ai-light-dance_drums_ft_pretrain_wav2vec2-base-new-13k_onset-drums_fold_2

This model is a fine-tuned version of gary109/ai-light-dance_drums_ft_pretrain_wav2vec2-base-new-13k_onset-drums_fold_1 on the GARY109/AI_LIGHT_DANCE - ONSET-DRUMS_FOLD_2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4180
• Wer: 0.1433

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
0.3238	0.99	69	0.4581	0.2081
0.275	1.99	138	0.6494	0.3343
0.2965	2.99	207	0.6193	0.2275
0.3406	3.99	276	0.6934	0.2615
0.3906	4.99	345	0.6265	0.1835
0.4643	5.99	414	0.5879	0.1899
0.4652	6.99	483	0.4961	0.1604
0.4512	7.99	552	0.5712	0.2801
0.5321	8.99	621	0.6898	0.2936
0.64	9.99	690	0.5916	0.2648
0.2959	10.99	759	0.5574	0.1745
0.2053	11.99	828	0.5216	0.2009
0.2433	12.99	897	0.4738	0.1643
0.2036	13.99	966	0.5063	0.1651
0.2654	14.99	1035	0.4904	0.1511
0.3641	15.99	1104	0.4660	0.1669
0.373	16.99	1173	0.5133	0.2106
0.4715	17.99	1242	0.5313	0.1912
0.4893	18.99	1311	0.5152	0.1712
0.4875	19.99	1380	0.5482	0.1718
0.1971	20.99	1449	0.4566	0.1449
0.1286	21.99	1518	0.4515	0.1478
0.1472	22.99	1587	0.5059	0.1418
0.1917	23.99	1656	0.5583	0.1457
0.2874	24.99	1725	0.5195	0.1503
0.2252	25.99	1794	0.4409	0.1506
0.3142	26.99	1863	0.4180	0.1433
0.385	27.99	1932	0.4708	0.1367
0.4296	28.99	2001	0.4740	0.1506
0.4404	29.99	2070	0.4652	0.1646
0.2466	30.99	2139	0.5013	0.1528
0.1017	31.99	2208	0.4578	0.1552
0.1383	32.99	2277	0.5026	0.1419
0.1719	33.99	2346	0.4651	0.1442
0.1808	34.99	2415	0.4499	0.1412
0.2429	35.99	2484	0.4523	0.1472
0.2651	36.99	2553	0.4544	0.1397
0.2748	37.99	2622	0.4181	0.1386
0.4171	38.99	2691	0.4385	0.1334
0.4119	39.99	2760	0.4568	0.1504
0.1453	40.99	2829	0.4425	0.1431
0.105	41.99	2898	0.4367	0.1353
0.1205	42.99	2967	0.4418	0.1340
0.2039	43.99	3036	0.4586	0.1379
0.1773	44.99	3105	0.4686	0.1391
0.2186	45.99	3174	0.4975	0.1446
0.2358	46.99	3243	0.4886	0.1448
0.3525	47.99	3312	0.4706	0.1398
0.3713	48.99	3381	0.4713	0.1388
0.3543	49.99	3450	0.4720	0.1388

Framework versions

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