fydhfzh/hubert-classifier-aug-fold-3-old

hubert-classifier-aug-fold-3

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: 0.4622
• Accuracy: 0.8962
• Precision: 0.9090
• Recall: 0.8962
• F1: 0.8963
• Binary: 0.9276

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
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Binary
No log	0.22	50	3.8684	0.0648	0.0195	0.0648	0.0184	0.3348
No log	0.43	100	3.4085	0.0891	0.0169	0.0891	0.0238	0.3588
No log	0.65	150	3.1493	0.1107	0.0376	0.1107	0.0422	0.3746
No log	0.86	200	2.8408	0.1822	0.1026	0.1822	0.1043	0.4273
3.6116	1.08	250	2.6147	0.2780	0.1840	0.2780	0.1853	0.4906
3.6116	1.29	300	2.4272	0.2982	0.2217	0.2982	0.2156	0.5023
3.6116	1.51	350	2.1582	0.4103	0.3356	0.4103	0.3247	0.5852
3.6116	1.72	400	2.0102	0.4143	0.3454	0.4143	0.3447	0.5892
3.6116	1.94	450	1.8796	0.4615	0.4433	0.4615	0.4066	0.6225
2.6485	2.16	500	1.6625	0.5412	0.5251	0.5412	0.4856	0.6771
2.6485	2.37	550	1.5422	0.5843	0.5780	0.5843	0.5307	0.7078
2.6485	2.59	600	1.4263	0.6073	0.5806	0.6073	0.5573	0.7246
2.6485	2.8	650	1.2985	0.6451	0.6244	0.6451	0.6039	0.7503
2.1146	3.02	700	1.2788	0.6613	0.6564	0.6613	0.6273	0.7614
2.1146	3.23	750	1.2186	0.6802	0.6820	0.6802	0.6499	0.7742
2.1146	3.45	800	1.1269	0.7152	0.7428	0.7152	0.6978	0.7991
2.1146	3.66	850	1.1179	0.6680	0.6970	0.6680	0.6377	0.7675
2.1146	3.88	900	1.0928	0.7031	0.7264	0.7031	0.6793	0.7889
1.8074	4.09	950	0.9427	0.7638	0.7780	0.7638	0.7513	0.8350
1.8074	4.31	1000	0.8876	0.7692	0.7952	0.7692	0.7603	0.8382
1.8074	4.53	1050	0.8686	0.7773	0.7852	0.7773	0.7664	0.8425
1.8074	4.74	1100	0.8814	0.7665	0.7770	0.7665	0.7497	0.8363
1.8074	4.96	1150	0.8280	0.7706	0.7896	0.7706	0.7603	0.8401
1.5857	5.17	1200	0.8050	0.7773	0.8023	0.7773	0.7720	0.8439
1.5857	5.39	1250	0.7475	0.8016	0.8114	0.8016	0.7976	0.8609
1.5857	5.6	1300	0.7396	0.7895	0.8187	0.7895	0.7859	0.8521
1.5857	5.82	1350	0.7637	0.8030	0.8177	0.8030	0.7953	0.8598
1.4411	6.03	1400	0.7511	0.7976	0.8157	0.7976	0.7934	0.8574
1.4411	6.25	1450	0.6479	0.8232	0.8392	0.8232	0.8185	0.8756
1.4411	6.47	1500	0.6521	0.8286	0.8494	0.8286	0.8233	0.8803
1.4411	6.68	1550	0.5778	0.8529	0.8637	0.8529	0.8501	0.8962
1.4411	6.9	1600	0.5898	0.8259	0.8428	0.8259	0.8249	0.8776
1.3162	7.11	1650	0.5784	0.8421	0.8614	0.8421	0.8404	0.8892
1.3162	7.33	1700	0.6395	0.8232	0.8407	0.8232	0.8170	0.8764
1.3162	7.54	1750	0.6334	0.8340	0.8519	0.8340	0.8320	0.8834
1.3162	7.76	1800	0.6133	0.8286	0.8513	0.8286	0.8274	0.8798
1.3162	7.97	1850	0.5488	0.8502	0.8663	0.8502	0.8496	0.8949
1.2312	8.19	1900	0.6521	0.8246	0.8411	0.8246	0.8227	0.8769
1.2312	8.41	1950	0.5706	0.8529	0.8669	0.8529	0.8528	0.8962
1.2312	8.62	2000	0.5822	0.8462	0.8596	0.8462	0.8448	0.8924
1.2312	8.84	2050	0.5332	0.8502	0.8646	0.8502	0.8498	0.8953
1.1409	9.05	2100	0.5226	0.8650	0.8743	0.8650	0.8631	0.9053
1.1409	9.27	2150	0.5451	0.8623	0.8750	0.8623	0.8617	0.9032
1.1409	9.48	2200	0.5940	0.8381	0.8510	0.8381	0.8365	0.8860
1.1409	9.7	2250	0.5303	0.8570	0.8686	0.8570	0.8568	0.8988
1.1409	9.91	2300	0.5706	0.8448	0.8622	0.8448	0.8429	0.8912
1.0865	10.13	2350	0.5140	0.8623	0.8780	0.8623	0.8635	0.9026
1.0865	10.34	2400	0.5106	0.8704	0.8811	0.8704	0.8692	0.9092
1.0865	10.56	2450	0.5478	0.8583	0.8753	0.8583	0.8570	0.9005
1.0865	10.78	2500	0.6036	0.8583	0.8694	0.8583	0.8548	0.9003
1.0865	10.99	2550	0.5360	0.8543	0.8712	0.8543	0.8498	0.8984
1.0383	11.21	2600	0.5426	0.8570	0.8691	0.8570	0.8558	0.8982
1.0383	11.42	2650	0.5124	0.8691	0.8777	0.8691	0.8673	0.9067
1.0383	11.64	2700	0.5676	0.8435	0.8554	0.8435	0.8422	0.8892
1.0383	11.85	2750	0.5387	0.8596	0.8700	0.8596	0.8590	0.9022
0.9938	12.07	2800	0.5402	0.8691	0.8778	0.8691	0.8675	0.9089
0.9938	12.28	2850	0.5814	0.8529	0.8603	0.8529	0.8496	0.8969
0.9938	12.5	2900	0.5124	0.8623	0.8705	0.8623	0.8594	0.9034
0.9938	12.72	2950	0.5077	0.8623	0.8739	0.8623	0.8604	0.9032
0.9938	12.93	3000	0.5305	0.8704	0.8785	0.8704	0.8675	0.9101
0.9526	13.15	3050	0.5455	0.8718	0.8849	0.8718	0.8707	0.9100
0.9526	13.36	3100	0.5153	0.8826	0.8939	0.8826	0.8822	0.9175
0.9526	13.58	3150	0.5218	0.8826	0.8902	0.8826	0.8813	0.9167
0.9526	13.79	3200	0.5361	0.8637	0.8756	0.8637	0.8634	0.9030
0.91	14.01	3250	0.5174	0.8785	0.8873	0.8785	0.8780	0.9139
0.91	14.22	3300	0.5346	0.8799	0.8892	0.8799	0.8787	0.9158
0.91	14.44	3350	0.5586	0.8650	0.8747	0.8650	0.8634	0.9050
0.91	14.66	3400	0.5504	0.8704	0.8816	0.8704	0.8698	0.9097
0.91	14.87	3450	0.5643	0.8718	0.8814	0.8718	0.8700	0.9101
0.8689	15.09	3500	0.5425	0.8650	0.8766	0.8650	0.8642	0.9043
0.8689	15.3	3550	0.5609	0.8623	0.8775	0.8623	0.8616	0.9038
0.8689	15.52	3600	0.5440	0.8745	0.8847	0.8745	0.8739	0.9116
0.8689	15.73	3650	0.5020	0.8718	0.8814	0.8718	0.8714	0.9103
0.8689	15.95	3700	0.5650	0.8718	0.8810	0.8718	0.8704	0.9099
0.8437	16.16	3750	0.5115	0.8785	0.8874	0.8785	0.8774	0.9146
0.8437	16.38	3800	0.5651	0.8596	0.8735	0.8596	0.8592	0.9022
0.8437	16.59	3850	0.4996	0.8920	0.9025	0.8920	0.8921	0.9242
0.8437	16.81	3900	0.5528	0.8772	0.8887	0.8772	0.8765	0.9134
0.8213	17.03	3950	0.5568	0.8677	0.8816	0.8677	0.8666	0.9074
0.8213	17.24	4000	0.5270	0.8812	0.8906	0.8812	0.8804	0.9167
0.8213	17.46	4050	0.5239	0.8812	0.8922	0.8812	0.8800	0.9162
0.8213	17.67	4100	0.4915	0.8839	0.8921	0.8839	0.8834	0.9181
0.8213	17.89	4150	0.5282	0.8812	0.8914	0.8812	0.8807	0.9152
0.7835	18.1	4200	0.5031	0.8866	0.8959	0.8866	0.8865	0.9194
0.7835	18.32	4250	0.4997	0.8812	0.8898	0.8812	0.8803	0.9158
0.7835	18.53	4300	0.5080	0.8826	0.8904	0.8826	0.8809	0.9167
0.7835	18.75	4350	0.5264	0.8812	0.8898	0.8812	0.8800	0.9158
0.7835	18.97	4400	0.5487	0.8718	0.8808	0.8718	0.8707	0.9105
0.7606	19.18	4450	0.5266	0.8772	0.8877	0.8772	0.8759	0.9139
0.7606	19.4	4500	0.5257	0.8772	0.8875	0.8772	0.8770	0.9139
0.7606	19.61	4550	0.5321	0.8880	0.8977	0.8880	0.8882	0.9215
0.7606	19.83	4600	0.5349	0.8772	0.8880	0.8772	0.8765	0.9139
0.7342	20.04	4650	0.5250	0.8880	0.8962	0.8880	0.8877	0.9219
0.7342	20.26	4700	0.5081	0.8907	0.8990	0.8907	0.8904	0.9232
0.7342	20.47	4750	0.4958	0.8839	0.8941	0.8839	0.8842	0.9171
0.7342	20.69	4800	0.5293	0.8826	0.8928	0.8826	0.8819	0.9181
0.7342	20.91	4850	0.5094	0.8812	0.8924	0.8812	0.8805	0.9167
0.7129	21.12	4900	0.4922	0.8920	0.8997	0.8920	0.8908	0.9242
0.7129	21.34	4950	0.5078	0.8907	0.9000	0.8907	0.8901	0.9238
0.7129	21.55	5000	0.5303	0.8799	0.8892	0.8799	0.8781	0.9167
0.7129	21.77	5050	0.5531	0.8731	0.8842	0.8731	0.8711	0.9115
0.7129	21.98	5100	0.5572	0.8799	0.8920	0.8799	0.8784	0.9158
0.7032	22.2	5150	0.5151	0.8799	0.8903	0.8799	0.8793	0.9167
0.7032	22.41	5200	0.5090	0.8812	0.8921	0.8812	0.8808	0.9177
0.7032	22.63	5250	0.5318	0.8799	0.8891	0.8799	0.8785	0.9158
0.7032	22.84	5300	0.5114	0.8826	0.8897	0.8826	0.8812	0.9171
0.6809	23.06	5350	0.5049	0.8866	0.8946	0.8866	0.8858	0.9209
0.6809	23.28	5400	0.5378	0.8799	0.8901	0.8799	0.8786	0.9152
0.6809	23.49	5450	0.5088	0.8812	0.8905	0.8812	0.8806	0.9158
0.6809	23.71	5500	0.4883	0.8920	0.9033	0.8920	0.8925	0.9252
0.6809	23.92	5550	0.5168	0.8799	0.8911	0.8799	0.8800	0.9152
0.6604	24.14	5600	0.5167	0.8799	0.8907	0.8799	0.8795	0.9148
0.6604	24.35	5650	0.5092	0.8866	0.9011	0.8866	0.8878	0.9200
0.6604	24.57	5700	0.5048	0.8961	0.9069	0.8961	0.8965	0.9270
0.6604	24.78	5750	0.5303	0.8839	0.8973	0.8839	0.8835	0.9186
0.6604	25.0	5800	0.4996	0.8934	0.9041	0.8934	0.8939	0.9242
0.6595	25.22	5850	0.5095	0.8934	0.9033	0.8934	0.8927	0.9242
0.6595	25.43	5900	0.5109	0.8920	0.9024	0.8920	0.8921	0.9232
0.6595	25.65	5950	0.4993	0.8893	0.8973	0.8893	0.8890	0.9219
0.6595	25.86	6000	0.4954	0.8934	0.9022	0.8934	0.8928	0.9247
0.6347	26.08	6050	0.4939	0.8988	0.9076	0.8988	0.8986	0.9279
0.6347	26.29	6100	0.4820	0.8974	0.9049	0.8974	0.8970	0.9270
0.6347	26.51	6150	0.5168	0.8880	0.8952	0.8880	0.8869	0.9205
0.6347	26.72	6200	0.5275	0.8839	0.8916	0.8839	0.8827	0.9181
0.6347	26.94	6250	0.5026	0.8907	0.8991	0.8907	0.8898	0.9219
0.6361	27.16	6300	0.5003	0.8988	0.9076	0.8988	0.8984	0.9275
0.6361	27.37	6350	0.4777	0.8988	0.9069	0.8988	0.8984	0.9275
0.6361	27.59	6400	0.4904	0.8988	0.9079	0.8988	0.8986	0.9275
0.6361	27.8	6450	0.4885	0.9001	0.9084	0.9001	0.8998	0.9285
0.631	28.02	6500	0.5134	0.8893	0.8973	0.8893	0.8882	0.9209
0.631	28.23	6550	0.5128	0.8920	0.9011	0.8920	0.8916	0.9232
0.631	28.45	6600	0.5136	0.8947	0.9032	0.8947	0.8942	0.9251
0.631	28.66	6650	0.5148	0.8907	0.8998	0.8907	0.8900	0.9219
0.631	28.88	6700	0.5143	0.8893	0.8971	0.8893	0.8883	0.9215
0.6104	29.09	6750	0.5237	0.8853	0.8952	0.8853	0.8844	0.9181
0.6104	29.31	6800	0.5187	0.8880	0.8976	0.8880	0.8873	0.9200
0.6104	29.53	6850	0.5183	0.8866	0.8964	0.8866	0.8860	0.9190
0.6104	29.74	6900	0.5172	0.8907	0.9006	0.8907	0.8899	0.9219
0.6104	29.96	6950	0.5141	0.8907	0.9001	0.8907	0.8902	0.9219

Framework versions

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