hubert-classifier-aug-fold-0

This model is a fine-tuned version of facebook/hubert-base-ls960 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5873
• Accuracy: 0.8787
• Precision: 0.8925
• Recall: 0.8787
• F1: 0.8784
• Binary: 0.9162

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Binary
No log	0.24	50	4.4206	0.0195	0.0007	0.0195	0.0014	0.1390
No log	0.48	100	4.3006	0.0442	0.0114	0.0442	0.0127	0.2528
No log	0.72	150	3.9867	0.0472	0.0033	0.0472	0.0061	0.3276
No log	0.96	200	3.6925	0.0712	0.0116	0.0712	0.0180	0.3447
4.2438	1.2	250	3.4305	0.0854	0.0508	0.0854	0.0319	0.3580
4.2438	1.44	300	3.2405	0.1071	0.0689	0.1071	0.0432	0.3730
4.2438	1.68	350	3.0535	0.1491	0.1053	0.1491	0.0823	0.3999
4.2438	1.92	400	2.7897	0.2419	0.2020	0.2419	0.1678	0.4667
3.3411	2.16	450	2.4987	0.3303	0.2416	0.3303	0.2457	0.5288
3.3411	2.4	500	2.1588	0.4779	0.3998	0.4779	0.4078	0.6354
3.3411	2.63	550	1.8909	0.5273	0.4768	0.5273	0.4604	0.6688
3.3411	2.87	600	1.6458	0.5708	0.5612	0.5708	0.5191	0.6994
2.4102	3.11	650	1.4630	0.6187	0.6002	0.6187	0.5757	0.7327
2.4102	3.35	700	1.2770	0.6764	0.6582	0.6764	0.6409	0.7730
2.4102	3.59	750	1.1875	0.6966	0.6830	0.6966	0.6696	0.7884
2.4102	3.83	800	1.0563	0.7228	0.7372	0.7228	0.7012	0.8073
1.6409	4.07	850	0.9471	0.7506	0.7688	0.7506	0.7322	0.8260
1.6409	4.31	900	0.9012	0.7588	0.7677	0.7588	0.7471	0.8313
1.6409	4.55	950	0.8540	0.7768	0.8025	0.7768	0.7685	0.8435
1.6409	4.79	1000	0.7910	0.7828	0.7915	0.7828	0.7723	0.8479
1.2621	5.03	1050	0.7229	0.7918	0.7952	0.7918	0.7804	0.8542
1.2621	5.27	1100	0.7388	0.8067	0.8250	0.8067	0.8031	0.8650
1.2621	5.51	1150	0.7315	0.8090	0.8298	0.8090	0.8029	0.8672
1.2621	5.75	1200	0.7357	0.7903	0.8053	0.7903	0.7856	0.8533
1.2621	5.99	1250	0.7088	0.8090	0.8240	0.8090	0.8037	0.8672
1.0138	6.23	1300	0.6828	0.8112	0.8209	0.8112	0.8077	0.8684
1.0138	6.47	1350	0.7561	0.8082	0.8229	0.8082	0.8032	0.8678
1.0138	6.71	1400	0.6640	0.8292	0.8415	0.8292	0.8250	0.8812
1.0138	6.95	1450	0.6330	0.8315	0.8453	0.8315	0.8282	0.8828
0.9058	7.19	1500	0.6482	0.8217	0.8331	0.8217	0.8189	0.8764
0.9058	7.43	1550	0.7005	0.8187	0.8330	0.8187	0.8135	0.8736
0.9058	7.66	1600	0.5902	0.8562	0.8645	0.8562	0.8533	0.8998
0.9058	7.9	1650	0.5481	0.8607	0.8723	0.8607	0.8594	0.9019
0.7905	8.14	1700	0.6131	0.8427	0.8534	0.8427	0.8394	0.8899
0.7905	8.38	1750	0.6664	0.8419	0.8541	0.8419	0.8394	0.8897
0.7905	8.62	1800	0.6453	0.8330	0.8473	0.8330	0.8293	0.8842
0.7905	8.86	1850	0.6178	0.8390	0.8553	0.8390	0.8362	0.8873
0.7208	9.1	1900	0.6779	0.8412	0.8540	0.8412	0.8379	0.8895
0.7208	9.34	1950	0.5752	0.8607	0.8690	0.8607	0.8581	0.9031
0.7208	9.58	2000	0.6717	0.8434	0.8544	0.8434	0.8408	0.8909
0.7208	9.82	2050	0.6790	0.8345	0.8500	0.8345	0.8321	0.8848
0.6476	10.06	2100	0.6429	0.8494	0.8631	0.8494	0.8472	0.8954
0.6476	10.3	2150	0.6006	0.8577	0.8668	0.8577	0.8558	0.9007
0.6476	10.54	2200	0.5987	0.8532	0.8634	0.8532	0.8519	0.8974
0.6476	10.78	2250	0.6524	0.8472	0.8594	0.8472	0.8443	0.8934
0.6156	11.02	2300	0.6748	0.8412	0.8529	0.8412	0.8386	0.8904
0.6156	11.26	2350	0.5571	0.8577	0.8644	0.8577	0.8547	0.9011
0.6156	11.5	2400	0.6081	0.8502	0.8607	0.8502	0.8468	0.8959
0.6156	11.74	2450	0.5866	0.8592	0.8692	0.8592	0.8575	0.9022
0.6156	11.98	2500	0.6205	0.8517	0.8630	0.8517	0.8501	0.8966
0.5738	12.22	2550	0.6544	0.8562	0.8704	0.8562	0.8549	0.8996
0.5738	12.46	2600	0.6792	0.8427	0.8545	0.8427	0.8385	0.8906
0.5738	12.69	2650	0.6009	0.8569	0.8676	0.8569	0.8557	0.9008
0.5738	12.93	2700	0.6580	0.8524	0.8621	0.8524	0.8490	0.8972
0.5416	13.17	2750	0.6781	0.8532	0.8639	0.8532	0.8504	0.8977
0.5416	13.41	2800	0.5903	0.8659	0.8749	0.8659	0.8646	0.9084
0.5416	13.65	2850	0.5766	0.8644	0.8728	0.8644	0.8620	0.9064
0.5416	13.89	2900	0.6674	0.8592	0.8688	0.8592	0.8565	0.9027
0.5213	14.13	2950	0.6256	0.8652	0.8751	0.8652	0.8635	0.9067
0.5213	14.37	3000	0.6518	0.8622	0.8704	0.8622	0.8602	0.9051
0.5213	14.61	3050	0.6694	0.8547	0.8661	0.8547	0.8531	0.8999
0.5213	14.85	3100	0.6153	0.8719	0.8799	0.8719	0.8710	0.9125
0.4856	15.09	3150	0.6067	0.8727	0.8821	0.8727	0.8715	0.9106
0.4856	15.33	3200	0.6354	0.8592	0.8712	0.8592	0.8581	0.9019
0.4856	15.57	3250	0.6773	0.8532	0.8623	0.8532	0.8507	0.8988
0.4856	15.81	3300	0.6356	0.8682	0.8759	0.8682	0.8660	0.9088
0.4631	16.05	3350	0.6139	0.8712	0.8783	0.8712	0.8700	0.9102
0.4631	16.29	3400	0.6589	0.8622	0.8730	0.8622	0.8612	0.9049
0.4631	16.53	3450	0.6439	0.8539	0.8660	0.8539	0.8516	0.8982
0.4631	16.77	3500	0.6727	0.8689	0.8757	0.8689	0.8673	0.9091
0.4605	17.01	3550	0.6359	0.8712	0.8793	0.8712	0.8703	0.9103
0.4605	17.25	3600	0.6926	0.8547	0.8647	0.8547	0.8534	0.8999
0.4605	17.49	3650	0.6937	0.8562	0.8687	0.8562	0.8544	0.9008
0.4605	17.72	3700	0.6625	0.8659	0.8777	0.8659	0.8649	0.9068
0.4605	17.96	3750	0.6542	0.8674	0.8784	0.8674	0.8655	0.9090
0.4371	18.2	3800	0.5719	0.8742	0.8831	0.8742	0.8727	0.9121
0.4371	18.44	3850	0.6245	0.8734	0.8811	0.8734	0.8727	0.9124
0.4371	18.68	3900	0.6993	0.8577	0.8680	0.8577	0.8559	0.9018
0.4371	18.92	3950	0.6896	0.8592	0.8681	0.8592	0.8573	0.9028
0.4277	19.16	4000	0.6869	0.8517	0.8640	0.8517	0.8507	0.8973
0.4277	19.4	4050	0.6963	0.8599	0.8692	0.8599	0.8587	0.9021
0.4277	19.64	4100	0.5527	0.8831	0.8898	0.8831	0.8819	0.9184
0.4277	19.88	4150	0.6925	0.8592	0.8699	0.8592	0.8580	0.9025
0.401	20.12	4200	0.6998	0.8592	0.8719	0.8592	0.8582	0.9040
0.401	20.36	4250	0.6390	0.8757	0.8849	0.8757	0.8743	0.9139
0.401	20.6	4300	0.6792	0.8659	0.8762	0.8659	0.8641	0.9075
0.401	20.84	4350	0.6946	0.8554	0.8662	0.8554	0.8529	0.8990
0.3945	21.08	4400	0.8223	0.8427	0.8559	0.8427	0.8409	0.8903
0.3945	21.32	4450	0.7841	0.8622	0.8710	0.8622	0.8599	0.9040
0.3945	21.56	4500	0.6545	0.8697	0.8766	0.8697	0.8687	0.9093
0.3945	21.8	4550	0.7135	0.8652	0.8710	0.8652	0.8630	0.9072
0.3829	22.04	4600	0.6901	0.8622	0.8705	0.8622	0.8610	0.9046
0.3829	22.28	4650	0.6960	0.8599	0.8688	0.8599	0.8579	0.9035
0.3829	22.51	4700	0.7047	0.8644	0.8752	0.8644	0.8630	0.9061
0.3829	22.75	4750	0.6855	0.8674	0.8784	0.8674	0.8662	0.9094
0.3829	22.99	4800	0.7315	0.8539	0.8652	0.8539	0.8516	0.8993
0.3695	23.23	4850	0.7299	0.8569	0.8663	0.8569	0.8545	0.9005

Framework versions

• Transformers 4.38.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1