ai-light-dance_drums_ft_pretrain_wav2vec2-base-new_onset-rbma13-2_7k

This model is a fine-tuned version of gary109/ai-light-dance_drums_pretrain_wav2vec2-base-new-7k on the GARY109/AI_LIGHT_DANCE - ONSET-RBMA13-2 dataset. It achieves the following results on the evaluation set:

• Loss: 2.3330
• Wer: 1.0

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: 30
• num_epochs: 100.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	1	68.1358	1.0
No log	2.0	2	68.1358	1.0
No log	3.0	3	68.1358	1.0
No log	4.0	4	68.0245	1.0
No log	5.0	5	67.7874	1.0
No log	6.0	6	67.4535	1.0
No log	7.0	7	67.0142	1.0
No log	8.0	8	67.0142	1.0
No log	9.0	9	66.4335	1.0
38.4011	10.0	10	65.7100	1.0
38.4011	11.0	11	64.8206	1.0
38.4011	12.0	12	63.8239	1.0
38.4011	13.0	13	62.6489	1.0
38.4011	14.0	14	61.3071	1.0
38.4011	15.0	15	59.7427	1.0
38.4011	16.0	16	58.0256	0.98
38.4011	17.0	17	56.0327	1.0
38.4011	18.0	18	53.7724	1.0
38.4011	19.0	19	51.2556	1.0
33.2554	20.0	20	48.4956	1.0
33.2554	21.0	21	45.4038	1.0
33.2554	22.0	22	41.9980	1.0
33.2554	23.0	23	41.9980	1.0
33.2554	24.0	24	38.2281	1.0
33.2554	25.0	25	34.1577	1.0
33.2554	26.0	26	29.7985	1.0
33.2554	27.0	27	25.1146	1.0
33.2554	28.0	28	20.2287	1.0
33.2554	29.0	29	15.3406	1.0
15.1206	30.0	30	10.7693	1.0
15.1206	31.0	31	6.8998	1.0
15.1206	32.0	32	4.5907	1.0
15.1206	33.0	33	3.3596	1.0
15.1206	34.0	34	2.7711	1.0
15.1206	35.0	35	2.5962	1.0
15.1206	36.0	36	2.9002	1.0
15.1206	37.0	37	3.0061	1.0
15.1206	38.0	38	2.8175	1.0
15.1206	39.0	39	2.4512	1.0
2.4298	40.0	40	2.3330	1.0
2.4298	41.0	41	2.3766	1.0
2.4298	42.0	42	2.5626	1.0
2.4298	43.0	43	2.9632	1.0
2.4298	44.0	44	3.2796	1.0
2.4298	45.0	45	3.4015	1.0
2.4298	46.0	46	3.2808	1.0
2.4298	47.0	47	3.2373	1.0
2.4298	48.0	48	3.2462	1.0
2.4298	49.0	49	3.6168	1.0
1.6143	50.0	50	3.6625	1.0
1.6143	51.0	51	3.7593	1.0
1.6143	52.0	52	3.9327	1.0
1.6143	53.0	53	3.7185	1.0
1.6143	54.0	54	3.9100	1.0
1.6143	55.0	55	4.3123	1.0
1.6143	56.0	56	4.2904	1.0
1.6143	57.0	57	3.9519	1.0
1.6143	58.0	58	3.4518	1.0
1.6143	59.0	59	3.0197	1.0
1.4054	60.0	60	2.8863	1.0
1.4054	61.0	61	2.9754	1.0
1.4054	62.0	62	3.2998	1.0
1.4054	63.0	63	3.8715	1.0
1.4054	64.0	64	4.1898	1.0
1.4054	65.0	65	4.1813	1.0
1.4054	66.0	66	3.9025	1.0
1.4054	67.0	67	3.4319	1.0
1.4054	68.0	68	3.2755	1.0
1.4054	69.0	69	3.3349	1.0
1.3121	70.0	70	3.5485	1.0
1.3121	71.0	71	3.9019	1.0
1.3121	72.0	72	4.0819	1.0
1.3121	73.0	73	3.9955	1.0
1.3121	74.0	74	3.7088	1.0
1.3121	75.0	75	3.2957	1.0
1.3121	76.0	76	3.1141	1.0
1.3121	77.0	77	3.0852	1.0
1.3121	78.0	78	3.1871	1.0
1.3121	79.0	79	3.4127	1.0
1.2576	80.0	80	3.6913	1.0
1.2576	81.0	81	3.8286	1.0
1.2576	82.0	82	3.8157	1.0
1.2576	83.0	83	3.6814	1.0
1.2576	84.0	84	3.4496	1.0
1.2576	85.0	85	3.2844	1.0
1.2576	86.0	86	3.2254	1.0
1.2576	87.0	87	3.2683	1.0
1.2576	88.0	88	3.3791	1.0
1.2576	89.0	89	3.5501	1.0
1.2373	90.0	90	3.6622	1.0
1.2373	91.0	91	3.7207	1.0
1.2373	92.0	92	3.6961	1.0
1.2373	93.0	93	3.6099	1.0
1.2373	94.0	94	3.5336	1.0
1.2373	95.0	95	3.4342	1.0
1.2373	96.0	96	3.3170	1.0
1.2373	97.0	97	3.2624	1.0
1.2373	98.0	98	3.2437	1.0
1.2373	99.0	99	3.2591	1.0
1.1952	100.0	100	3.2927	1.0

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.8.1+cu111
• Datasets 2.7.1.dev0
• Tokenizers 0.13.2