dense-gpt2-wikitext-ba16-lr1e-04-dense

This model is a fine-tuned version of openai-community/gpt2 on the wikitext wikitext-103-raw-v1 dataset. It achieves the following results on the evaluation set:

• Loss: 2.7293
• Accuracy: 0.4656

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.2449	0.08	500	3.1233	0.4140
3.1952	0.16	1000	3.0566	0.4234
3.1208	0.24	1500	3.0195	0.4287
3.1156	0.32	2000	2.9971	0.4314
3.0824	0.4	2500	2.9783	0.4339
3.0628	0.48	3000	2.9608	0.4360
3.0411	0.56	3500	2.9438	0.4381
3.0414	0.64	4000	2.9328	0.4392
3.0605	0.72	4500	2.9228	0.4408
3.0287	0.8	5000	2.9153	0.4419
2.9961	0.88	5500	2.9057	0.4420
3.0192	0.96	6000	2.8979	0.4432
2.9678	1.04	6500	2.8916	0.4439
2.9533	1.12	7000	2.8849	0.4444
2.9363	1.2	7500	2.8760	0.4459
2.9718	1.28	8000	2.8733	0.4458
2.9535	1.3600	8500	2.8641	0.4471
2.9188	1.44	9000	2.8575	0.4480
2.9261	1.52	9500	2.8517	0.4490
2.9132	1.6	10000	2.8482	0.4489
2.9444	1.6800	10500	2.8434	0.4494
2.9378	1.76	11000	2.8367	0.4502
2.9475	1.8400	11500	2.8351	0.4510
2.9319	1.92	12000	2.8304	0.4513
2.9338	2.0	12500	2.8276	0.4518
2.8819	2.08	13000	2.8244	0.4516
2.8785	2.16	13500	2.8207	0.4524
2.853	2.24	14000	2.8200	0.4531
2.8795	2.32	14500	2.8169	0.4530
2.8492	2.4	15000	2.8138	0.4536
2.8781	2.48	15500	2.8114	0.4538
2.8634	2.56	16000	2.8084	0.4546
2.8682	2.64	16500	2.8038	0.4544
2.8695	2.7200	17000	2.8015	0.4544
2.8789	2.8	17500	2.7991	0.4551
2.8274	2.88	18000	2.7965	0.4558
2.8529	2.96	18500	2.7947	0.4560
2.8234	3.04	19000	2.7948	0.4557
2.7951	3.12	19500	2.7933	0.4558
2.8365	3.2	20000	2.7913	0.4554
2.8156	3.2800	20500	2.7879	0.4564
2.7865	3.36	21000	2.7883	0.4566
2.8368	3.44	21500	2.7848	0.4570
2.8124	3.52	22000	2.7820	0.4574
2.818	3.6	22500	2.7816	0.4575
2.8121	3.68	23000	2.7804	0.4576
2.8037	3.76	23500	2.7766	0.4577
2.802	3.84	24000	2.7731	0.4584
2.8611	3.92	24500	2.7733	0.4580
2.8149	4.0	25000	2.7747	0.4581
2.7639	4.08	25500	2.7760	0.4584
2.7486	4.16	26000	2.7751	0.4584
2.7613	4.24	26500	2.7693	0.4590
2.7939	4.32	27000	2.7688	0.4595
2.777	4.4	27500	2.7672	0.4598
2.8233	4.48	28000	2.7654	0.4598
2.7838	4.5600	28500	2.7671	0.4603
2.7777	4.64	29000	2.7660	0.4602
2.7736	4.72	29500	2.7645	0.4596
2.7716	4.8	30000	2.7616	0.4600
2.787	4.88	30500	2.7599	0.4605
2.8111	4.96	31000	2.7578	0.4604
2.7401	5.04	31500	2.7611	0.4608
2.7138	5.12	32000	2.7599	0.4613
2.7371	5.2	32500	2.7604	0.4614
2.7458	5.28	33000	2.7570	0.4613
2.7478	5.36	33500	2.7576	0.4615
2.7533	5.44	34000	2.7550	0.4611
2.7667	5.52	34500	2.7525	0.4618
2.7922	5.6	35000	2.7533	0.4622
2.7632	5.68	35500	2.7550	0.4612
2.7091	5.76	36000	2.7517	0.4617
2.7834	5.84	36500	2.7501	0.4622
2.7505	5.92	37000	2.7490	0.4625
2.7423	6.0	37500	2.7508	0.4626
2.6821	6.08	38000	2.7547	0.4616
2.7103	6.16	38500	2.7527	0.4620
2.6849	6.24	39000	2.7485	0.4624
2.7469	6.32	39500	2.7511	0.4625
2.7172	6.4	40000	2.7509	0.4624
2.727	6.48	40500	2.7489	0.4624
2.7301	6.5600	41000	2.7474	0.4629
2.7134	6.64	41500	2.7448	0.4632
2.7299	6.72	42000	2.7430	0.4624
2.7641	6.8	42500	2.7429	0.4632
2.7476	6.88	43000	2.7412	0.4634
2.7158	6.96	43500	2.7402	0.4639
2.6965	7.04	44000	2.7454	0.4634
2.6981	7.12	44500	2.7432	0.4637
2.6896	7.2	45000	2.7469	0.4625
2.6465	7.28	45500	2.7393	0.4638
2.7134	7.36	46000	2.7408	0.4637
2.7305	7.44	46500	2.7444	0.4634
2.6816	7.52	47000	2.7390	0.4641
2.6531	7.6	47500	2.7380	0.4642
2.6938	7.68	48000	2.7416	0.4639
2.731	7.76	48500	2.7396	0.4647
2.712	7.84	49000	2.7367	0.4647
2.7444	7.92	49500	2.7344	0.4645
2.6864	8.0	50000	2.7337	0.4651
2.6703	8.08	50500	2.7403	0.4637
2.6821	8.16	51000	2.7411	0.4642
2.6883	8.24	51500	2.7361	0.4648
2.6857	8.32	52000	2.7389	0.4641
2.6669	8.4	52500	2.7377	0.4648
2.7408	8.48	53000	2.7353	0.4652
2.6846	8.56	53500	2.7387	0.4645
2.6899	8.64	54000	2.7366	0.4656
2.6424	8.72	54500	2.7346	0.4647
2.6663	8.8	55000	2.7323	0.4651
2.717	8.88	55500	2.7322	0.4652
2.6908	8.96	56000	2.7303	0.4654
2.6235	9.04	56500	2.7360	0.4653
2.6426	9.12	57000	2.7359	0.4655
2.6455	9.2	57500	2.7342	0.4654
2.6243	9.28	58000	2.7365	0.4655
2.6507	9.36	58500	2.7358	0.4652
2.649	9.44	59000	2.7341	0.4656
2.6463	9.52	59500	2.7337	0.4652
2.6706	9.6	60000	2.7329	0.4657
2.6536	9.68	60500	2.7339	0.4655
2.6781	9.76	61000	2.7290	0.4663
2.6706	9.84	61500	2.7300	0.4655
2.6771	9.92	62000	2.7314	0.4655
2.6747	10.0	62500	2.7293	0.4656

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1