Edit model card

pos_final_xlm_fr

This model is a fine-tuned version of xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1022
  • Precision: 0.9744
  • Recall: 0.9746
  • F1: 0.9745
  • Accuracy: 0.9769

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: 256
  • eval_batch_size: 256
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 1024
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 40.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
No log 0.95 14 3.5537 0.0 0.0 0.0 0.0026
No log 1.95 28 3.4536 0.0153 0.0024 0.0042 0.0049
No log 2.95 42 3.1247 0.2395 0.1816 0.2066 0.2843
No log 3.95 56 2.5988 0.4342 0.3539 0.3900 0.4543
No log 4.95 70 2.0168 0.5125 0.4086 0.4547 0.5148
No log 5.95 84 1.4838 0.5959 0.5180 0.5543 0.6086
No log 6.95 98 0.9300 0.7905 0.7619 0.7759 0.7981
No log 7.95 112 0.4874 0.9111 0.9078 0.9094 0.9147
No log 8.95 126 0.2940 0.9372 0.9368 0.9370 0.9396
No log 9.95 140 0.2086 0.9471 0.9482 0.9476 0.9490
No log 10.95 154 0.1688 0.9594 0.9610 0.9602 0.9627
No log 11.95 168 0.1450 0.9624 0.9641 0.9632 0.9659
No log 12.95 182 0.1334 0.9651 0.9669 0.9660 0.9686
No log 13.95 196 0.1213 0.9674 0.9685 0.9679 0.9702
No log 14.95 210 0.1155 0.9684 0.9696 0.9690 0.9718
No log 15.95 224 0.1093 0.9707 0.9712 0.9709 0.9734
No log 16.95 238 0.1059 0.9710 0.9716 0.9713 0.9739
No log 17.95 252 0.1046 0.9711 0.9716 0.9714 0.9740
No log 18.95 266 0.1014 0.9719 0.9724 0.9722 0.9745
No log 19.95 280 0.1003 0.9715 0.9722 0.9718 0.9742
No log 20.95 294 0.0987 0.9724 0.9730 0.9727 0.9751
No log 21.95 308 0.0971 0.9722 0.9728 0.9725 0.9750
No log 22.95 322 0.0968 0.9724 0.9735 0.9730 0.9754
No log 23.95 336 0.0954 0.9728 0.9736 0.9732 0.9756
No log 24.95 350 0.0967 0.9722 0.9731 0.9727 0.9752
No log 25.95 364 0.0965 0.9735 0.9744 0.9739 0.9763
No log 26.95 378 0.0963 0.9725 0.9735 0.9730 0.9757
No log 27.95 392 0.0972 0.9728 0.9738 0.9733 0.9759
No log 28.95 406 0.0987 0.9736 0.9745 0.9740 0.9766
No log 29.95 420 0.0994 0.9737 0.9742 0.9740 0.9764
No log 30.95 434 0.0985 0.9737 0.9741 0.9739 0.9764
No log 31.95 448 0.1022 0.9744 0.9746 0.9745 0.9769
No log 32.95 462 0.1020 0.9740 0.9744 0.9742 0.9767
No log 33.95 476 0.1055 0.9730 0.9738 0.9734 0.9758
No log 34.95 490 0.1068 0.9732 0.9742 0.9737 0.9760
0.6768 35.95 504 0.1085 0.9737 0.9740 0.9739 0.9764
0.6768 36.95 518 0.1088 0.9735 0.9743 0.9739 0.9764
0.6768 37.95 532 0.1100 0.9739 0.9744 0.9742 0.9768
0.6768 38.95 546 0.1107 0.9739 0.9745 0.9742 0.9767
0.6768 39.95 560 0.1115 0.9740 0.9747 0.9744 0.9769

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.12.0
  • Datasets 2.18.0
  • Tokenizers 0.13.2
Downloads last month
5
Inference API
This model can be loaded on Inference API (serverless).

Collection including pranaydeeps/lettuce_pos_fr_xlm