scenario-KD-PR-MSV-D2_data-cl-cardiff_cl_only_delta-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: 17.1445
• Accuracy: 0.4005
• F1: 0.3981

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: 7777
• 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	1.09	250	11.9607	0.3603	0.3165
13.9882	2.17	500	11.5568	0.375	0.3692
13.9882	3.26	750	11.3711	0.4090	0.4091
11.3923	4.35	1000	11.5830	0.3974	0.3972
11.3923	5.43	1250	11.6052	0.4074	0.4019
9.8435	6.52	1500	11.6147	0.4128	0.4100
9.8435	7.61	1750	12.7909	0.3966	0.3837
8.6495	8.7	2000	11.8204	0.4128	0.4046
8.6495	9.78	2250	12.4905	0.3889	0.3824
7.3906	10.87	2500	13.3672	0.4113	0.4097
7.3906	11.96	2750	15.0103	0.4020	0.3925
6.4797	13.04	3000	14.2666	0.3789	0.3791
6.4797	14.13	3250	14.9095	0.3904	0.3825
5.5862	15.22	3500	14.1357	0.4028	0.3995
5.5862	16.3	3750	15.0102	0.4005	0.3997
5.0029	17.39	4000	15.1376	0.4051	0.4035
5.0029	18.48	4250	15.2697	0.3958	0.3935
4.4091	19.57	4500	15.8735	0.3935	0.3905
4.4091	20.65	4750	15.5799	0.4028	0.4009
3.9734	21.74	5000	16.0068	0.4151	0.4137
3.9734	22.83	5250	15.9701	0.3796	0.3797
3.569	23.91	5500	16.0636	0.3850	0.3783
3.569	25.0	5750	16.2960	0.3904	0.3878
3.2517	26.09	6000	16.5689	0.3897	0.3850
3.2517	27.17	6250	17.1440	0.3858	0.3843
3.1047	28.26	6500	17.0026	0.4074	0.4028
3.1047	29.35	6750	17.1445	0.4005	0.3981

Framework versions

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