You can deploy this model and start earning money today!

Lguyogiro/
corrected_mms-1b-NhiFromAzz-adapterft-ilv_fold1

Name: corrected_mms-1b-NhiFromAzz-adapterft-ilv_fold1
Availability: InStock

tensorboard

safetensors

wav2vec2

generated_from_trainer

audiofolder

Lguyogiro/wav2vec2-large-mms-1b-azz-adapter-ilv-all_data

finetune

cc-by-nc-4.0

model-index

Views

No views yet

corrected_mms-1b-NhiFromAzz-adapterft-ilv_fold1

This model is a fine-tuned version of Lguyogiro/wav2vec2-large-mms-1b-azz-adapter-ilv-all_data on the audiofolder dataset. It achieves the following results on the evaluation set:

Loss: 0.5502
Wer: 0.4105
Cer: 0.1209

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: 100
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.8911	1.6529	200	0.7129	0.6360	0.1941
0.759	3.3058	400	0.6331	0.5892	0.1746
0.7224	4.9587	600	0.6067	0.5560	0.1677
0.6701	6.6116	800	0.5838	0.5206	0.1535
0.6349	8.2645	1000	0.5786	0.5289	0.1531
0.6315	9.9174	1200	0.5660	0.4868	0.1442
0.5995	11.5702	1400	0.5445	0.4826	0.1406
0.587	13.2231	1600	0.5597	0.4912	0.1439
0.5524	14.8760	1800	0.5235	0.4660	0.1377
0.5337	16.5289	2000	0.5227	0.4788	0.1400
0.531	18.1818	2200	0.5305	0.4600	0.1400
0.5088	19.8347	2400	0.5290	0.4632	0.1383
0.5065	21.4876	2600	0.5459	0.4654	0.1370
0.4905	23.1405	2800	0.5212	0.4542	0.1339
0.4812	24.7934	3000	0.5352	0.4386	0.1316
0.4724	26.4463	3200	0.5143	0.4440	0.1333
0.464	28.0992	3400	0.5247	0.4456	0.1325
0.459	29.7521	3600	0.5038	0.4514	0.1310
0.447	31.4050	3800	0.5139	0.4376	0.1288
0.4409	33.0579	4000	0.5164	0.4408	0.1294
0.4271	34.7107	4200	0.5018	0.4469	0.1289
0.4173	36.3636	4400	0.5208	0.4402	0.1300
0.4002	38.0165	4600	0.5242	0.4424	0.1313
0.4032	39.6694	4800	0.5149	0.4389	0.1285
0.3874	41.3223	5000	0.5130	0.4341	0.1264
0.3961	42.9752	5200	0.5228	0.4354	0.1282
0.3686	44.6281	5400	0.5244	0.4380	0.1286
0.3672	46.2810	5600	0.5130	0.4300	0.1266
0.3706	47.9339	5800	0.5150	0.4325	0.1259
0.363	49.5868	6000	0.5161	0.4281	0.1253
0.3481	51.2397	6200	0.5275	0.4252	0.1265
0.3528	52.8926	6400	0.5255	0.4211	0.1258
0.3444	54.5455	6600	0.5270	0.4214	0.1249
0.3415	56.1983	6800	0.5329	0.4201	0.1236
0.3459	57.8512	7000	0.5223	0.4201	0.1242
0.3327	59.5041	7200	0.5368	0.4195	0.1231
0.3226	61.1570	7400	0.5374	0.4159	0.1237
0.3313	62.8099	7600	0.5185	0.4105	0.1221
0.3118	64.4628	7800	0.5506	0.4144	0.1228
0.3108	66.1157	8000	0.5412	0.4093	0.1215
0.3181	67.7686	8200	0.5448	0.4172	0.1254
0.3108	69.4215	8400	0.5449	0.4144	0.1223
0.3076	71.0744	8600	0.5508	0.4099	0.1215
0.3043	72.7273	8800	0.5486	0.4083	0.1198
0.2961	74.3802	9000	0.5379	0.4153	0.1219
0.3011	76.0331	9200	0.5406	0.4105	0.1202
0.28	77.6860	9400	0.5507	0.4144	0.1222
0.2861	79.3388	9600	0.5502	0.4140	0.1214
0.2883	80.9917	9800	0.5611	0.4195	0.1241
0.2798	82.6446	10000	0.5592	0.4144	0.1228
0.2783	84.2975	10200	0.5383	0.4150	0.1230
0.2811	85.9504	10400	0.5455	0.4067	0.1222
0.2716	87.6033	10600	0.5416	0.4102	0.1218
0.2739	89.2562	10800	0.5434	0.4124	0.1215
0.2644	90.9091	11000	0.5411	0.4124	0.1208
0.2633	92.5620	11200	0.5492	0.4105	0.1211
0.2632	94.2149	11400	0.5462	0.4064	0.1197
0.2685	95.8678	11600	0.5485	0.4064	0.1199
0.2682	97.5207	11800	0.5483	0.4093	0.1209
0.261	99.1736	12000	0.5502	0.4105	0.1209

Framework versions

Transformers 4.41.2
Pytorch 2.4.0
Datasets 2.19.1
Tokenizers 0.19.1

Lguyogiro/corrected_mms-1b-NhiFromAzz-adapterft-ilv_fold1

corrected_mms-1b-NhiFromAzz-adapterft-ilv_fold1

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Lguyogiro/
corrected_mms-1b-NhiFromAzz-adapterft-ilv_fold1