scenario-KD-PR-CDF-EN-FROM-CL-D2_data-en-cardiff_eng_only_beta-jason

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

• Loss: 15.8985
• Accuracy: 0.4109
• F1: 0.4052

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: 6666
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.72	100	12.7770	0.3289	0.2140
No log	3.45	200	12.5133	0.3611	0.3597
No log	5.17	300	12.8380	0.3854	0.3580
No log	6.9	400	12.8508	0.4061	0.4051
13.1287	8.62	500	13.5318	0.4017	0.3945
13.1287	10.34	600	13.1743	0.3973	0.3971
13.1287	12.07	700	13.8236	0.4114	0.4114
13.1287	13.79	800	14.0940	0.3854	0.3717
13.1287	15.52	900	14.3671	0.4008	0.3905
9.0164	17.24	1000	14.6749	0.4078	0.3961
9.0164	18.97	1100	14.8097	0.3924	0.3850
9.0164	20.69	1200	15.4684	0.3867	0.3827
9.0164	22.41	1300	15.2537	0.4078	0.4008
9.0164	24.14	1400	16.0142	0.4039	0.3995
6.3951	25.86	1500	15.7916	0.4105	0.4106
6.3951	27.59	1600	16.2415	0.3827	0.3681
6.3951	29.31	1700	15.8985	0.4109	0.4052

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3