Modelpublic

muhtasham/tiny-mlm-glue-cola-from-scratch-custom-tokenizer-target-glue-qnli

sourceHugging Faceupdated 4y agoView on Hugging Face

0likes27downloads

Model Card

tiny-mlm-glue-cola-from-scratch-custom-tokenizer-target-glue-qnli

This model is a fine-tuned version of muhtasham/tiny-mlm-glue-cola-from-scratch-custom-tokenizer on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.6513
—Accuracy: 0.6129

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: 3e-05
—trainbatchsize: 32
—evalbatchsize: 32
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: constant
—training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6926	0.15	500	0.6887	0.5599
0.6814	0.31	1000	0.6675	0.5867
0.6728	0.46	1500	0.6621	0.5997
0.6665	0.61	2000	0.6609	0.6022
0.6614	0.76	2500	0.6589	0.6028
0.6627	0.92	3000	0.6566	0.6039
0.6552	1.07	3500	0.6562	0.6046
0.659	1.22	4000	0.6533	0.6077
0.6536	1.37	4500	0.6519	0.6114
0.6553	1.53	5000	0.6513	0.6129

Framework versions

—Transformers 4.26.0.dev0
—Pytorch 1.13.0+cu116
—Datasets 2.8.1.dev0
—Tokenizers 0.13.2