xlm-roberta-base-sun-MICRO

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

  • Loss: 0.2865
  • F1: 0.7748
  • Roc Auc: 0.8504
  • Accuracy: 0.6122

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: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 100
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss F1 Roc Auc Accuracy
0.3514 1.0 103 0.3448 0.6745 0.7760 0.5707
0.3815 2.0 206 0.3401 0.6745 0.7760 0.5707
0.3462 3.0 309 0.3261 0.6745 0.7760 0.5707
0.2737 4.0 412 0.2872 0.7345 0.8100 0.6293
0.2769 5.0 515 0.2814 0.7457 0.8167 0.6390
0.2756 6.0 618 0.2670 0.7297 0.8008 0.6049
0.2228 7.0 721 0.2704 0.7489 0.8193 0.6244
0.2371 8.0 824 0.2540 0.7570 0.8204 0.6390
0.2084 9.0 927 0.2684 0.7613 0.8415 0.5902
0.1813 10.0 1030 0.2678 0.7678 0.8415 0.6268
0.1993 11.0 1133 0.2720 0.7654 0.8396 0.6122
0.1738 12.0 1236 0.2725 0.7698 0.8427 0.6122
0.1558 13.0 1339 0.2886 0.7626 0.8381 0.6122
0.1579 14.0 1442 0.2784 0.7624 0.8413 0.6
0.1581 15.0 1545 0.2874 0.7700 0.8493 0.6024
0.1276 16.0 1648 0.2868 0.7697 0.8470 0.6024
0.1325 17.0 1751 0.2865 0.7748 0.8504 0.6122
0.1254 18.0 1854 0.2881 0.7671 0.8440 0.6049
0.1453 19.0 1957 0.2882 0.7695 0.8459 0.6073
0.1161 20.0 2060 0.2881 0.7663 0.8437 0.6024

Framework versions

  • Transformers 4.45.1
  • Pytorch 2.4.0
  • Datasets 3.0.1
  • Tokenizers 0.20.0
Downloads last month
6
Safetensors
Model size
278M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Model tree for sercetexam9/xlm-roberta-base-sun-MICRO

Finetuned
(2874)
this model