Edit model card

run-1

This model is a fine-tuned version of roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 2.3480
  • Accuracy: 0.73
  • Precision: 0.6930
  • Recall: 0.6829
  • F1: 0.6871

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: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall F1
1.0042 1.0 50 0.8281 0.665 0.6105 0.6240 0.6016
0.8062 2.0 100 0.9313 0.665 0.6513 0.6069 0.5505
0.627 3.0 150 0.8275 0.72 0.6713 0.6598 0.6638
0.4692 4.0 200 0.8289 0.68 0.6368 0.6447 0.6398
0.2766 5.0 250 1.1263 0.72 0.6893 0.6431 0.6417
0.1868 6.0 300 1.2901 0.725 0.6823 0.6727 0.6764
0.1054 7.0 350 1.6742 0.68 0.6696 0.6427 0.6384
0.0837 8.0 400 1.6199 0.72 0.6826 0.6735 0.6772
0.0451 9.0 450 1.8324 0.735 0.7029 0.6726 0.6727
0.0532 10.0 500 2.1136 0.705 0.6949 0.6725 0.6671
0.0178 11.0 550 2.1136 0.73 0.6931 0.6810 0.6832
0.0111 12.0 600 2.2740 0.69 0.6505 0.6430 0.6461
0.0205 13.0 650 2.3026 0.725 0.6965 0.6685 0.6716
0.0181 14.0 700 2.2901 0.735 0.7045 0.6806 0.6876
0.0074 15.0 750 2.2277 0.74 0.7075 0.6923 0.6978
0.0063 16.0 800 2.2720 0.75 0.7229 0.7051 0.7105
0.0156 17.0 850 2.1237 0.73 0.6908 0.6841 0.6854
0.0027 18.0 900 2.2376 0.73 0.6936 0.6837 0.6874
0.003 19.0 950 2.3359 0.735 0.6992 0.6897 0.6937
0.0012 20.0 1000 2.3480 0.73 0.6930 0.6829 0.6871

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.13.1+cu116
  • Tokenizers 0.13.2
Downloads last month
8
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for SkyR/roberta-base-ours-run-1

Finetuned
(1303)
this model