Modelpublic

kugler/gbert-large-AmDi.small-synset-classifier

sourceHugging Facemitupdated 1y agoView on Hugging Face

0likes4downloads

Model Card

gbertsynsetclassifieramdismall

This model is a fine-tuned version of deepset/gbert-large on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.6371
—Accuracy: 0.8443
—F1: 0.8414
—Precision: 0.8523
—Recall: 0.8443
—F1 Macro: 0.7742
—Precision Macro: 0.7539
—Recall Macro: 0.8118
—F1 Micro: 0.8443
—Precision Micro: 0.8443
—Recall Micro: 0.8443

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: 32
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 128
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—lrschedulerwarmup_steps: 50
—num_epochs: 10
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	F1 Macro	Precision Macro	Recall Macro	F1 Micro	Precision Micro	Recall Micro
3.1817	0.6483	100	1.7424	0.6200	0.5455	0.5576	0.6200	0.2894	0.3465	0.2954	0.6200	0.6200	0.6200
1.0711	1.2966	200	0.7171	0.8140	0.7971	0.7992	0.8140	0.5958	0.5870	0.6238	0.8140	0.8140	0.8140
0.649	1.9449	300	0.6003	0.8275	0.8184	0.8282	0.8275	0.6797	0.6812	0.7138	0.8275	0.8275	0.8275
0.4903	2.5932	400	0.5668	0.8336	0.8268	0.8375	0.8336	0.6942	0.6869	0.7271	0.8336	0.8336	0.8336
0.4095	3.2415	500	0.5511	0.8387	0.8351	0.8398	0.8387	0.7224	0.7198	0.7414	0.8387	0.8387	0.8387
0.3586	3.8898	600	0.5313	0.8415	0.8360	0.8452	0.8415	0.7188	0.7075	0.7481	0.8415	0.8415	0.8415
0.2813	4.5381	700	0.5442	0.8485	0.8451	0.8502	0.8485	0.7290	0.7355	0.7419	0.8485	0.8485	0.8485
0.2543	5.1864	800	0.5736	0.8494	0.8461	0.8515	0.8494	0.7812	0.7708	0.8047	0.8494	0.8494	0.8494
0.1928	5.8347	900	0.5791	0.8448	0.8419	0.8484	0.8448	0.7646	0.7536	0.7899	0.8448	0.8448	0.8448
0.1645	6.4830	1000	0.6371	0.8443	0.8414	0.8523	0.8443	0.7742	0.7539	0.8118	0.8443	0.8443	0.8443

Framework versions

—Transformers 4.45.2
—Pytorch 2.3.1+cu121
—Datasets 2.20.0
—Tokenizers 0.20.3