jialicheng/unlearn_speech_commands_wav2vec2-large_random_label_6_42

superb_ks_42

This model is a fine-tuned version of facebook/wav2vec2-large on the superb dataset. It achieves the following results on the evaluation set:

• Loss: 0.2130
• Accuracy: 0.9743
• Test Accuracy: 0.9743
• Df Accuracy: 0.9638
• Unlearn Overall Accuracy: 0.5052
• Unlearn Time: 8878.6492

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
0.4081	1.0	1597	0.1837	0.9746	0.5038	0.5038	-1
0.3772	2.0	3194	0.1881	0.9693	0.5043	0.5043	-1
0.4035	3.0	4791	0.2130	0.9638	0.5052	0.5052	-1
0.4123	4.0	6388	0.2009	0.9684	0.5038	0.5038	-1
0.4061	5.0	7985	0.1858	0.9680	0.5039	0.5039	-1
0.3944	6.0	9582	0.1756	0.9749	0.5005	0.5005	-1
0.3775	7.0	11179	0.2013	0.9667	0.5041	0.5041	-1
0.3778	8.0	12776	0.2132	0.9716	0.5004	0.5004	-1
0.3679	9.0	14373	0.1963	0.9755	0.5000	0.5000	-1
0.3612	10.0	15970	0.1786	0.9759	0.5012	0.5012	-1
0.3534	11.0	17567	0.1659	0.9759	0.5013	0.5013	-1
0.3401	12.0	19164	0.1557	0.9759	0.5018	0.5018	-1
0.3222	13.0	20761	0.1708	0.9778	0.5000	0.5000	-1
0.3239	14.0	22358	0.1617	0.9791	0.5015	0.5015	-1
0.3163	15.0	23955	0.1592	0.9781	0.5017	0.5017	-1
0.3049	16.0	25552	0.1513	0.9804	0.5010	0.5010	-1
0.2928	17.0	27149	0.1451	0.9801	0.5013	0.5013	-1
0.2912	18.0	28746	0.1539	0.9804	0.5007	0.5007	-1
0.2846	19.0	30343	0.1540	0.9808	0.5009	0.5009	-1
0.2851	20.0	31940	0.1580	0.9808	0.5009	0.5009	-1

Framework versions

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