brabrant-xvii-ner

This model is a fine-tuned version of FacebookAI/xlm-roberta-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0769
• Precision: 0.7559
• Recall: 0.7996
• F1: 0.7771

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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
No log	1.0	54	0.3302	0.0	0.0	0.0
No log	2.0	108	0.3255	0.0	0.0	0.0
No log	3.0	162	0.1749	0.2513	0.3489	0.2922
No log	4.0	216	0.0689	0.6692	0.7637	0.7134
No log	5.0	270	0.0610	0.7266	0.7776	0.7512
No log	6.0	324	0.0583	0.7556	0.7915	0.7731
No log	7.0	378	0.0585	0.7658	0.7988	0.7820
No log	8.0	432	0.0683	0.7043	0.8069	0.7521
No log	9.0	486	0.0698	0.7442	0.8171	0.7789
0.1553	10.0	540	0.0675	0.7409	0.8178	0.7775
0.1553	11.0	594	0.0689	0.7526	0.7835	0.7677
0.1553	12.0	648	0.0714	0.7372	0.7981	0.7664
0.1553	13.0	702	0.0717	0.7517	0.8018	0.7759
0.1553	14.0	756	0.0765	0.7606	0.8018	0.7806
0.1553	15.0	810	0.0769	0.7559	0.7996	0.7771

Framework versions

• Transformers 4.48.3
• Pytorch 2.1.2
• Datasets 3.3.0
• Tokenizers 0.21.0