ai-light-dance_drums_ft_pretrain_wav2vec2-base

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

• Loss: 1.8991
• Wer: 0.6046

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: 0.0004
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 20
• num_epochs: 200.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.9	8	2.0434	0.6226
0.4739	1.9	16	2.1024	0.6247
0.4693	2.9	24	1.9824	0.6211
0.5139	3.9	32	2.2962	0.6429
0.5081	4.9	40	2.2201	0.6292
0.5081	5.9	48	2.1399	0.6208
0.5785	6.9	56	2.1451	0.6417
0.533	7.9	64	2.1184	0.6330
0.5141	8.9	72	2.0230	0.6342
0.4971	9.9	80	2.2137	0.6381
0.4971	10.9	88	2.1159	0.6253
0.5645	11.9	96	2.0966	0.6247
0.4932	12.9	104	1.9249	0.6223
0.4918	13.9	112	2.0445	0.6235
0.5053	14.9	120	2.1317	0.6304
0.5053	15.9	128	2.0723	0.6256
0.5565	16.9	136	2.1390	0.6402
0.4819	17.9	144	1.9556	0.6321
0.5131	18.9	152	1.9886	0.6333
0.4798	19.9	160	1.9700	0.6259
0.4798	20.9	168	1.9771	0.6295
0.5221	21.9	176	1.9880	0.6235
0.4862	22.9	184	2.0994	0.6298
0.4831	23.9	192	2.0521	0.6205
0.4952	24.9	200	1.9838	0.6064
0.4952	25.9	208	2.0319	0.6103
0.5119	26.9	216	2.0419	0.6160
0.4996	27.9	224	2.0073	0.6178
0.488	28.9	232	2.1740	0.6304
0.4978	29.9	240	2.2731	0.6163
0.4978	30.9	248	2.2420	0.6205
0.5259	31.9	256	2.0561	0.6184
0.47	32.9	264	1.9455	0.6136
0.5132	33.9	272	1.9307	0.6043
0.4972	34.9	280	2.0536	0.6127
0.4972	35.9	288	1.9113	0.6223
0.5147	36.9	296	1.9317	0.6286
0.4914	37.9	304	2.1810	0.6241
0.472	38.9	312	2.1403	0.6160
0.4825	39.9	320	2.1141	0.6094
0.4825	40.9	328	2.2870	0.6031
0.5138	41.9	336	2.1404	0.6181
0.48	42.9	344	2.0243	0.6265
0.4598	43.9	352	2.1117	0.6199
0.474	44.9	360	2.0378	0.6321
0.474	45.9	368	2.1919	0.6211
0.4933	46.9	376	2.3645	0.6109
0.4692	47.9	384	2.1920	0.6076
0.4716	48.9	392	2.3663	0.6034
0.4601	49.9	400	2.2838	0.6280
0.4601	50.9	408	2.0287	0.6148
0.4891	51.9	416	2.1346	0.6130
0.4506	52.9	424	2.1556	0.6181
0.4581	53.9	432	2.0560	0.6229
0.4485	54.9	440	1.9944	0.5971
0.4485	55.9	448	1.9791	0.6097
0.4942	56.9	456	2.1166	0.6070
0.4748	57.9	464	2.0271	0.6124
0.4229	58.9	472	2.0437	0.6229
0.45	59.9	480	2.1012	0.6142
0.45	60.9	488	1.9151	0.6049
0.4936	61.9	496	1.8991	0.6046
0.4602	62.9	504	1.9813	0.6112
0.4626	63.9	512	1.9372	0.6136
0.445	64.9	520	1.9060	0.6154
0.445	65.9	528	1.9574	0.6151
0.4907	66.9	536	2.0947	0.6022
0.4723	67.9	544	2.0061	0.6010
0.4103	68.9	552	1.9557	0.6094
0.4808	69.9	560	2.1042	0.6088
0.4808	70.9	568	2.1360	0.6073
0.4682	71.9	576	2.1290	0.6013
0.4472	72.9	584	1.9454	0.5989
0.4259	73.9	592	2.0937	0.6043
0.4464	74.9	600	2.0822	0.6058
0.4464	75.9	608	2.0128	0.6058
0.4775	76.9	616	1.9744	0.6094
0.4394	77.9	624	1.9992	0.6010
0.418	78.9	632	2.1693	0.5947
0.4384	79.9	640	2.1326	0.5923
0.4384	80.9	648	2.1151	0.5950
0.4971	81.9	656	2.1581	0.5923
0.4176	82.9	664	2.0876	0.6013
0.4312	83.9	672	2.1316	0.5935
0.4408	84.9	680	2.2627	0.5971
0.4408	85.9	688	2.2799	0.6112
0.4678	86.9	696	2.1239	0.5989
0.4288	87.9	704	2.1574	0.5983
0.4157	88.9	712	2.2125	0.5908
0.444	89.9	720	2.0542	0.5986
0.444	90.9	728	2.0899	0.5920
0.4694	91.9	736	2.1122	0.6076
0.4314	92.9	744	2.0634	0.5950
0.4348	93.9	752	2.0333	0.6046
0.4558	94.9	760	2.1188	0.5956
0.4558	95.9	768	2.0606	0.5995
0.461	96.9	776	2.0600	0.5971
0.4258	97.9	784	2.0479	0.6040
0.4395	98.9	792	2.1282	0.6055
0.4282	99.9	800	2.0593	0.6043
0.4282	100.9	808	2.0592	0.5920
0.4623	101.9	816	2.0852	0.5944
0.4392	102.9	824	2.2024	0.5920
0.4308	103.9	832	2.1786	0.5935
0.4375	104.9	840	2.1085	0.5911
0.4375	105.9	848	2.0724	0.5974
0.4501	106.9	856	2.1306	0.5881
0.4273	107.9	864	2.1340	0.5899
0.4234	108.9	872	2.1125	0.5980
0.4289	109.9	880	2.0526	0.6007
0.4289	110.9	888	2.0955	0.5884
0.478	111.9	896	2.1146	0.5872
0.4143	112.9	904	2.2310	0.5899
0.4193	113.9	912	2.2165	0.5899
0.4159	114.9	920	2.1631	0.5941
0.4159	115.9	928	2.1371	0.5938
0.4776	116.9	936	2.0972	0.5935
0.4143	117.9	944	2.1248	0.5917
0.4022	118.9	952	2.1317	0.5956
0.4346	119.9	960	2.1237	0.5992
0.4346	120.9	968	2.0684	0.5935
0.4564	121.9	976	2.0722	0.5947
0.4243	122.9	984	2.1361	0.5884
0.413	123.9	992	2.1207	0.5893
0.4113	124.9	1000	2.0697	0.5837
0.4113	125.9	1008	2.1005	0.5875
0.4426	126.9	1016	2.0822	0.5870
0.4255	127.9	1024	2.0572	0.5959
0.4214	128.9	1032	2.0343	0.5935
0.4042	129.9	1040	2.0282	0.5902
0.4042	130.9	1048	2.0314	0.5846
0.4515	131.9	1056	2.0621	0.5870
0.4138	132.9	1064	2.0704	0.5938
0.4289	133.9	1072	2.0222	0.5896
0.3908	134.9	1080	2.0879	0.5855
0.3908	135.9	1088	2.1068	0.5822
0.4489	136.9	1096	2.0702	0.5837
0.4191	137.9	1104	2.1093	0.5881
0.4149	138.9	1112	2.1046	0.5819
0.4127	139.9	1120	2.1729	0.5777
0.4127	140.9	1128	2.1636	0.5810
0.4449	141.9	1136	2.1515	0.5786
0.3977	142.9	1144	2.1531	0.5774
0.4121	143.9	1152	2.0857	0.5816
0.4363	144.9	1160	2.1372	0.5822
0.4363	145.9	1168	2.1902	0.5828
0.4318	146.9	1176	2.1465	0.5831
0.4112	147.9	1184	2.0697	0.5858
0.4292	148.9	1192	2.0850	0.5837
0.4182	149.9	1200	2.1171	0.5846
0.4182	150.9	1208	2.1020	0.5867
0.4381	151.9	1216	2.1052	0.5849
0.4235	152.9	1224	2.1430	0.5864
0.4173	153.9	1232	2.1131	0.5834
0.3927	154.9	1240	2.1134	0.5846
0.3927	155.9	1248	2.1173	0.5846
0.4492	156.9	1256	2.0772	0.5801
0.4313	157.9	1264	2.0309	0.5861
0.4015	158.9	1272	2.0887	0.5819
0.4268	159.9	1280	2.1812	0.5849
0.4268	160.9	1288	2.1568	0.5881
0.4496	161.9	1296	2.0805	0.5801
0.4121	162.9	1304	2.0461	0.5872
0.401	163.9	1312	2.0377	0.5864
0.4192	164.9	1320	2.0183	0.5872
0.4192	165.9	1328	2.0107	0.5855
0.4466	166.9	1336	2.0528	0.5881
0.3981	167.9	1344	2.0511	0.5878
0.3967	168.9	1352	2.0374	0.5867
0.4072	169.9	1360	2.0554	0.5867
0.4072	170.9	1368	2.0388	0.5858
0.4581	171.9	1376	2.0188	0.5914
0.3937	172.9	1384	1.9999	0.5852
0.4074	173.9	1392	1.9738	0.5840
0.4085	174.9	1400	2.0090	0.5843
0.4085	175.9	1408	1.9990	0.5864
0.4224	176.9	1416	2.0391	0.5852
0.4471	177.9	1424	2.0262	0.5855
0.4233	178.9	1432	2.0621	0.5801
0.409	179.9	1440	2.0486	0.5846
0.409	180.9	1448	2.0508	0.5807
0.4518	181.9	1456	2.0241	0.5887
0.4077	182.9	1464	2.0169	0.5843
0.4197	183.9	1472	2.0014	0.5896
0.4237	184.9	1480	2.0189	0.5843
0.4237	185.9	1488	2.0095	0.5867
0.4394	186.9	1496	1.9993	0.5884
0.4299	187.9	1504	2.0097	0.5899
0.4198	188.9	1512	2.0049	0.5870
0.4116	189.9	1520	1.9899	0.5875
0.4116	190.9	1528	1.9814	0.5881
0.445	191.9	1536	1.9820	0.5887
0.4198	192.9	1544	1.9838	0.5881
0.4065	193.9	1552	1.9849	0.5884
0.3917	194.9	1560	1.9803	0.5867
0.3917	195.9	1568	1.9777	0.5881
0.4239	196.9	1576	1.9752	0.5875
0.4183	197.9	1584	1.9766	0.5872
0.3965	198.9	1592	1.9773	0.5872
0.4144	199.9	1600	1.9781	0.5872

Framework versions

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