roberta-large-aces

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

• Loss: 0.5257
• Precision: 0.8561
• Recall: 0.8594
• F1: 0.8553
• Accuracy: 0.8594
• F1 Who: 0.8494
• F1 What: 0.8391
• F1 Where: 0.7558
• F1 How: 0.9208

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy	F1 Who	F1 What	F1 Where	F1 How
0.4619	1.0	87	0.5447	0.8247	0.8416	0.8308	0.8416	0.8309	0.8188	0.6973	0.9244
0.4358	2.0	174	0.4662	0.8522	0.8571	0.8517	0.8571	0.8314	0.8446	0.7613	0.9238
0.3793	3.0	261	0.4892	0.8507	0.8622	0.8556	0.8622	0.8321	0.8418	0.7725	0.9280
0.2875	4.0	348	0.5034	0.8702	0.8641	0.8593	0.8641	0.8471	0.8441	0.7715	0.9225
0.1847	5.0	435	0.5257	0.8561	0.8594	0.8553	0.8594	0.8494	0.8391	0.7558	0.9208

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.1+cu117
• Datasets 2.8.0
• Tokenizers 0.13.2