ai-light-dance_drums_ft_pretrain_wav2vec2-base-new-13k_onset-drums_fold_3

This model is a fine-tuned version of gary109/ai-light-dance_drums_ft_pretrain_wav2vec2-base-new-13k_onset-drums_fold_2 on the GARY109/AI_LIGHT_DANCE - ONSET-DRUMS_FOLD_3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4093
• Wer: 0.1250

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
0.4557	1.0	70	0.5794	0.1197
0.6796	2.0	140	0.5726	0.1388
0.4511	3.0	210	0.6290	0.1242
0.609	4.0	280	0.7112	0.1187
0.4082	5.0	350	0.8275	0.1965
0.4638	6.0	420	0.4767	0.1524
0.4446	7.0	490	0.5091	0.1376
0.4337	8.0	560	0.6622	0.1170
0.4604	9.0	630	0.7242	0.1600
0.4462	10.0	700	0.7298	0.1383
0.4201	11.0	770	0.8058	0.1362
0.4204	12.0	840	0.6255	0.1099
0.461	13.0	910	0.5204	0.1109
0.3779	14.0	980	0.6911	0.1125
0.3403	15.0	1050	0.5863	0.1188
0.6223	16.0	1120	0.6367	0.1147
0.3827	17.0	1190	0.6266	0.1293
0.3055	18.0	1260	0.4866	0.1095
0.3917	19.0	1330	0.4093	0.1250
0.3912	20.0	1400	0.4514	0.1077
0.3861	21.0	1470	0.5043	0.1156
0.3659	22.0	1540	0.5680	0.1091
0.3536	23.0	1610	0.7940	0.1029
0.3559	24.0	1680	0.5877	0.1101
0.3274	25.0	1750	0.4461	0.1059
0.5232	26.0	1820	1.2051	0.1068
0.3241	27.0	1890	0.8716	0.1099
0.3169	28.0	1960	0.6752	0.1082
0.2938	29.0	2030	0.6023	0.1071
0.3022	30.0	2100	0.6122	0.1146
0.4245	31.0	2170	0.5735	0.1102
0.3095	32.0	2240	0.4476	0.1042
0.4062	33.0	2310	0.6339	0.1130
0.3202	34.0	2380	0.4101	0.1077
0.2952	35.0	2450	0.4825	0.1076
0.2945	36.0	2520	0.4998	0.1058
0.336	37.0	2590	0.5490	0.1061
0.2912	38.0	2660	0.4804	0.1038
0.282	39.0	2730	0.4776	0.1022
0.4359	40.0	2800	0.4376	0.1044
0.2698	41.0	2870	0.5609	0.1098
0.3004	42.0	2940	0.5258	0.1083
0.2873	43.0	3010	0.4810	0.1069
0.3413	44.0	3080	0.4961	0.1080
0.2802	45.0	3150	0.6850	0.1076
0.2584	46.0	3220	0.7210	0.1082
0.3282	47.0	3290	0.6179	0.1053
0.2666	48.0	3360	0.7673	0.1075
0.2989	49.0	3430	0.7710	0.1079
0.2676	50.0	3500	0.7655	0.1076

Framework versions

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