scenario-KD-PO-CDF-CL-D2_data-cl-cardiff_cl_only_alpha-jason

This model is a fine-tuned version of haryoaw/scenario-TCR_data-cl-cardiff_cl_only2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 16.1676
• Accuracy: 0.4005
• F1: 0.3988

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: 2222
• 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	11.7802	0.3727	0.3689
13.6353	2.17	500	10.9269	0.4120	0.4013
13.6353	3.26	750	10.7530	0.4066	0.4034
11.1498	4.35	1000	11.4698	0.4059	0.3994
11.1498	5.43	1250	11.1047	0.4182	0.4140
9.4007	6.52	1500	11.6114	0.4028	0.3911
9.4007	7.61	1750	12.1035	0.3935	0.3911
8.1024	8.7	2000	13.1654	0.4090	0.4035
8.1024	9.78	2250	12.8799	0.4020	0.4001
7.094	10.87	2500	12.5580	0.4082	0.3972
7.094	11.96	2750	12.7991	0.4228	0.4214
6.1021	13.04	3000	13.4827	0.3920	0.3906
6.1021	14.13	3250	14.8695	0.4159	0.4135
5.1725	15.22	3500	14.0881	0.4090	0.4081
5.1725	16.3	3750	14.4576	0.3866	0.3782
4.616	17.39	4000	14.4197	0.3819	0.3777
4.616	18.48	4250	15.2137	0.3997	0.3986
4.0484	19.57	4500	15.2744	0.3943	0.3939
4.0484	20.65	4750	15.3068	0.3858	0.3840
3.6271	21.74	5000	16.0691	0.4059	0.4030
3.6271	22.83	5250	15.7583	0.4120	0.4120
3.3297	23.91	5500	15.9934	0.4028	0.4002
3.3297	25.0	5750	16.3662	0.4097	0.4092
3.0762	26.09	6000	16.4914	0.4051	0.4040
3.0762	27.17	6250	16.4281	0.4051	0.4043
2.8363	28.26	6500	16.3025	0.4151	0.4149
2.8363	29.35	6750	16.1676	0.4005	0.3988

Framework versions

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