CoolFace
Modelpublic

Alvin-Nahabwe/mms-1b-all-Sagalee-orm-85hrs-4

sourceHugging Facecc-by-nc-4.0updated 2y agoView on Hugging Face
0likes13downloads
Model Card

<!-- 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. -->

mms-1b-all-Sagalee-orm-85hrs-4

This model is a fine-tuned version of facebook/mms-1b-all on an unknown dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.2759
  • —Wer: 0.1432
  • —Cer: 0.0298

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: 3e-05
  • —trainbatchsize: 8
  • —evalbatchsize: 4
  • —seed: 42
  • —optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
  • —lrschedulertype: linear
  • —lrschedulerwarmup_ratio: 0.05
  • —num_epochs: 100
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWerCer
0.8191.056130.22900.29180.0555
0.2442.0112260.21560.26510.0499
0.21773.0168390.21110.25560.0470
0.20664.0224520.21150.24980.0470
0.19885.0280650.20300.25160.0474
0.18616.0336780.19390.24170.0452
0.16927.0392910.19480.23030.0437
0.15078.0449040.18930.21810.0417
0.13339.0505170.19470.21140.0406
0.116410.0561300.19980.20930.0405
0.10111.0617430.18860.19820.0384
0.087812.0673560.18780.19610.0380
0.07713.0729690.20610.19260.0377
0.068914.0785820.21100.19060.0374
0.062515.0841950.21190.18110.0358
0.05716.0898080.22490.17990.0356
0.053817.0954210.22060.18020.0360
0.0518.01010340.22460.17910.0353
0.046719.01066470.22490.17850.0355
0.044520.01122600.22100.17390.0348
0.042121.01178730.22980.17540.0352
0.039822.01234860.23310.16780.0340
0.037623.01290990.22120.16280.0332
0.036224.01347120.23040.16700.0340
0.034325.01403250.23070.16800.0341
0.032526.01459380.24040.16480.0331
0.031327.01515510.23780.16730.0338
0.0328.01571640.24180.16160.0330
0.02929.01627770.24350.16270.0329
0.027830.01683900.24600.16420.0334
0.027531.01740030.24200.16330.0332
0.026232.01796160.25370.16070.0324
0.025933.01852290.24510.15750.0324
0.02534.01908420.24320.15740.0319
0.02435.01964550.25840.16030.0327
0.023136.02020680.25190.16040.0329
0.022437.02076810.25060.16250.0327
0.021538.02132940.24730.15710.0320
0.021839.02189070.25510.15510.0318
0.020540.02245200.25010.15470.0317
0.019841.02301330.25900.15200.0312
0.019142.02357460.25360.15280.0317
0.019243.02413590.24530.15260.0311
0.018444.02469720.25710.15580.0319
0.017845.02525850.24440.15470.0314
0.017246.02581980.25640.15260.0313
0.016847.02638110.25640.15230.0312
0.016148.02694240.26560.14940.0308
0.015449.02750370.25050.15120.0310
0.015150.02806500.26600.14850.0305
0.014551.02862630.24960.14730.0303
0.014352.02918760.25890.14800.0306
0.013453.02974890.26450.14770.0303
0.013354.03031020.25500.15050.0309
0.013355.03087150.27070.14580.0299
0.012556.03143280.26330.14460.0298
0.012457.03199410.26910.14850.0309
0.011958.03255540.26940.14620.0299
0.011659.03311670.27370.14690.0306
0.011260.03367800.26260.14540.0300
0.010861.03423930.26860.14360.0299
0.010662.03480060.27340.14510.0300
0.010763.03536190.26930.14400.0298
0.010164.03592320.27610.14670.0300
0.009865.03648450.27140.14510.0301
0.009466.03704580.27590.14320.0298

Framework versions

  • —Transformers 4.49.0
  • —Pytorch 2.5.1+cu121
  • —Datasets 2.20.0
  • —Tokenizers 0.21.0