You can deploy this model and start earning money today!

alexziweiwang/
exp15-F01-both

Name: exp15-F01-both
Availability: InStock

transformers

pytorch

wav2vec2

automatic-speech-recognition

generated_from_trainer

endpoints_compatible

Views

No views yet

exp15-F01-both

This model is a fine-tuned version of yongjian/wav2vec2-large-a on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.6149
Wer: 1.0154

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: 0.0001
train_batch_size: 4
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: 1000
num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
45.2628	0.33	500	2.9838	1.0
3.1304	0.67	1000	2.8311	1.0
2.9607	1.0	1500	2.6426	1.0039
2.7429	1.33	2000	2.5365	1.2046
2.5496	1.66	2500	2.2169	1.3050
2.3134	2.0	3000	2.0450	1.3127
2.1189	2.33	3500	1.8677	1.2780
2.0075	2.66	4000	1.7450	1.2703
1.9014	3.0	4500	1.8381	1.2664
1.7246	3.33	5000	1.7980	1.2510
1.6783	3.66	5500	1.7269	1.2510
1.589	3.99	6000	1.5640	1.2664
1.4085	4.33	6500	1.7296	1.2355
1.4126	4.66	7000	1.5208	1.2317
1.3506	4.99	7500	1.6253	1.2317
1.2276	5.33	8000	1.6222	1.2239
1.1842	5.66	8500	1.4836	1.1969
1.1445	5.99	9000	1.5313	1.2046
1.0254	6.32	9500	1.9130	1.2046
1.0214	6.66	10000	1.8944	1.2085
0.9677	6.99	10500	1.9039	1.1853
0.8822	7.32	11000	1.7036	1.1892
0.8824	7.66	11500	1.6062	1.1815
0.8695	7.99	12000	1.7019	1.1853
0.7536	8.32	12500	1.9117	1.1737
0.775	8.66	13000	1.8778	1.1815
0.7409	8.99	13500	1.7534	1.1776
0.7035	9.32	14000	1.9860	1.1853
0.6905	9.65	14500	1.9141	1.1892
0.6536	9.99	15000	1.7848	1.1737
0.6237	10.32	15500	2.0624	1.1544
0.5986	10.65	16000	1.9958	1.1544
0.5838	10.99	16500	1.8005	1.1622
0.5231	11.32	17000	1.5967	1.1351
0.5452	11.65	17500	1.8145	1.1274
0.5446	11.98	18000	2.0214	1.1429
0.4727	12.32	18500	1.8989	1.1313
0.4908	12.65	19000	1.7152	1.1467
0.483	12.98	19500	1.7354	1.1429
0.4455	13.32	20000	1.9493	1.1506
0.4456	13.65	20500	2.0869	1.1197
0.4306	13.98	21000	1.9248	1.1236
0.3827	14.31	21500	1.9245	1.1274
0.4059	14.65	22000	1.9478	1.1313
0.3941	14.98	22500	2.2373	1.1197
0.4094	15.31	23000	2.0268	1.1158
0.3584	15.65	23500	1.9292	1.1313
0.3615	15.98	24000	2.1744	1.0965
0.3564	16.31	24500	2.4167	1.0927
0.3202	16.64	25000	2.6332	1.1081
0.3099	16.98	25500	2.9448	1.1004
0.3126	17.31	26000	2.4662	1.0927
0.3189	17.64	26500	2.3619	1.0772
0.3929	17.98	27000	2.3571	1.0618
0.27	18.31	27500	2.2457	1.0734
0.2664	18.64	28000	2.5133	1.0772
0.2875	18.97	28500	2.2798	1.0618
0.2336	19.31	29000	2.3515	1.0347
0.2597	19.64	29500	2.3072	1.0463
0.2573	19.97	30000	2.1702	1.0425
0.2431	20.31	30500	2.2727	1.0618
0.2362	20.64	31000	2.3082	1.0772
0.2377	20.97	31500	2.5453	1.0734
0.228	21.3	32000	2.6838	1.0618
0.2082	21.64	32500	2.7629	1.0695
0.2041	21.97	33000	2.4433	1.0347
0.2208	22.3	33500	2.2516	1.0463
0.2505	22.64	34000	2.4056	1.0541
0.187	22.97	34500	2.6017	1.0347
0.1987	23.3	35000	2.5061	1.0425
0.1952	23.64	35500	2.4440	1.0463
0.1777	23.97	36000	2.4333	1.0463
0.1981	24.3	36500	2.4327	1.0309
0.1729	24.63	37000	2.4114	1.0309
0.1895	24.97	37500	2.3885	1.0347
0.1766	25.3	38000	2.2978	1.0154
0.1603	25.63	38500	2.3070	1.0039
0.1764	25.97	39000	2.4975	1.0154
0.1502	26.3	39500	2.3422	0.9923
0.1574	26.63	40000	2.5013	1.0077
0.1794	26.96	40500	2.4088	1.0039
0.1481	27.3	41000	2.3456	1.0077
0.1594	27.63	41500	2.4916	1.0154
0.1384	27.96	42000	2.4173	1.0077
0.1649	28.3	42500	2.5922	1.0116
0.145	28.63	43000	2.5461	1.0039
0.1654	28.96	43500	2.5312	1.0039
0.1389	29.29	44000	2.5974	1.0077
0.1592	29.63	44500	2.6050	1.0193
0.1055	29.96	45000	2.6149	1.0154

Framework versions

Transformers 4.23.1
Pytorch 1.12.1+cu113
Datasets 1.18.3
Tokenizers 0.13.2

alexziweiwang/exp15-F01-both

exp15-F01-both

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

alexziweiwang/
exp15-F01-both