Modelpublic

PHILIPPUNI/distilbert-amazon-software-reviews-finetuned

sourceHugging Faceapache-2.0updated 3y agoView on Hugging Face

0likes6downloads

Model Card

results

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

—Loss: 2.2385
—Accuracy: 0.6475
—F1 Score: 0.5149
—Precision Score: 0.5166
—Recall Score: 0.5186

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: 8
—evalbatchsize: 8
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Score	Precision Score	Recall Score
0.8908	1.0	1000	0.8550	0.6775	0.4135	0.5308	0.4387
0.7222	2.0	2000	0.8526	0.68	0.4892	0.5139	0.4846
0.5898	3.0	3000	0.9706	0.659	0.5017	0.5050	0.4995
0.4364	4.0	4000	1.0946	0.669	0.5143	0.5231	0.5074
0.2925	5.0	5000	1.5019	0.6385	0.5190	0.5275	0.5281
0.2378	6.0	6000	1.6785	0.639	0.5095	0.5204	0.5122
0.1715	7.0	7000	1.8847	0.6535	0.5156	0.5163	0.5189
0.1177	8.0	8000	2.1249	0.6425	0.5232	0.5251	0.5309
0.0968	9.0	9000	2.1572	0.659	0.5226	0.5220	0.5288
0.0555	10.0	10000	2.2385	0.6475	0.5149	0.5166	0.5186

Framework versions

—Transformers 4.35.2
—Pytorch 2.1.0+cu121
—Datasets 2.16.1
—Tokenizers 0.15.0