Modelpublic

gokuls/hbertv1-small-wt-48-Massive-intent-emb-comp

sourceHugging Faceupdated 3y agoView on Hugging Face

0likes25downloads

Model Card

hbertv1-small-wt-48-Massive-intent-emb-comp

This model is a fine-tuned version of gokuls/model_v1_complete_training_wt_init_48_small_emb_comp on the massive dataset. It achieves the following results on the evaluation set:

—Loss: 0.8462
—Accuracy: 0.8505

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: 5e-05
—trainbatchsize: 64
—evalbatchsize: 64
—seed: 33
—distributed_type: multi-GPU
—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
2.1467	1.0	180	1.0602	0.7393
0.8554	2.0	360	0.7646	0.7964
0.5593	3.0	540	0.6846	0.8239
0.3868	4.0	720	0.6673	0.8278
0.2613	5.0	900	0.6909	0.8259
0.1681	6.0	1080	0.7123	0.8278
0.1096	7.0	1260	0.7193	0.8318
0.0687	8.0	1440	0.7653	0.8337
0.0405	9.0	1620	0.7966	0.8308
0.0255	10.0	1800	0.8047	0.8441
0.0145	11.0	1980	0.8415	0.8426
0.0092	12.0	2160	0.8462	0.8505
0.0053	13.0	2340	0.8635	0.8465
0.0031	14.0	2520	0.8625	0.8475
0.0023	15.0	2700	0.8632	0.8480

Framework versions

—Transformers 4.31.0
—Pytorch 1.14.0a0+410ce96
—Datasets 2.13.1
—Tokenizers 0.13.3