gpt2_small_summarized

This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.6935
• Accuracy: 0.79
• Precision: 0.2632
• Recall: 0.1515
• F1: 0.1923
• D-index: 1.4571

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1600
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	D-index
No log	1.0	200	0.9451	0.77	0.0667	0.0303	0.0417	1.3827
No log	2.0	400	0.7625	0.81	0.0	0.0	0.0	1.4265
2.1174	3.0	600	0.7145	0.835	0.0	0.0	0.0	1.4607
2.1174	4.0	800	1.0087	0.835	0.0	0.0	0.0	1.4607
0.6744	5.0	1000	0.6728	0.825	0.0	0.0	0.0	1.4471
0.6744	6.0	1200	0.7295	0.725	0.1944	0.2121	0.2029	1.3899
0.6744	7.0	1400	2.0423	0.825	0.3333	0.0606	0.1026	1.4704
0.3582	8.0	1600	2.5923	0.685	0.1591	0.2121	0.1818	1.3332
0.3582	9.0	1800	2.9312	0.605	0.1974	0.4545	0.2752	1.3098
0.1089	10.0	2000	3.0778	0.81	0.0	0.0	0.0	1.4265
0.1089	11.0	2200	3.0158	0.785	0.25	0.1515	0.1887	1.4503
0.1089	12.0	2400	3.0195	0.8	0.3333	0.2121	0.2593	1.4934
0.0376	13.0	2600	3.3198	0.77	0.2593	0.2121	0.2333	1.4525
0.0376	14.0	2800	3.4092	0.77	0.2593	0.2121	0.2333	1.4525
0.0012	15.0	3000	3.5722	0.76	0.2	0.1515	0.1724	1.4157
0.0012	16.0	3200	3.5919	0.775	0.25	0.1818	0.2105	1.4480
0.0012	17.0	3400	3.5835	0.795	0.2778	0.1515	0.1961	1.4639
0.0045	18.0	3600	3.6829	0.785	0.25	0.1515	0.1887	1.4503
0.0045	19.0	3800	3.6905	0.785	0.25	0.1515	0.1887	1.4503
0.0008	20.0	4000	3.6935	0.79	0.2632	0.1515	0.1923	1.4571

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3