Modelpublic

wiorz/gpt2_sm_cv_summarized_4

sourceHugging Facemitupdated 2y agoView on Hugging Face

0likes9downloads

Model Card

gpt2smcvsummarized4

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

—Loss: 2.4906
—Accuracy: 0.743
—Precision: 0.3038
—Recall: 0.2462
—F1: 0.2720
—D-index: 1.4383

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
—trainbatchsize: 16
—evalbatchsize: 16
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—lrschedulerwarmup_steps: 8000
—num_epochs: 20
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	D-index
No log	1.0	250	1.0827	0.673	0.2155	0.2564	0.2342	1.3419
2.5863	2.0	500	0.6038	0.761	0.2381	0.1026	0.1434	1.4124
2.5863	3.0	750	0.5238	0.794	0.3333	0.0564	0.0965	1.4421
0.5205	4.0	1000	0.5206	0.798	0.3158	0.0308	0.0561	1.4384
0.5205	5.0	1250	0.5068	0.802	0.4286	0.0462	0.0833	1.4495
0.4619	6.0	1500	0.5153	0.795	0.4074	0.1128	0.1767	1.4637
0.4619	7.0	1750	0.5246	0.795	0.3611	0.0667	0.1126	1.4472
0.3988	8.0	2000	0.6671	0.797	0.3333	0.0410	0.0731	1.4407
0.3988	9.0	2250	0.6091	0.763	0.35	0.2513	0.2925	1.4680
0.3128	10.0	2500	0.7342	0.759	0.3284	0.2256	0.2675	1.4535
0.3128	11.0	2750	0.8236	0.741	0.3049	0.2564	0.2786	1.4391
0.2198	12.0	3000	0.9349	0.742	0.3297	0.3128	0.3211	1.4601
0.2198	13.0	3250	1.1979	0.764	0.3481	0.2410	0.2848	1.4658
0.1522	14.0	3500	1.3995	0.758	0.3464	0.2718	0.3046	1.4682
0.1522	15.0	3750	2.3304	0.779	0.3143	0.1128	0.1660	1.4414
0.1137	16.0	4000	2.0930	0.762	0.2991	0.1641	0.2119	1.4359
0.1137	17.0	4250	2.6869	0.787	0.3714	0.1333	0.1962	1.4598
0.0904	18.0	4500	2.2347	0.678	0.2818	0.4205	0.3374	1.4071
0.0904	19.0	4750	2.4580	0.752	0.3221	0.2462	0.2791	1.4509
0.0737	20.0	5000	2.4906	0.743	0.3038	0.2462	0.2720	1.4383

Framework versions

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