README.md

metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
model-index:
  - name: RoBERTa-RILE
    results: []

RoBERTa-RILE

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

• Loss: 0.6757
• Accuracy: 0.7350
• F1-micro: 0.7350
• F1-macro: 0.7241
• F1-weighted: 0.7347
• Precision: 0.7350
• Recall: 0.7350

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1-micro	F1-macro	F1-weighted	Precision	Recall
0.7442	1.0	1812	0.6827	0.7120	0.7120	0.7007	0.7126	0.7120	0.7120
0.6447	2.0	3624	0.6618	0.7281	0.7281	0.7169	0.7281	0.7281	0.7281
0.5467	3.0	5436	0.6657	0.7309	0.7309	0.7176	0.7295	0.7309	0.7309
0.5179	4.0	7248	0.6654	0.7346	0.7346	0.7240	0.7345	0.7346	0.7346
0.4787	5.0	9060	0.6757	0.7350	0.7350	0.7241	0.7347	0.7350	0.7350

Framework versions

• Transformers 4.16.2
• Pytorch 1.9.0+cu102
• Datasets 1.8.0
• Tokenizers 0.10.3