CoolFace
Modelpublic

Plim/xls-r-300m-lm-fr

sourceHugging Faceupdated 5y agoView on Hugging Face
0likes27downloads
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. -->

This model is a fine-tuned version of ./checkpoint-6000 on the MOZILLA-FOUNDATION/COMMONVOICE7_0 - FR dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.2619
  • —Wer: 0.2457

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: 7.5e-05
  • —trainbatchsize: 16
  • —evalbatchsize: 16
  • —seed: 42
  • —gradientaccumulationsteps: 8
  • —totaltrainbatch_size: 128
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 2000
  • —num_epochs: 2.0
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
3.4950.165003.38831.0
2.90950.3210002.91521.0000
1.84340.4915001.04730.7446
1.42980.6520000.57290.5130
1.19370.8125000.37950.3450
1.12480.9730000.33210.3052
1.08351.1335000.30380.2805
1.04791.340000.29100.2689
1.04131.4645000.27980.2593
1.0141.6250000.27270.2512
1.0041.7855000.26460.2471
0.99491.9460000.26190.2457

Framework versions

  • —Transformers 4.17.0.dev0
  • —Pytorch 1.10.2+cu102
  • —Datasets 1.18.2.dev0
  • —Tokenizers 0.11.0