xlm-roberta-large-en-15

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

• Loss: 1.8635
• F1: 0.0986

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: 10
• eval_batch_size: 10
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.6633	1.0	2000	0.4470	0.6619
0.4004	2.0	4000	0.4114	0.7385
0.3302	3.0	6000	0.3452	0.7693
0.6865	4.0	8000	1.4687	0.1242
1.5637	5.0	10000	1.7529	0.0
1.5523	6.0	12000	1.8619	0.0986
1.5439	7.0	14000	1.7808	0.0986
1.5551	8.0	16000	1.7549	0.0986
1.5403	9.0	18000	1.8007	0.0986
1.5366	10.0	20000	1.8351	0.0986
1.5324	11.0	22000	1.8882	0.0986
1.5291	12.0	24000	1.8566	0.0986
1.5285	13.0	26000	1.8071	0.0986
1.5273	14.0	28000	1.8546	0.0986
1.5518	15.0	30000	1.6967	0.0
1.5424	16.0	32000	1.8714	0.0965
1.5356	17.0	34000	1.8270	0.0986
1.5324	18.0	36000	1.8352	0.0986
1.5286	19.0	38000	1.8234	0.0986
1.5318	20.0	40000	1.8017	0.0986
1.5285	21.0	42000	1.9042	0.0986
1.528	22.0	44000	1.8549	0.0986
1.5269	23.0	46000	1.8606	0.0986
1.525	24.0	48000	1.8404	0.0986
1.5243	25.0	50000	1.8635	0.0986

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1