Modelpublic

nop1006/gte-base-zh-finetuned-emotion

sourceHugging Facemitupdated 3y agoView on Hugging Face

0likes11downloads

Model Card

gte-base-zh-finetuned-emotion

This model is a fine-tuned version of thenlper/gte-base-zh on the None dataset. It achieves the following results on the evaluation set:

—Loss: 1.3958
—Accuracy: 0.8272
—F1: 0.8189

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
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.4103	1.0	570	0.3675	0.8333	0.8271
0.3452	2.0	1140	0.3796	0.8290	0.8180
0.2784	3.0	1710	0.3930	0.8397	0.8346
0.1904	4.0	2280	0.5113	0.8364	0.8301
0.1239	5.0	2850	0.6590	0.8232	0.8100
0.0828	6.0	3420	0.8153	0.8254	0.8241
0.0624	7.0	3990	0.8672	0.8250	0.8210
0.0413	8.0	4560	0.9244	0.8255	0.8159
0.0303	9.0	5130	1.0888	0.8199	0.8068
0.0233	10.0	5700	1.1171	0.8250	0.8194
0.0159	11.0	6270	1.2642	0.8241	0.8115
0.009	12.0	6840	1.2930	0.8265	0.8169
0.0056	13.0	7410	1.3720	0.8260	0.8150
0.0019	14.0	7980	1.3878	0.8255	0.8168
0.003	15.0	8550	1.3958	0.8272	0.8189

Framework versions

—Transformers 4.38.2
—Pytorch 2.1.0+cu121
—Datasets 2.18.0
—Tokenizers 0.15.2