Ir para o conteúdo
ModelVerificadoExterno
Baixar

Wav2vec2 Xls R 300m Okinoerabu Lr 3e 4

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set: Loss: 0.2382 Cer: 0.0746 Wer: 0.3453

Fontehuggingface.co/larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4/tree/9510da288345c42e85e015e768172ce6443b6350
TipoOtherModelo basefacebook/wav2vec2-xls-r-300mVersãoRevision 9510da2Publicador@larrycmuCnota

Fonte do modelo

Tipo
Other
Tamanho do arquivo
1,18 GB
Modelo base
facebook/wav2vec2-xls-r-300m
Versão
vRevision 9510da2
Atualizado
4 de out. de 2026
Publicador
@larrycmu
Downloads
20
Parâmetros
300M
Tarefas
Vídeo para vídeo
Fonte
Hugging Face

Descrição da fonte

wav2vec2-xls-r-300m-okinoerabu-lr-3e-4

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2382
  • Cer: 0.0746
  • Wer: 0.3453

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: 5e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 32
  • optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 100

Training results

Training LossEpochStepValidation LossCerWer
131.59630.740710046.40790.99991.0
71.00361.481520023.69390.99991.0
41.53792.222230014.31870.99991.0
22.46542.96304007.11780.99991.0
10.98293.70375004.30450.99991.0
8.67484.44446004.18530.99991.0
8.40735.18527004.10680.99991.0
8.26995.92598004.07330.99991.0
8.12016.66679004.03410.99991.0
8.06087.407410004.02570.99991.0
8.06078.148111004.02500.99991.0
8.06838.888912004.01830.99991.0
8.02769.629613004.00600.99991.0
8.067410.370414004.00260.99991.0
7.994411.111115003.98100.99991.0
7.842211.851916003.89680.99991.0
7.807512.592617004.04290.99991.0
7.772313.333318003.94070.99991.0
7.742314.074119003.86310.99991.0
7.624614.814820003.77440.99991.0
7.085215.555621003.02020.92321.0
4.692616.296322001.46900.37700.9971
2.730617.037023000.95290.24110.9055
1.931317.777824000.67340.15140.6376
1.497618.518525000.53980.13030.5314
1.264019.259326000.45330.12100.5166
1.047120.027000.37330.10920.4789
0.894020.740728000.34310.10370.4603
0.789221.481529000.31140.10180.4458
0.751122.222230000.28960.09990.4518
0.697822.963031000.27010.09610.4435
0.651223.703732000.25180.09040.4127
0.635324.444433000.23880.09000.4167
0.587725.185234000.23580.08790.4081
0.558125.925935000.22680.08650.4055
0.559726.666736000.23150.08710.4047
0.514527.407437000.21720.08430.3853
0.504528.148138000.21850.08590.4030
0.475628.888939000.21110.08560.3958
0.485529.629640000.20290.08180.3838
0.450930.370441000.20240.08050.3710
0.440731.111142000.20040.08080.3736
0.421731.851943000.19560.08080.3761
0.412032.592644000.19140.07960.3656
0.387933.333345000.19150.07900.3616
0.395834.074146000.18870.07880.3630
0.400734.814847000.19590.07920.3673
0.398135.555648000.19000.08000.3776
0.365536.296349000.18700.07910.3707
0.378337.037050000.18410.07790.3576
0.348437.777851000.19090.07890.3639
0.347338.518552000.19020.07890.3650
0.357039.259353000.21050.08500.4018
0.312840.054000.18760.07990.3701
0.328840.740755000.18680.07860.3587
0.326641.481556000.19230.07840.3656
0.320942.222257000.20730.08240.3827
0.302942.963058000.20260.08000.3704
0.294143.703759000.20730.08120.3759
0.299944.444460000.19270.07870.3679
0.291045.185261000.20420.08130.3659
0.298645.925962000.19640.07720.3610
0.289446.666763000.20380.07990.3679
0.271447.407464000.19900.07740.3607
0.287148.148165000.20230.07770.3636
0.267348.888966000.19730.07630.3599
0.268849.629667000.19440.07740.3624
0.256150.370468000.20420.07720.3593
0.245551.111169000.19520.07750.3647
0.240351.851970000.19450.07670.3604
0.244052.592671000.19370.07640.3567
0.232153.333372000.20210.07930.3647
0.212654.074173000.20220.07650.3590
0.219454.814874000.19930.07770.3604
0.228655.555675000.19520.07620.3556
0.219256.296376000.21410.07840.3690
0.215357.037077000.20490.07590.3527
0.214857.777878000.20820.07680.3582
0.218758.518579000.20120.07770.3610
0.204259.259380000.20800.07570.3527
0.198160.081000.21310.07900.3673
0.211860.740782000.21440.07800.3604
0.201161.481583000.21560.07730.3564
0.197862.222284000.20950.07590.3545
0.185962.963085000.22940.07880.3604
0.196063.703786000.20880.07620.3533
0.189564.444487000.22430.07950.3650
0.189565.185288000.21740.07500.3527
0.185965.925989000.22500.07920.3619
0.180066.666790000.21290.07670.3522
0.185167.407491000.21390.07680.3542
0.176768.148192000.21210.07460.3479
0.180668.888993000.21520.07540.3510
0.186169.629694000.22010.07510.3522
0.180570.370495000.21340.07600.3513
0.184271.111196000.21960.07650.3539
0.168871.851997000.21920.07630.3547
0.164772.592698000.22150.07460.3467
0.180473.333399000.22160.07430.3482
0.155474.0741100000.22840.07590.3536
0.164774.8148101000.21850.07580.3485
0.165575.5556102000.21680.07570.3545
0.152276.2963103000.22230.07550.3505
0.156677.0370104000.22000.07460.3459
0.154477.7778105000.22790.07550.3493
0.158678.5185106000.22540.07540.3513
0.147179.2593107000.22290.07540.3536
0.153180.0108000.22480.07600.3516
0.153980.7407109000.23300.07600.3487
0.154381.4815110000.22510.07550.3533
0.154682.2222111000.23180.07570.3490
0.145382.9630112000.22940.07580.3539
0.149283.7037113000.23010.07590.3513
0.156984.4444114000.23240.07410.3462
0.147485.1852115000.23350.07510.3482
0.138985.9259116000.23220.07500.3507
0.145086.6667117000.23760.07440.3462
0.144887.4074118000.24130.07530.3490
0.144088.1481119000.23780.07490.3465
0.138488.8889120000.23960.07410.3465
0.148689.6296121000.23750.07490.3473
0.140490.3704122000.23490.07470.3487
0.137791.1111123000.23620.07470.3479
0.142891.8519124000.23440.07490.3476
0.144692.5926125000.23270.07460.3465
0.137893.3333126000.23370.07480.3467
0.133794.0741127000.23570.07440.3447
0.137694.8148128000.23460.07460.3470
0.131195.5556129000.23770.07490.3485
0.131896.2963130000.23700.07500.3465
0.135997.0370131000.23850.07470.3473
0.141897.7778132000.23830.07430.3445
0.136498.5185133000.23830.07470.3450
0.119699.2593134000.23800.07420.3436
0.1342100.0135000.23820.07460.3453

Framework versions

  • Transformers 5.8.1
  • Pytorch 2.11.0+cu128
  • Datasets 3.6.0
  • Tokenizers 0.22.2

Fontes

1 fonte
Hugging FacePrimárioVersão 9510da288345c42e85e015e768172ce6443b6350
huggingface.co/larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4/tree/9510da288345c42e85e015e768172ce6443b6350

Verificado 4 de out.

Artefatos de modelo

1 artefato

model.safetensors

safetensors · 1,18 GB · SHA-256 c1dac920146a…afad · Hugging Face

Baixar

Trechos de fonte

3 trechos
Readme11 de jul.huggingface.co/larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4

--- library_name: transformers license: apache-2.0 base_model: facebook/wav2vec2-xls-r-300m tags: - generated_from_trainer metrics: - wer model-index: - name: wav2vec2-xls-r-300m-okinoerabu-lr-3e-4 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # wav2vec2-xls-r-300m-okinoerabu-lr-3e-4 This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2382 - Cer: 0.0746 - Wer: 0.3453 ## 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: 5e-05 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 32 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 500 - num_epochs: 100 ### Training results | Training Loss | Epoch | Step | Validation Loss | Cer | Wer | |:-------------:|:-------:|:-----:|:---------------:|:------:|:------:| | 131.5963 | 0.7407 | 100 | 46.4079 | 0.9999 | 1.0 | | 71.0036 | 1.4815 | 200 | 23.6939 | 0.9999 | 1.0 | | 41.5379 | 2.2222 | 300 | 14.3187 | 0.9999 | 1.0 | | 22.4654 | 2.9630 | 400 | 7.1178 | 0.9999 | 1.0 | | 10.9829 | 3.7037 | 500 | 4.3045 | 0.9999 | 1.0 | | 8.6748 | 4.4444 | 600 | 4.1853 | 0.9999 | 1.0 | | 8.4073 | 5.1852 | 700 | 4.1068 | 0.9999 | 1.0 | | 8.2699 | 5.9259 | 800 | 4.0733 | 0.9999 | 1.0 | | 8.1201 | 6.6667 | 900 | 4.0341 | 0.9999 | 1.0 | | 8.0608 | 7.4074 | 1000 | 4.0257 | 0.9999 | 1.0 | | 8.0607 | 8.1481 | 1100 | 4.0250 | 0.9999 | 1.0 | | 8.0683 | 8.8889 | 1200 | 4.0183 | 0.9999 | 1.0 | | 8.0276 | 9.6296 | 1300 | 4.0060 | 0.9999 | 1.0 | | 8.0674 | 10.3704 | 1400 | 4.0026 | 0.9999 | 1.0 | | 7.9944 | 11.1111 | 1500 | 3.9810 | 0.9999 | 1.0 | | 7.8422 | 11.8519 | 1600 | 3.8968 | 0.9999 | 1.0 | | 7.8075 | 12.5926 | 1700 | 4.0429 | 0.9999 | 1.0...

Note4 de out.huggingface.co/larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4

Source context: 20 downloads · 0 likes · Pipeline automatic-speech-recognition · Library transformers · Repo larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4

Note17 de jul.huggingface.co/larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4

Source context: 1075 downloads · 0 likes · Pipeline automatic-speech-recognition · Library transformers · Repo larrycmu/wav2vec2-xls-r-300m-okinoerabu-lr-3e-4