Hugging Face's logo

Dandan0K
/
xls_1b_decoding_fr_decoding_test_iter

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Model card Files Community
xls_1b_decoding_fr_decoding_test_iter / README.md
Dandan0K's picture
Dandan0K
End of training
875c16e verified
preview code
|
raw
history blame
5.23 kB
metadata
license: apache-2.0
base_model: jonatasgrosman/wav2vec2-xls-r-1b-french
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: xls_1b_decoding_fr_decoding_test_iter
    results: []

Visualize in Weights & Biases Visualize in Weights & Biases Visualize in Weights & Biases Visualize in Weights & Biases

xls_1b_decoding_fr_decoding_test_iter

This model is a fine-tuned version of jonatasgrosman/wav2vec2-xls-r-1b-french on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7122
  • Wer: 0.4249

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: 1e-05
  • train_batch_size: 4
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 40
  • num_epochs: 30
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
1.2189 0.6452 40 0.6929 0.6997
0.6612 1.2903 80 0.5628 0.5886
0.5586 1.9355 120 0.4900 0.5202
0.4528 2.5806 160 0.4671 0.4960
0.3799 3.2258 200 0.4555 0.4812
0.3638 3.8710 240 0.4534 0.4686
0.3035 4.5161 280 0.4709 0.4575
0.2905 5.1613 320 0.4640 0.4551
0.2599 5.8065 360 0.4629 0.4444
0.2095 6.4516 400 0.4966 0.4598
0.2206 7.0968 440 0.4958 0.4496
0.1921 7.7419 480 0.4944 0.4389
0.1946 8.3871 520 0.5035 0.4542
0.1629 9.0323 560 0.4978 0.4430
0.15 9.6774 600 0.5143 0.4449
0.1402 10.3226 640 0.5550 0.4351
0.1351 10.9677 680 0.5548 0.4319
0.1212 11.6129 720 0.5455 0.4291
0.1243 12.2581 760 0.5773 0.4300
0.1035 12.9032 800 0.5636 0.4407
0.1103 13.5484 840 0.6062 0.4245
0.0879 14.1935 880 0.5990 0.4384
0.0947 14.8387 920 0.5905 0.4426
0.0804 15.4839 960 0.6118 0.4412
0.0921 16.1290 1000 0.6040 0.4435
0.0816 16.7742 1040 0.6188 0.4170
0.0715 17.4194 1080 0.6463 0.4268
0.0799 18.0645 1120 0.6326 0.4351
0.0631 18.7097 1160 0.6526 0.4314
0.0643 19.3548 1200 0.6502 0.4254
0.0537 20.0 1240 0.6922 0.4310
0.0628 20.6452 1280 0.6778 0.4286
0.0527 21.2903 1320 0.6765 0.4324
0.0566 21.9355 1360 0.6843 0.4249
0.0533 22.5806 1400 0.7073 0.4277
0.052 23.2258 1440 0.7048 0.4296
0.0473 23.8710 1480 0.6886 0.4226
0.0502 24.5161 1520 0.6940 0.4258
0.0496 25.1613 1560 0.6839 0.4240
0.0435 25.8065 1600 0.6931 0.4207
0.0394 26.4516 1640 0.7002 0.4235
0.047 27.0968 1680 0.7086 0.4212
0.0439 27.7419 1720 0.7124 0.4272
0.0375 28.3871 1760 0.7166 0.4245
0.0444 29.0323 1800 0.7149 0.4240
0.0421 29.6774 1840 0.7122 0.4249

Framework versions

  • Transformers 4.43.0.dev0
  • Pytorch 2.3.1+cu118
  • Datasets 2.20.0
  • Tokenizers 0.19.1