wav2vec2-xlsr-53-ft-btb-ccv-cy

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-CLEAN-WITH-CCV - DEFAULT dataset. It achieves the following results on the evaluation set:

• Loss: 0.5067
• Wer: 0.3522

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 30000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.0194	100	3.5677	1.0
No log	0.0387	200	3.0472	1.0
No log	0.0581	300	2.9665	1.0
No log	0.0774	400	2.4643	0.9813
4.1279	0.0968	500	1.6253	0.9345
4.1279	0.1161	600	1.2481	0.8191
4.1279	0.1355	700	1.0997	0.7770
4.1279	0.1549	800	1.0475	0.7340
4.1279	0.1742	900	0.9693	0.7013
1.0598	0.1936	1000	0.9115	0.6749
1.0598	0.2129	1100	0.8824	0.6563
1.0598	0.2323	1200	0.8610	0.6431
1.0598	0.2516	1300	0.8330	0.6114
1.0598	0.2710	1400	0.8173	0.6017
0.8546	0.2904	1500	0.8103	0.6139
0.8546	0.3097	1600	0.7860	0.6078
0.8546	0.3291	1700	0.8576	0.5990
0.8546	0.3484	1800	0.7556	0.5773
0.8546	0.3678	1900	0.7365	0.5826
0.7776	0.3871	2000	0.7292	0.5552
0.7776	0.4065	2100	0.7166	0.5386
0.7776	0.4259	2200	0.7117	0.5402
0.7776	0.4452	2300	0.7061	0.5388
0.7776	0.4646	2400	0.7045	0.5364
0.706	0.4839	2500	0.7063	0.5429
0.706	0.5033	2600	0.6941	0.5434
0.706	0.5226	2700	0.6840	0.5203
0.706	0.5420	2800	0.6902	0.5594
0.706	0.5614	2900	0.6595	0.5149
0.7002	0.5807	3000	0.6768	0.5253
0.7002	0.6001	3100	0.6657	0.5064
0.7002	0.6194	3200	0.6759	0.5409
0.7002	0.6388	3300	0.6709	0.5091
0.7002	0.6581	3400	0.6479	0.5037
0.685	0.6775	3500	0.6378	0.5034
0.685	0.6969	3600	0.6493	0.4988
0.685	0.7162	3700	0.6340	0.4833
0.685	0.7356	3800	0.6227	0.4735
0.685	0.7549	3900	0.6257	0.4907
0.6655	0.7743	4000	0.6420	0.4999
0.6655	0.7937	4100	0.6111	0.4791
0.6655	0.8130	4200	0.6136	0.4807
0.6655	0.8324	4300	0.6218	0.4860
0.6655	0.8517	4400	0.6084	0.4585
0.6431	0.8711	4500	0.6009	0.4628
0.6431	0.8904	4600	0.6010	0.4630
0.6431	0.9098	4700	0.5823	0.4504
0.6431	0.9292	4800	0.6119	0.4630
0.6431	0.9485	4900	0.6002	0.4600
0.6235	0.9679	5000	0.5892	0.4552
0.6235	0.9872	5100	0.5674	0.4489
0.6235	1.0066	5200	0.5792	0.4317
0.6235	1.0259	5300	0.5753	0.4333
0.6235	1.0453	5400	0.5699	0.4462
0.5527	1.0647	5500	0.5667	0.4364
0.5527	1.0840	5600	0.5558	0.4295
0.5527	1.1034	5700	0.5602	0.4223
0.5527	1.1227	5800	0.5591	0.4194
0.5527	1.1421	5900	0.5399	0.4188
0.533	1.1614	6000	0.5459	0.4311
0.533	1.1808	6100	0.5348	0.4118
0.533	1.2002	6200	0.5454	0.4176
0.533	1.2195	6300	0.5443	0.4216
0.533	1.2389	6400	0.5383	0.4096
0.5228	1.2582	6500	0.5407	0.4126
0.5228	1.2776	6600	0.5527	0.4143
0.5228	1.2969	6700	0.5313	0.4081
0.5228	1.3163	6800	0.5339	0.4150
0.5228	1.3357	6900	0.5236	0.4121
0.5204	1.3550	7000	0.5528	0.4166
0.5204	1.3744	7100	0.5331	0.4056
0.5204	1.3937	7200	0.5242	0.4059
0.5204	1.4131	7300	0.5310	0.4093
0.5204	1.4324	7400	0.5278	0.4063
0.5199	1.4518	7500	0.5168	0.3956
0.5199	1.4712	7600	0.5237	0.4024
0.5199	1.4905	7700	0.5316	0.4179
0.5199	1.5099	7800	0.5182	0.4033
0.5199	1.5292	7900	0.5175	0.3984
0.5066	1.5486	8000	0.5138	0.3949
0.5066	1.5679	8100	0.5156	0.4016
0.5066	1.5873	8200	0.5131	0.3979
0.5066	1.6067	8300	0.5140	0.3941
0.5066	1.6260	8400	0.5224	0.3985
0.502	1.6454	8500	0.5275	0.4002
0.502	1.6647	8600	0.5054	0.3861
0.502	1.6841	8700	0.5144	0.3913
0.502	1.7034	8800	0.5018	0.3861
0.502	1.7228	8900	0.5002	0.3998
0.4965	1.7422	9000	0.5075	0.3890
0.4965	1.7615	9100	0.4929	0.3865
0.4965	1.7809	9200	0.4962	0.3856
0.4965	1.8002	9300	0.4904	0.3760
0.4965	1.8196	9400	0.4996	0.3901
0.4776	1.8389	9500	0.4899	0.3762
0.4776	1.8583	9600	0.4918	0.3795
0.4776	1.8777	9700	0.4915	0.3798
0.4776	1.8970	9800	0.4841	0.3706
0.4776	1.9164	9900	0.4834	0.3773
0.4752	1.9357	10000	0.4832	0.3712
0.4752	1.9551	10100	0.4891	0.3783
0.4752	1.9744	10200	0.4787	0.3783
0.4752	1.9938	10300	0.4726	0.3714
0.4752	2.0132	10400	0.4917	0.3732
0.4587	2.0325	10500	0.4802	0.3726
0.4587	2.0519	10600	0.4884	0.3825
0.4587	2.0712	10700	0.4841	0.3785
0.4587	2.0906	10800	0.4809	0.3738
0.4587	2.1099	10900	0.4797	0.3713
0.3967	2.1293	11000	0.4866	0.3750
0.3967	2.1487	11100	0.4938	0.3749
0.3967	2.1680	11200	0.4860	0.3680
0.3967	2.1874	11300	0.4849	0.3700
0.3967	2.2067	11400	0.4908	0.3638
0.406	2.2261	11500	0.4797	0.3679
0.406	2.2455	11600	0.4812	0.3759
0.406	2.2648	11700	0.4704	0.3613
0.406	2.2842	11800	0.4716	0.3635
0.406	2.3035	11900	0.4677	0.3635
0.4088	2.3229	12000	0.4705	0.3640
0.4088	2.3422	12100	0.4782	0.3591
0.4088	2.3616	12200	0.4796	0.3613
0.4088	2.3810	12300	0.4713	0.3558
0.4088	2.4003	12400	0.4763	0.3589
0.407	2.4197	12500	0.4690	0.3565
0.407	2.4390	12600	0.4686	0.3577
0.407	2.4584	12700	0.4677	0.3584
0.407	2.4777	12800	0.4614	0.3577
0.407	2.4971	12900	0.4558	0.3599
0.3855	2.5165	13000	0.4556	0.3565
0.3855	2.5358	13100	0.4601	0.3558
0.3855	2.5552	13200	0.4650	0.3543
0.3855	2.5745	13300	0.4737	0.3548
0.3855	2.5939	13400	0.4506	0.3534
0.3748	2.6132	13500	0.4607	0.3589
0.3748	2.6326	13600	0.4549	0.3537
0.3748	2.6520	13700	0.4563	0.3641
0.3748	2.6713	13800	0.4467	0.3437
0.3748	2.6907	13900	0.4536	0.3545
0.3888	2.7100	14000	0.4504	0.3510
0.3888	2.7294	14100	0.4470	0.3602
0.3888	2.7487	14200	0.4564	0.3539
0.3888	2.7681	14300	0.4521	0.3562
0.3888	2.7875	14400	0.4453	0.3522
0.376	2.8068	14500	0.4552	0.3517
0.376	2.8262	14600	0.4534	0.3550
0.376	2.8455	14700	0.4552	0.3405
0.376	2.8649	14800	0.4556	0.3514
0.376	2.8842	14900	0.4423	0.3468
0.379	2.9036	15000	0.4387	0.3427
0.379	2.9230	15100	0.4373	0.3436
0.379	2.9423	15200	0.4399	0.3387
0.379	2.9617	15300	0.4438	0.3380
0.379	2.9810	15400	0.4370	0.3431
0.3731	3.0004	15500	0.4381	0.3341
0.3731	3.0197	15600	0.4514	0.3286
0.3731	3.0391	15700	0.4378	0.3340
0.3731	3.0585	15800	0.4434	0.3441
0.3731	3.0778	15900	0.4419	0.3399
0.3176	3.0972	16000	0.4408	0.3335
0.3176	3.1165	16100	0.4368	0.3358
0.3176	3.1359	16200	0.4478	0.3401
0.3176	3.1552	16300	0.4414	0.3374
0.3176	3.1746	16400	0.4477	0.3350
0.3201	3.1940	16500	0.4306	0.3292
0.3201	3.2133	16600	0.4535	0.3294
0.3201	3.2327	16700	0.4380	0.3341
0.3201	3.2520	16800	0.4368	0.3325
0.3201	3.2714	16900	0.4360	0.3304
0.3101	3.2907	17000	0.4347	0.3281
0.3101	3.3101	17100	0.4375	0.3285
0.3101	3.3295	17200	0.4492	0.3303
0.3101	3.3488	17300	0.4268	0.3285
0.3101	3.3682	17400	0.4377	0.3270
0.2963	3.3875	17500	0.4249	0.3323
0.2963	3.4069	17600	0.4405	0.3339
0.2963	3.4262	17700	0.4364	0.3286
0.2963	3.4456	17800	0.4351	0.3309
0.2963	3.4650	17900	0.4300	0.3229
0.3062	3.4843	18000	0.4231	0.3252
0.3062	3.5037	18100	0.4326	0.3233
0.3062	3.5230	18200	0.4314	0.3282
0.3062	3.5424	18300	0.4344	0.3289
0.3062	3.5617	18400	0.4266	0.3221
0.2968	3.5811	18500	0.4306	0.3216
0.2968	3.6005	18600	0.4319	0.3239
0.2968	3.6198	18700	0.4271	0.3232
0.2968	3.6392	18800	0.4184	0.3264
0.2968	3.6585	18900	0.4238	0.3200
0.3191	3.6779	19000	0.4139	0.3226
0.3191	3.6973	19100	0.4238	0.3160
0.3191	3.7166	19200	0.4176	0.3193
0.3191	3.7360	19300	0.4196	0.3203
0.3191	3.7553	19400	0.4095	0.3182
0.2921	3.7747	19500	0.4121	0.3167
0.2921	3.7940	19600	0.4113	0.3146
0.2921	3.8134	19700	0.4094	0.3161
0.2921	3.8328	19800	0.4093	0.3139
0.2921	3.8521	19900	0.4112	0.3173
0.3007	3.8715	20000	0.4093	0.3159
0.3007	3.8908	20100	0.4148	0.3157
0.3007	3.9102	20200	0.4114	0.3150
0.3007	3.9295	20300	0.4155	0.3146
0.3007	3.9489	20400	0.4076	0.3136
0.296	3.9683	20500	0.4067	0.3126
0.296	3.9876	20600	0.4084	0.3150
0.296	4.0070	20700	0.4150	0.3124
0.296	4.0263	20800	0.4132	0.3133
0.296	4.0457	20900	0.4183	0.3146
0.2611	4.0650	21000	0.4184	0.3095
0.2611	4.0844	21100	0.4168	0.3085
0.2611	4.1038	21200	0.4224	0.3102
0.2611	4.1231	21300	0.4187	0.3046
0.2611	4.1425	21400	0.4145	0.3110
0.2431	4.1618	21500	0.4272	0.3107
0.2431	4.1812	21600	0.4174	0.3070
0.2431	4.2005	21700	0.4190	0.3086
0.2431	4.2199	21800	0.4164	0.3051
0.2431	4.2393	21900	0.4196	0.3078
0.2453	4.2586	22000	0.4249	0.3092
0.2453	4.2780	22100	0.4246	0.3074
0.2453	4.2973	22200	0.4166	0.3074
0.2453	4.3167	22300	0.4192	0.3028
0.2453	4.3360	22400	0.4186	0.3021
0.2336	4.3554	22500	0.4268	0.3084
0.2336	4.3748	22600	0.4347	0.3071
0.2336	4.3941	22700	0.4753	0.3209
0.2336	4.4135	22800	0.5824	0.4154
0.2336	4.4328	22900	0.5074	0.3415
0.3426	4.4522	23000	0.6242	0.4198
0.3426	4.4715	23100	0.5862	0.4201
0.3426	4.4909	23200	0.6151	0.3964
0.3426	4.5103	23300	0.5640	0.3686
0.3426	4.5296	23400	0.6590	0.4647
0.4541	4.5490	23500	0.6011	0.3960
0.4541	4.5683	23600	0.5803	0.3951
0.4541	4.5877	23700	0.5763	0.3911
0.4541	4.6070	23800	0.5418	0.3655
0.4541	4.6264	23900	0.5547	0.3888
0.4145	4.6458	24000	0.5301	0.3608
0.4145	4.6651	24100	0.5739	0.3993
0.4145	4.6845	24200	0.5776	0.3982
0.4145	4.7038	24300	0.5412	0.3708
0.4145	4.7232	24400	0.5329	0.3704
0.3834	4.7425	24500	0.5299	0.3732
0.3834	4.7619	24600	0.5425	0.3929
0.3834	4.7813	24700	0.5111	0.3585
0.3834	4.8006	24800	0.5076	0.3503
0.3834	4.8200	24900	0.5262	0.3681
0.3719	4.8393	25000	0.5474	0.3833
0.3719	4.8587	25100	0.5747	0.4039
0.3719	4.8780	25200	0.5188	0.3503
0.3719	4.8974	25300	0.5523	0.3866
0.3719	4.9168	25400	0.5302	0.3645
0.3798	4.9361	25500	0.5099	0.3500
0.3798	4.9555	25600	0.4823	0.3376
0.3798	4.9748	25700	0.4806	0.3357
0.3798	4.9942	25800	0.4943	0.3509
0.3798	5.0136	25900	0.4953	0.3525
0.3158	5.0329	26000	0.4853	0.3470
0.3158	5.0523	26100	0.5205	0.3618
0.3158	5.0716	26200	0.5013	0.3510
0.3158	5.0910	26300	0.4863	0.3397
0.3158	5.1103	26400	0.4715	0.3285
0.2993	5.1297	26500	0.4816	0.3327
0.2993	5.1491	26600	0.4806	0.3381
0.2993	5.1684	26700	0.4854	0.3342
0.2993	5.1878	26800	0.4955	0.3433
0.2993	5.2071	26900	0.4863	0.3434
0.2902	5.2265	27000	0.4867	0.3449
0.2902	5.2458	27100	0.4787	0.3378
0.2902	5.2652	27200	0.4862	0.3379
0.2902	5.2846	27300	0.4954	0.3468
0.2902	5.3039	27400	0.5726	0.4142
0.305	5.3233	27500	0.5180	0.3574
0.305	5.3426	27600	0.4997	0.3452
0.305	5.3620	27700	0.4950	0.3413
0.305	5.3813	27800	0.5071	0.3492
0.305	5.4007	27900	0.5096	0.3545
0.3163	5.4201	28000	0.5129	0.3566
0.3163	5.4394	28100	0.5067	0.3506
0.3163	5.4588	28200	0.5053	0.3500
0.3163	5.4781	28300	0.5078	0.3519
0.3163	5.4975	28400	0.4845	0.3375
0.3136	5.5168	28500	0.4930	0.3440
0.3136	5.5362	28600	0.5026	0.3512
0.3136	5.5556	28700	0.5056	0.3519
0.3136	5.5749	28800	0.5091	0.3546
0.3136	5.5943	28900	0.5028	0.3495
0.3092	5.6136	29000	0.5057	0.3510
0.3092	5.6330	29100	0.5086	0.3533
0.3092	5.6523	29200	0.5055	0.3514
0.3092	5.6717	29300	0.5133	0.3577
0.3092	5.6911	29400	0.5130	0.3570
0.3152	5.7104	29500	0.5148	0.3581
0.3152	5.7298	29600	0.5115	0.3555
0.3152	5.7491	29700	0.5054	0.3526
0.3152	5.7685	29800	0.5081	0.3536
0.3152	5.7878	29900	0.5077	0.3535
0.3085	5.8072	30000	0.5067	0.3522

Framework versions

• Transformers 4.40.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1