jialicheng/unlearn_speech_commands_hubert-base_random_label_2_42

superb_ks_42

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

• Loss: 0.1141
• Accuracy: 0.9818
• Test Accuracy: 0.9818
• Df Accuracy: 0.9736
• Unlearn Overall Accuracy: 0.5041
• Unlearn Time: 3242.6132

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: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Overall Accuracy	Unlearn Overall Accuracy	Time
0.1998	1.0	1597	0.1072	0.9775	0.5018	0.5018	-1
0.1835	2.0	3194	0.1141	0.9736	0.5041	0.5041	-1
0.1964	3.0	4791	0.0950	0.9775	0.5025	0.5025	-1
0.189	4.0	6388	0.1102	0.9726	0.5039	0.5039	-1
0.1827	5.0	7985	0.1138	0.9765	0.5017	0.5017	-1
0.1775	6.0	9582	0.1098	0.9775	0.5027	0.5027	-1
0.1696	7.0	11179	0.1009	0.9794	0.5022	0.5022	-1
0.1749	8.0	12776	0.1219	0.9804	0.5007	0.5007	-1
0.1651	9.0	14373	0.1012	0.9814	0.5014	0.5014	-1
0.1654	10.0	15970	0.1182	0.9794	0.5015	0.5015	-1
0.1568	11.0	17567	0.1028	0.9824	0.5007	0.5007	-1
0.1544	12.0	19164	0.0921	0.9814	0.5017	0.5017	-1
0.145	13.0	20761	0.1060	0.9814	0.5014	0.5014	-1
0.1473	14.0	22358	0.1043	0.9824	0.5013	0.5013	-1
0.1418	15.0	23955	0.1248	0.9794	0.5007	0.5007	-1
0.1401	16.0	25552	0.1007	0.9824	0.5009	0.5009	-1
0.1351	17.0	27149	0.1111	0.9833	0.5001	0.5001	-1
0.1305	18.0	28746	0.1101	0.9833	0.5001	0.5001	-1
0.1257	19.0	30343	0.1100	0.9833	0.4999	0.4999	-1
0.1297	20.0	31940	0.1106	0.9833	0.4999	0.4999	-1

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu118
• Datasets 2.18.0
• Tokenizers 0.15.2