You can deploy this model and start earning money today!

tuanio/
fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g1.0-0.05_10_0.004_40

Name: fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g1.0-0.05_10_0.004_40
Availability: InStock

transformers

pytorch

wav2vec2

generated_from_trainer

nguyenvulebinh/wav2vec2-base-vietnamese-250h

finetune

cc-by-nc-4.0

endpoints_compatible

Views

No views yet

fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g1.0-0.05_10_0.004_40

This model is a fine-tuned version of nguyenvulebinh/wav2vec2-base-vietnamese-250h on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.0608
Wer: 0.1012

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: 2e-05
train_batch_size: 16
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 4
total_train_batch_size: 64
total_eval_batch_size: 32
optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
533.9511	0.94	50	256.7028	15.6243
337.9142	1.89	100	82.7848	0.9952
61.5907	2.83	150	23.0722	1.0
28.765	3.77	200	21.1465	1.0
27.4031	4.72	250	20.4765	1.0
26.4393	5.66	300	19.7334	1.0
25.6404	6.6	350	19.2024	1.0
24.6513	7.55	400	18.8785	1.0
24.1773	8.49	450	18.6720	1.0
24.4131	9.43	500	18.5339	1.0
24.0393	10.38	550	18.2257	1.0
23.0434	11.32	600	14.5791	0.7926
15.8664	12.26	650	7.8034	0.4329
9.3173	13.21	700	4.4018	0.2799
6.3563	14.15	750	3.2101	0.2358
5.0252	15.09	800	2.6043	0.2082
4.2585	16.04	850	2.2239	0.1825
3.7273	16.98	900	1.9758	0.1832
3.3365	17.92	950	1.7976	0.1559
3.1388	18.87	1000	1.6666	0.1595
2.8685	19.81	1050	1.5991	0.1669
2.6594	20.75	1100	1.4857	0.1446
2.6179	21.7	1150	1.4291	0.1417
2.5211	22.64	1200	1.4383	0.1519
2.3276	23.58	1250	1.3807	0.1406
2.1864	24.53	1300	1.3551	0.1362
2.2149	25.47	1350	1.2799	0.1335
2.0461	26.42	1400	1.2590	0.1269
2.0038	27.36	1450	1.2624	0.1343
1.9107	28.3	1500	1.2001	0.1253
1.9049	29.25	1550	1.1936	0.1251
1.8057	30.19	1600	1.1986	0.1314
1.7086	31.13	1650	1.1642	0.1225
1.7464	32.08	1700	1.1177	0.1195
1.6634	33.02	1750	1.1247	0.1140
1.6189	33.96	1800	1.1151	0.1114
1.5062	34.91	1850	1.1218	0.1118
1.5323	35.85	1900	1.0949	0.1062
1.5779	36.79	1950	1.0786	0.1104
1.4826	37.74	2000	1.0774	0.1175
1.5034	38.68	2050	1.0891	0.1187
1.4051	39.62	2100	1.0873	0.1229
1.4084	40.57	2150	1.0893	0.1147
1.3231	41.51	2200	1.0818	0.1086
1.3182	42.45	2250	1.0795	0.1145
1.2747	43.4	2300	1.0833	0.1109
1.2657	44.34	2350	1.0797	0.1095
1.2867	45.28	2400	1.0753	0.1053
1.2034	46.23	2450	1.0782	0.1070
1.1649	47.17	2500	1.0685	0.1034
1.1314	48.11	2550	1.0622	0.0979
1.158	49.06	2600	1.0880	0.1032
1.0918	50.0	2650	1.0677	0.1022
1.0786	50.94	2700	1.0708	0.0980
1.1275	51.89	2750	1.0576	0.0989
0.9832	52.83	2800	1.0594	0.1048
1.0832	53.77	2850	1.0528	0.1026
1.0483	54.72	2900	1.0524	0.1072
0.9776	55.66	2950	1.0491	0.1049
0.972	56.6	3000	1.0471	0.1064
1.0257	57.55	3050	1.0680	0.1104
0.9965	58.49	3100	1.0723	0.1157
0.961	59.43	3150	1.0600	0.1040
0.9893	60.38	3200	1.0720	0.1123
0.8888	61.32	3250	1.0598	0.1060
0.9583	62.26	3300	1.0703	0.1057
0.8763	63.21	3350	1.0754	0.1114
0.9151	64.15	3400	1.0769	0.1045
0.8981	65.09	3450	1.0714	0.1014
0.8937	66.04	3500	1.0753	0.1049
0.8897	66.98	3550	1.0775	0.1071
0.8903	67.92	3600	1.0775	0.1039
0.8655	68.87	3650	1.0844	0.1026
0.8845	69.81	3700	1.0831	0.1020
0.854	70.75	3750	1.0877	0.1022
0.8797	71.7	3800	1.0739	0.1007
0.8357	72.64	3850	1.0680	0.1018
0.822	73.58	3900	1.0628	0.1026
0.7696	74.53	3950	1.0585	0.0995
0.8735	75.47	4000	1.0596	0.0994
0.8212	76.42	4050	1.0574	0.0978
0.8341	77.36	4100	1.0632	0.0998
0.8275	78.3	4150	1.0620	0.1019
0.7727	79.25	4200	1.0592	0.1002
0.8182	80.19	4250	1.0523	0.0969
0.813	81.13	4300	1.0672	0.1039
0.7961	82.08	4350	1.0689	0.1014
0.7956	83.02	4400	1.0666	0.0999
0.7853	83.96	4450	1.0648	0.1001
0.8167	84.91	4500	1.0634	0.0986
0.779	85.85	4550	1.0633	0.1012
0.8055	86.79	4600	1.0583	0.1004
0.7847	87.74	4650	1.0618	0.1016
0.7961	88.68	4700	1.0615	0.1016
0.7911	89.62	4750	1.0586	0.1017
0.7098	90.57	4800	1.0613	0.1031
0.8353	91.51	4850	1.0596	0.1015
0.7127	92.45	4900	1.0582	0.1012
0.824	93.4	4950	1.0597	0.1009
0.762	94.34	5000	1.0587	0.1014
0.7424	95.28	5050	1.0596	0.1013
0.7701	96.23	5100	1.0605	0.1011
0.7544	97.17	5150	1.0609	0.1014
0.7844	98.11	5200	1.0606	0.1014
0.7769	99.06	5250	1.0607	0.1012
0.7914	100.0	5300	1.0608	0.1012

Framework versions

Transformers 4.34.0
Pytorch 2.0.1
Datasets 2.14.5
Tokenizers 0.14.1

tuanio/fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g1.0-0.05_10_0.004_40

fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g1.0-0.05_10_0.004_40

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

tuanio/
fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g1.0-0.05_10_0.004_40