wav2vec2-large-mms-1b-azz-adapter-all_data_10epochs

This model is a fine-tuned version of facebook/mms-1b-all on the audiofolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.4129
• Wer: 0.3069
• Cer: 0.0941

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: 0.001
• train_batch_size: 20
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.0574	0.1294	200	0.9724	0.6573	0.1971
0.9322	0.2589	400	0.7179	0.5833	0.1726
0.9043	0.3883	600	0.7195	0.5468	0.1631
0.8735	0.5178	800	0.6374	0.5238	0.1556
0.773	0.6472	1000	0.6531	0.5076	0.1517
0.7799	0.7767	1200	0.6242	0.5147	0.1520
0.9846	0.9061	1400	0.5878	0.5035	0.1466
0.8117	1.0356	1600	0.5782	0.4827	0.1430
0.746	1.1650	1800	0.5716	0.4745	0.1407
0.7802	1.2945	2000	0.5586	0.4717	0.1390
0.9001	1.4239	2200	0.5548	0.4672	0.1400
0.7452	1.5534	2400	0.5631	0.4572	0.1366
0.676	1.6828	2600	0.5342	0.4490	0.1339
0.7153	1.8123	2800	0.5681	0.4486	0.1334
0.7538	1.9417	3000	0.5299	0.4455	0.1326
0.681	2.0712	3200	0.5385	0.4448	0.1316
0.652	2.2006	3400	0.5315	0.4323	0.1280
0.6037	2.3301	3600	0.5700	0.4439	0.1322
0.6534	2.4595	3800	0.5563	0.4281	0.1288
0.6403	2.5890	4000	0.5152	0.4275	0.1284
0.674	2.7184	4200	0.5150	0.4247	0.1265
0.7599	2.8479	4400	0.5017	0.4218	0.1259
0.6729	2.9773	4600	0.5041	0.4059	0.1220
0.5941	3.1068	4800	0.5145	0.4202	0.1272
0.6739	3.2362	5000	0.5028	0.4193	0.1241
0.6256	3.3657	5200	0.5290	0.4175	0.1227
0.7149	3.4951	5400	0.4998	0.4088	0.1226
0.6998	3.6246	5600	0.4919	0.4124	0.1231
0.6055	3.7540	5800	0.4959	0.4096	0.1206
0.6154	3.8835	6000	0.4943	0.4031	0.1216
0.7407	4.0129	6200	0.5014	0.4283	0.1234
0.6168	4.1424	6400	0.4956	0.3929	0.1173
0.6959	4.2718	6600	0.4954	0.3898	0.1174
0.6978	4.4013	6800	0.5512	0.3913	0.1179
0.9123	4.5307	7000	0.4915	0.3923	0.1170
0.5721	4.6602	7200	0.4913	0.4000	0.1209
0.7162	4.7896	7400	0.4745	0.3923	0.1167
0.6501	4.9191	7600	0.4815	0.3801	0.1145
0.6957	5.0485	7800	0.4814	0.4044	0.1183
0.5791	5.1780	8000	0.4766	0.3839	0.1152
0.6509	5.3074	8200	0.4840	0.4042	0.1203
0.6117	5.4369	8400	0.4800	0.3892	0.1163
0.7194	5.5663	8600	0.4920	0.3815	0.1149
0.8941	5.6958	8800	0.4678	0.3752	0.1130
0.5965	5.8252	9000	0.4702	0.3815	0.1143
0.6351	5.9547	9200	0.4703	0.3816	0.1148
0.5887	6.0841	9400	0.4665	0.3759	0.1138
0.5603	6.2136	9600	0.4866	0.3698	0.1112
0.6516	6.3430	9800	0.4685	0.3717	0.1124
0.6041	6.4725	10000	0.4708	0.3757	0.1131
0.5621	6.6019	10200	0.4669	0.3638	0.1102
0.6136	6.7314	10400	0.4792	0.3687	0.1113
0.5835	6.8608	10600	0.4657	0.3707	0.1119
0.5732	6.9903	10800	0.4723	0.3654	0.1109
0.6285	7.1197	11000	0.4668	0.3661	0.1094
0.6128	7.2492	11200	0.4785	0.3695	0.1129
0.5489	7.3786	11400	0.5141	0.3643	0.1105
0.5681	7.5081	11600	0.4582	0.3612	0.1093
0.4984	7.6375	11800	0.4705	0.3602	0.1083
0.8323	7.7670	12000	0.4689	0.3560	0.1073
0.5723	7.8964	12200	0.4647	0.3558	0.1064
0.62	8.0259	12400	0.4581	0.3555	0.1075
0.5197	8.1553	12600	0.4551	0.3538	0.1072
0.6087	8.2848	12800	0.4591	0.3643	0.1090
0.583	8.4142	13000	0.4526	0.3500	0.1066
0.7788	8.5437	13200	0.4548	0.3618	0.1084
0.6503	8.6731	13400	0.4511	0.3545	0.1056
0.7021	8.8026	13600	0.4653	0.3519	0.1066
0.5428	8.9320	13800	0.4473	0.3523	0.1056
0.5716	9.0615	14000	0.4517	0.3513	0.1070
0.5345	9.1909	14200	0.4431	0.3503	0.1057
0.6278	9.3204	14400	0.4400	0.3489	0.1056
0.5128	9.4498	14600	0.4501	0.3427	0.1032
0.5278	9.5793	14800	0.4649	0.3462	0.1058
0.6367	9.7087	15000	0.4427	0.3563	0.1072
0.5131	9.8382	15200	0.4422	0.3492	0.1053
0.5187	9.9676	15400	0.4361	0.3452	0.1044
0.4976	10.0971	15600	0.4317	0.3445	0.1041
0.5494	10.2265	15800	0.4462	0.3416	0.1032
0.5362	10.3560	16000	0.4295	0.3404	0.1032
0.5069	10.4854	16200	0.4403	0.3418	0.1026
0.5938	10.6149	16400	0.4305	0.3375	0.1025
0.5548	10.7443	16600	0.4394	0.3369	0.1023
0.5127	10.8738	16800	0.4429	0.3407	0.1025
0.5588	11.0032	17000	0.4441	0.3463	0.1045
0.517	11.1327	17200	0.4326	0.3357	0.1014
0.5102	11.2621	17400	0.4562	0.3330	0.1017
0.6477	11.3916	17600	0.4327	0.3358	0.1028
0.5468	11.5210	17800	0.4289	0.3360	0.1019
0.5697	11.6505	18000	0.4340	0.3333	0.1010
0.5501	11.7799	18200	0.4476	0.3372	0.1016
0.5557	11.9094	18400	0.4474	0.3315	0.1006
0.5543	12.0388	18600	0.4251	0.3365	0.1024
0.7196	12.1683	18800	0.4364	0.3306	0.1005
0.4728	12.2977	19000	0.4313	0.3295	0.1009
0.5342	12.4272	19200	0.4267	0.3365	0.1017
0.5437	12.5566	19400	0.4339	0.3323	0.1011
0.5251	12.6861	19600	0.4206	0.3332	0.1015
0.4648	12.8155	19800	0.4297	0.3324	0.1004
0.5792	12.9450	20000	0.4347	0.3259	0.0998
0.4869	13.0744	20200	0.4296	0.3256	0.0988
0.5087	13.2039	20400	0.4338	0.3254	0.0991
0.6692	13.3333	20600	0.4212	0.3256	0.0994
0.5254	13.4628	20800	0.4253	0.3228	0.0985
0.5053	13.5922	21000	0.4294	0.3263	0.0989
0.5353	13.7217	21200	0.4273	0.3217	0.0979
0.5015	13.8511	21400	0.4223	0.3280	0.0997
0.5073	13.9806	21600	0.4308	0.3204	0.0983
0.5079	14.1100	21800	0.4306	0.3220	0.0986
0.5243	14.2395	22000	0.4302	0.3218	0.0980
0.4713	14.3689	22200	0.4391	0.3213	0.0981
0.475	14.4984	22400	0.4197	0.3248	0.0992
0.5342	14.6278	22600	0.4215	0.3214	0.0979
0.4778	14.7573	22800	0.4230	0.3223	0.0975
0.5256	14.8867	23000	0.4285	0.3213	0.0975
0.4872	15.0162	23200	0.4250	0.3187	0.0974
0.7514	15.1456	23400	0.4163	0.3283	0.0985
0.5381	15.2751	23600	0.4219	0.3192	0.0968
0.4458	15.4045	23800	0.4266	0.3241	0.0980
0.474	15.5340	24000	0.4292	0.3164	0.0966
0.4808	15.6634	24200	0.4171	0.3228	0.0983
0.4972	15.7929	24400	0.4156	0.3187	0.0970
0.5605	15.9223	24600	0.4204	0.3196	0.0966
0.4825	16.0518	24800	0.4194	0.3209	0.0972
0.5396	16.1812	25000	0.4109	0.3183	0.0967
0.4674	16.3107	25200	0.4227	0.3130	0.0956
0.5416	16.4401	25400	0.4178	0.3147	0.0962
0.575	16.5696	25600	0.4213	0.3153	0.0961
0.4864	16.6990	25800	0.4116	0.3127	0.0953
0.4773	16.8285	26000	0.4151	0.3137	0.0956
0.5163	16.9579	26200	0.4213	0.3130	0.0954
0.4877	17.0874	26400	0.4152	0.3108	0.0951
0.4689	17.2168	26600	0.4075	0.3147	0.0957
0.5275	17.3463	26800	0.4172	0.3127	0.0956
0.5345	17.4757	27000	0.4218	0.3139	0.0955
0.4446	17.6052	27200	0.4206	0.3105	0.0949
0.435	17.7346	27400	0.4152	0.3121	0.0956
0.5809	17.8641	27600	0.4131	0.3107	0.0952
0.4526	17.9935	27800	0.4105	0.3109	0.0950
0.468	18.1230	28000	0.4153	0.3075	0.0946
0.4247	18.2524	28200	0.4174	0.3077	0.0944
0.5555	18.3819	28400	0.4171	0.3095	0.0947
0.4383	18.5113	28600	0.4162	0.3077	0.0942
0.4817	18.6408	28800	0.4103	0.3085	0.0946
0.4931	18.7702	29000	0.4098	0.3091	0.0944
0.73	18.8997	29200	0.4126	0.3060	0.0941
0.4573	19.0291	29400	0.4119	0.3081	0.0945
0.4527	19.1586	29600	0.4124	0.3067	0.0942
0.446	19.2880	29800	0.4130	0.3074	0.0942
0.4757	19.4175	30000	0.4109	0.3070	0.0940
0.5375	19.5469	30200	0.4133	0.3059	0.0939
0.4405	19.6764	30400	0.4141	0.3067	0.0941
0.5613	19.8058	30600	0.4130	0.3072	0.0941
0.4159	19.9353	30800	0.4129	0.3069	0.0941

Framework versions

• Transformers 4.41.2
• Pytorch 2.4.0
• Datasets 2.19.1
• Tokenizers 0.19.1