Modelpublic

paulo037/longformer-schema-linking

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face

0likes5downloads

Model Card

longformer-schema-linking

This model is a fine-tuned version of allenai/longformer-base-4096 on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.1273
—Accuracy: 0.9652
—F1: 0.9354
—Precision: 0.9325
—Recall: 0.9383

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: 32
—evalbatchsize: 64
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 200
—num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.4973	0.2251	400	0.2696	0.9303	0.8552	0.9688	0.7654
0.2865	0.4502	800	0.1839	0.9502	0.9020	0.9583	0.8519
0.2234	0.6753	1200	0.1078	0.9635	0.9337	0.9118	0.9568
0.1829	0.9004	1600	0.1148	0.9635	0.9325	0.9268	0.9383
0.1506	1.1255	2000	0.1339	0.9701	0.9448	0.9390	0.9506
0.1315	1.3506	2400	0.1094	0.9652	0.9362	0.9222	0.9506
0.1208	1.5757	2800	0.1312	0.9635	0.9317	0.9375	0.9259
0.1071	1.8008	3200	0.1273	0.9652	0.9354	0.9325	0.9383

Framework versions

—Transformers 4.47.1
—Pytorch 2.5.1+cu121
—Datasets 3.2.0
—Tokenizers 0.21.0