Modelpublic

alexgshaw/hyperpartisan-classifier

sourceHugging Faceapache-2.0updated 4y agoView on Hugging Face

0likes8downloads

Model Card

hyperpartisan-classifier

This model is a fine-tuned version of distilbert-base-uncased on the hyperpartisannewsdetection dataset. It achieves the following results on the evaluation set:

—Loss: 0.0036
—Accuracy: 0.9988

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
—trainbatchsize: 64
—evalbatchsize: 64
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.1441	0.11	1000	0.1391	0.9453
0.1248	0.21	2000	0.1042	0.9595
0.1027	0.32	3000	0.0913	0.9647
0.0928	0.43	4000	0.0827	0.9688
0.0992	0.53	5000	0.0799	0.9698
0.0881	0.64	6000	0.0710	0.9741
0.078	0.75	7000	0.0640	0.9762
0.0708	0.85	8000	0.0626	0.9764
0.0696	0.96	9000	0.0564	0.9792
0.0586	1.07	10000	0.0516	0.9813
0.0558	1.17	11000	0.0507	0.9815
0.0531	1.28	12000	0.0463	0.9829
0.0585	1.39	13000	0.0468	0.9831
0.0488	1.49	14000	0.0403	0.9854
0.057	1.6	15000	0.0393	0.9865
0.0514	1.71	16000	0.0349	0.9879
0.052	1.81	17000	0.0366	0.9868
0.0572	1.92	18000	0.0300	0.9895
0.0311	2.03	19000	0.0309	0.9893
0.0332	2.13	20000	0.0262	0.9908
0.0396	2.24	21000	0.0250	0.9914
0.0314	2.35	22000	0.0223	0.9924
0.0361	2.45	23000	0.0236	0.9919
0.0289	2.56	24000	0.0197	0.9933
0.0322	2.67	25000	0.0182	0.9939
0.0416	2.77	26000	0.0183	0.9937
0.0273	2.88	27000	0.0159	0.9946
0.0317	2.99	28000	0.0152	0.9949
0.0203	3.09	29000	0.0132	0.9957
0.0182	3.2	30000	0.0146	0.9953
0.0165	3.31	31000	0.0123	0.9961
0.0184	3.41	32000	0.0105	0.9968
0.0208	3.52	33000	0.0103	0.9967
0.0187	3.63	34000	0.0083	0.9973
0.0183	3.73	35000	0.0076	0.9977
0.0258	3.84	36000	0.0073	0.9977
0.0114	3.95	37000	0.0066	0.9979
0.007	4.05	38000	0.0052	0.9983
0.0094	4.16	39000	0.0061	0.9981
0.0106	4.27	40000	0.0053	0.9983
0.0134	4.37	41000	0.0052	0.9984
0.0087	4.48	42000	0.0040	0.9987
0.018	4.59	43000	0.0047	0.9985
0.0118	4.69	44000	0.0041	0.9987
0.012	4.8	45000	0.0038	0.9988
0.0165	4.91	46000	0.0036	0.9988

Framework versions

—Transformers 4.27.0.dev0
—Pytorch 1.13.1+cu116
—Datasets 2.9.0
—Tokenizers 0.13.2