FULL-8epoch-XLMRoBERTa-finetuned-CEFR_ner-60000news

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

• Loss: 0.0719
• Accuracy: 0.3207
• Precision: 0.6441
• Recall: 0.8378
• F1: 0.5910

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.5383	0.9994	781	0.1595	0.3114	0.7302	0.7636	0.6299
0.1366	2.0	1563	0.1091	0.3169	0.6560	0.8090	0.5905
0.1105	2.9994	2344	0.0876	0.3191	0.6749	0.8239	0.6079
0.0872	4.0	3126	0.0806	0.3198	0.6504	0.8313	0.5937
0.0794	4.9994	3907	0.0758	0.3202	0.6445	0.8345	0.5910
0.0705	6.0	4689	0.0732	0.3206	0.6505	0.8350	0.5942
0.0678	6.9994	5470	0.0724	0.3207	0.6389	0.8381	0.5879
0.0634	7.9949	6248	0.0719	0.3207	0.6441	0.8378	0.5910

Framework versions

• Transformers 4.41.2
• Pytorch 2.2.1
• Datasets 2.19.2
• Tokenizers 0.19.1