9680	0.9824	0.7676	0.5129	0.7676	0.4819	0.7524
0.7364	2.0	1286	0.6755	0.8071	0.8088	0.8071	0.7425	0.6972	0.0174	0.0168	0.9751	0.9855	0.8071	0.6972	0.8071	0.7019	0.8013
0.6021	3.0	1929	0.8443	0.8064	0.8016	0.8064	0.7270	0.7262	0.0176	0.0169	0.9718	0.9852	0.8064	0.7262	0.8064	0.7229	0.8014
0.4361	4.0	2572	0.8850	0.8002	0.8001	0.8002	0.7167	0.7048	0.0180	0.0175	0.9731	0.9849	0.8002	0.7048	0.8002	0.7051	0.7971
0.3359	5.0	3215	1.1264	0.8017	0.7981	0.8017	0.6531	0.6681	0.0181	0.0174	0.9732	0.9850	0.8017	0.6681	0.8017	0.6459	0.7962
0.2827	6.0	3858	1.1471	0.7994	0.8092	0.7994	0.7389	0.6922	0.0183	0.0176	0.9686	0.9845	0.7994	0.6922	0.7994	0.7042	0.7952
0.1945	7.0	4501	1.1841	0.8149	0.8129	0.8149	0.7850	0.7598	0.0166	0.0160	0.9746	0.9860	0.8149	0.7598	0.8149	0.7667	0.8122
0.1286	8.0	5144	1.3231	0.8079	0.8105	0.8079	0.7630	0.7216	0.0171	0.0167	0.9757	0.9856	0.8079	0.7216	0.8079	0.7283	0.8067
0.1304	9.0	5787	1.3869	0.8102	0.8118	0.8102	0.7705	0.7603	0.0171	0.0165	0.9741	0.9856	0.8102	0.7603	0.8102	0.7570	0.8088
0.0875	10.0	6430	1.6901	0.7823	0.7932	0.7823	0.7601	0.7020	0.0199	0.0195	0.9680	0.9834	0.7823	0.7020	0.7823	0.7192	0.7817
0.1075	11.0	7073	1.6517	0.7978	0.8021	0.7978	0.7513	0.7567	0.0183	0.0178	0.9758	0.9849	0.7978	0.7567	0.7978	0.7470	0.7935
0.0632	12.0	7716	1.5290	0.8149	0.8184	0.8149	0.7746	0.7772	0.0167	0.0160	0.9738	0.9859	0.8149	0.7772	0.8149	0.7707	0.8150
0.0565	13.0	8359	1.5766	0.8064	0.8107	0.8064	0.7528	0.7628	0.0174	0.0169	0.9769	0.9856	0.8064	0.7628	0.8064	0.7537	0.8061
0.0504	14.0	9002	1.7548	0.8048	0.8100	0.8048	0.7569	0.7702	0.0174	0.0170	0.9765	0.9854	0.8048	0.7702	0.8048	0.7553	0.8046
0.0295	15.0	9645	1.7570	0.8102	0.8226	0.8102	0.7705	0.7611	0.0168	0.0165	0.9770	0.9858	0.8102	0.7611	0.8102	0.7610	0.8141
0.0338	16.0	10288	1.7394	0.8110	0.8138	0.8110	0.7639	0.7659	0.0168	0.0164	0.9775	0.9859	0.8110	0.7659	0.8110	0.7613	0.8100
0.0444	17.0	10931	1.7975	0.8118	0.8201	0.8118	0.7511	0.7610	0.0168	0.0163	0.9775	0.9859	0.8118	0.7610	0.8118	0.7457	0.8129
0.0397	18.0	11574	1.6921	0.8149	0.8203	0.8149	0.7540	0.7854	0.0165	0.0160	0.9780	0.9862	0.8149	0.7854	0.8149	0.7553	0.8130
0.0356	19.0	12217	1.6908	0.8273	0.8307	0.8273	0.7764	0.7992	0.0152	0.0147	0.9784	0.9870	0.8273	0.7992	0.8273	0.7814	0.8265
0.0306	20.0	12860	1.8374	0.8180	0.8208	0.8180	0.7635	0.7756	0.0162	0.0156	0.9771	0.9863	0.8180	0.7756	0.8180	0.7620	0.8166
0.0234	21.0	13503	1.7738	0.8195	0.8185	0.8195	0.7947	0.7602	0.0160	0.0155	0.9760	0.9864	0.8195	0.7602	0.8195	0.7713	0.8174
0.0091	22.0	14146	1.8537	0.8172	0.8167	0.8172	0.7732	0.7646	0.0163	0.0157	0.9764	0.9862	0.8172	0.7646	0.8172	0.7654	0.8143
0.0138	23.0	14789	1.8306	0.8102	0.8173	0.8102	0.7729	0.7569	0.0167	0.0165	0.9757	0.9857	0.8102	0.7569	0.8102	0.7625	0.8125
0.0213	24.0	15432	1.9363	0.8125	0.8149	0.8125	0.7777	0.7540	0.0168	0.0162	0.9739	0.9858	0.8125	0.7540	0.8125	0.7622	0.8115
0.0034	25.0	16075	1.9552	0.8156	0.8179	0.8156	0.7843	0.7583	0.0165	0.0159	0.9740	0.9860	0.8156	0.7583	0.8156	0.7657	0.8147
0.0028	26.0	16718	1.9404	0.8172	0.8163	0.8172	0.7884	0.7591	0.0164	0.0157	0.9747	0.9861	0.8172	0.7591	0.8172	0.7656	0.8137
0.0105	27.0	17361	1.9156	0.8180	0.8132	0.8180	0.7848	0.7575	0.0164	0.0156	0.9742	0.9861	0.8180	0.7575	0.8180	0.7667	0.8140
0.0048	28.0	18004	1.9104	0.8203	0.8196	0.8203	0.7877	0.7615	0.0160	0.0154	0.9758	0.9864	0.8203	0.7615	0.8203	0.7658	0.8175
0.0011	29.0	18647	1.9312	0.8203	0.8185	0.8203	0.7888	0.7600	0.0161	0.0154	0.9755	0.9864	0.8203	0.7600	0.8203	0.7664	0.8173
0.0004	30.0	19290	1.9234	0.8218	0.8189	0.8218	0.7836	0.7606	0.0159	0.0152	0.9756	0.9865	0.8218	0.7606	0.8218	0.7664	0.8189

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.19.0
• Tokenizers 0.15.1