scenario-NON-KD-PO-COPY-CDF-CL-D2_data-cl-cardiff_cl_only_alpha

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: 5.2297
• Accuracy: 0.4707
• F1: 0.4689

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: 1123
• 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	2.1549	0.4576	0.4554
0.5221	2.17	500	2.1917	0.4630	0.4639
0.5221	3.26	750	2.7826	0.4599	0.4541
0.2081	4.35	1000	3.2368	0.4406	0.4175
0.2081	5.43	1250	3.1572	0.4614	0.4550
0.1395	6.52	1500	3.2183	0.4622	0.4586
0.1395	7.61	1750	3.9808	0.4537	0.4467
0.0856	8.7	2000	4.0962	0.4560	0.4568
0.0856	9.78	2250	4.1215	0.4552	0.4500
0.0646	10.87	2500	4.5642	0.4429	0.4380
0.0646	11.96	2750	4.5945	0.4529	0.4511
0.0437	13.04	3000	4.9790	0.4514	0.4442
0.0437	14.13	3250	4.6107	0.4653	0.4618
0.0415	15.22	3500	4.9568	0.4522	0.4511
0.0415	16.3	3750	4.5385	0.4568	0.4554
0.0283	17.39	4000	5.1431	0.4437	0.4358
0.0283	18.48	4250	4.9139	0.4668	0.4657
0.0233	19.57	4500	5.0528	0.4676	0.4636
0.0233	20.65	4750	5.0386	0.4792	0.4785
0.0194	21.74	5000	5.4248	0.4560	0.4489
0.0194	22.83	5250	5.0333	0.4699	0.4678
0.017	23.91	5500	4.9202	0.4761	0.4747
0.017	25.0	5750	5.2043	0.4684	0.4667
0.0088	26.09	6000	5.1802	0.4630	0.4596
0.0088	27.17	6250	5.1366	0.4707	0.4697
0.0066	28.26	6500	5.2244	0.4691	0.4675
0.0066	29.35	6750	5.2297	0.4707	0.4689

Framework versions

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