Edit model card

run-5

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

  • Loss: 2.2694
  • Accuracy: 0.745
  • Precision: 0.7091
  • Recall: 0.7017
  • F1: 0.7043

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
0.9558 1.0 50 0.8587 0.665 0.6541 0.6084 0.5787
0.7752 2.0 100 0.8892 0.655 0.6416 0.5835 0.5790
0.5771 3.0 150 0.7066 0.715 0.6884 0.7026 0.6915
0.3738 4.0 200 1.0130 0.705 0.6578 0.6409 0.6455
0.253 5.0 250 1.1405 0.74 0.7132 0.7018 0.7059
0.1604 6.0 300 1.1993 0.69 0.6334 0.6244 0.6261
0.1265 7.0 350 1.5984 0.705 0.6875 0.6775 0.6764
0.0741 8.0 400 1.4755 0.745 0.7116 0.7132 0.7114
0.0505 9.0 450 2.2514 0.71 0.6791 0.6427 0.6524
0.0372 10.0 500 2.2234 0.71 0.6675 0.6503 0.6488
0.0161 11.0 550 2.1070 0.72 0.6783 0.6712 0.6718
0.016 12.0 600 2.0232 0.72 0.6737 0.6659 0.6688
0.0197 13.0 650 2.0224 0.74 0.7065 0.6954 0.6895
0.01 14.0 700 2.1777 0.74 0.7023 0.6904 0.6936
0.0173 15.0 750 2.3227 0.72 0.6761 0.6590 0.6638
0.0066 16.0 800 2.2131 0.735 0.6983 0.6912 0.6923
0.0043 17.0 850 2.1196 0.76 0.7278 0.7207 0.7191
0.0039 18.0 900 2.4087 0.72 0.6791 0.6590 0.6650
0.0041 19.0 950 2.1487 0.73 0.6889 0.6860 0.6873
0.0024 20.0 1000 2.2694 0.745 0.7091 0.7017 0.7043

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.13.1+cu116
  • Tokenizers 0.13.2
Downloads last month
12
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-5

Finetuned
(1300)
this model