gpt2-large-NaturalQuestions_2000-ep20

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

• Loss: 1.5184

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: 2e-05
• train_batch_size: 12
• eval_batch_size: 24
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
1.2377	0.3	50	1.0267
1.0948	0.6	100	0.9831
1.0673	0.9	150	0.9405
0.7058	1.2	200	1.0023
0.5501	1.5	250	1.0274
0.5649	1.8	300	0.9887
0.4972	2.1	350	1.0855
0.2934	2.4	400	1.1109
0.2999	2.69	450	1.0877
0.2907	2.99	500	1.0879
0.1533	3.29	550	1.2041
0.1553	3.59	600	1.1832
0.1681	3.89	650	1.1806
0.107	4.19	700	1.2764
0.0914	4.49	750	1.2541
0.1001	4.79	800	1.2589
0.0816	5.09	850	1.3118
0.0548	5.39	900	1.3473
0.068	5.69	950	1.2907
0.0615	5.99	1000	1.3007
0.0358	6.29	1050	1.4036
0.0409	6.59	1100	1.3794
0.0473	6.89	1150	1.3592
0.0413	7.19	1200	1.4133
0.0438	7.49	1250	1.3690
0.034	7.78	1300	1.3665
0.0307	8.08	1350	1.4268
0.0211	8.38	1400	1.4637
0.0302	8.68	1450	1.4456
0.0288	8.98	1500	1.4584
0.0232	9.28	1550	1.4431
0.017	9.58	1600	1.4756
0.02	9.88	1650	1.4988
0.0207	10.18	1700	1.5071
0.0182	10.48	1750	1.4956
0.0155	10.78	1800	1.5102
0.0163	11.08	1850	1.5207
0.0122	11.38	1900	1.5392
0.0156	11.68	1950	1.5124
0.0149	11.98	2000	1.5184
0.0128	12.28	2050	1.5435
0.0107	12.57	2100	1.5686
0.0118	12.87	2150	1.5301
0.0094	13.17	2200	1.5828
0.0173	13.47	2250	1.5810

Framework versions

• Transformers 4.29.2
• Pytorch 1.10.0+cu111
• Datasets 2.5.1
• Tokenizers 0.13.3