scenario-KD-PO-CDF-EN-FROM-EN-D2_data-en-cardiff_eng_only_gamma-jason

This model is a fine-tuned version of haryoaw/scenario-TCR_data-en-cardiff_eng_only2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 26.1249
• Accuracy: 0.3880
• F1: 0.3851

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: 88458
• 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	21.3705	0.3571	0.3491
No log	3.45	200	21.4480	0.3726	0.3483
No log	5.17	300	22.0203	0.3796	0.3543
No log	6.9	400	21.4652	0.3955	0.3889
21.8455	8.62	500	22.0346	0.4105	0.4068
21.8455	10.34	600	22.5203	0.4158	0.4064
21.8455	12.07	700	22.5508	0.3951	0.3893
21.8455	13.79	800	23.1432	0.3889	0.3760
21.8455	15.52	900	23.8503	0.3946	0.3841
15.7725	17.24	1000	24.0330	0.3964	0.3792
15.7725	18.97	1100	24.0211	0.4101	0.4097
15.7725	20.69	1200	25.0036	0.3973	0.3846
15.7725	22.41	1300	25.3511	0.3955	0.3880
15.7725	24.14	1400	25.6258	0.3867	0.3765
11.4934	25.86	1500	25.5123	0.3920	0.3904
11.4934	27.59	1600	25.4662	0.4021	0.3990
11.4934	29.31	1700	26.1249	0.3880	0.3851

Framework versions

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