ai-light-dance_drums_ft_pretrain_wav2vec2-base-new-13k_onset-drums_fold_1

This model is a fine-tuned version of gary109/ai-light-dance_drums_pretrain_wav2vec2-base-new-13k on the GARY109/AI_LIGHT_DANCE - ONSET-DRUMS_FOLD_1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4630
• Wer: 0.2145

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: 0.0003
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 50.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.0614	0.99	69	5.1275	1.0
1.8291	1.99	138	2.2008	1.0
1.4664	2.99	207	1.6821	1.0
1.287	3.99	276	1.5681	1.0
1.2642	4.99	345	1.5074	1.0
1.2702	5.99	414	1.4650	1.0
1.2245	6.99	483	1.3027	1.0
1.3461	7.99	552	1.3109	1.0
1.2903	8.99	621	1.3107	1.0
1.2741	9.99	690	1.1842	1.0
1.1446	10.99	759	1.1754	1.0
1.0746	11.99	828	1.1469	0.9999
0.8203	12.99	897	0.9071	0.6202
0.5996	13.99	966	0.7047	0.4234
0.5672	14.99	1035	0.5369	0.2567
0.4965	15.99	1104	0.4644	0.2861
0.5639	16.99	1173	0.4630	0.2145
0.6272	17.99	1242	0.6848	0.2667
0.6764	18.99	1311	0.6074	0.2508
0.7205	19.99	1380	0.6452	0.2184
0.346	20.99	1449	0.5962	0.2457
0.2212	21.99	1518	0.5236	0.2068
0.1646	22.99	1587	0.6130	0.2198
0.3148	23.99	1656	0.5592	0.2620
0.3061	24.99	1725	0.5577	0.2560
0.3137	25.99	1794	0.5247	0.2227
0.389	26.99	1863	0.5799	0.2081
0.4168	27.99	1932	0.5850	0.1818
0.4403	28.99	2001	0.5687	0.2053
0.4936	29.99	2070	0.5511	0.2065
0.2196	30.99	2139	0.5438	0.1706
0.1683	31.99	2208	0.6066	0.1855
0.1552	32.99	2277	0.5248	0.1930
0.1682	33.99	2346	0.5440	0.1783
0.2162	34.99	2415	0.6079	0.1778
0.3041	35.99	2484	0.5608	0.1834
0.3188	36.99	2553	0.6039	0.2007
0.3692	37.99	2622	0.5437	0.1769
0.4446	38.99	2691	0.6475	0.1881
0.386	39.99	2760	0.6468	0.1894
0.1995	40.99	2829	0.6398	0.1906
0.1174	41.99	2898	0.5987	0.1936
0.1288	42.99	2967	0.6133	0.1871
0.1857	43.99	3036	0.6976	0.1995
0.2025	44.99	3105	0.6356	0.1902
0.2922	45.99	3174	0.6324	0.2055
0.3575	46.99	3243	0.6338	0.1862
0.4019	47.99	3312	0.6113	0.1898
0.4211	48.99	3381	0.6320	0.1948
0.4323	49.99	3450	0.6307	0.1917

Framework versions

• Transformers 4.24.0.dev0
• Pytorch 1.12.1+cu113
• Datasets 2.6.1
• Tokenizers 0.13.1