bert-petco-filtered_fontsize-ctr

This model is a fine-tuned version of google-bert/bert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0022
• Mse: 0.0022
• Rmse: 0.0464
• Mae: 0.0364
• R2: 0.4468
• Accuracy: 0.8

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
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mse	Rmse	Mae	R2	Accuracy
0.0403	1.0	15	0.0119	0.0119	0.1091	0.0921	-2.0564	0.3
0.0167	2.0	30	0.0031	0.0031	0.0555	0.0449	0.2100	0.6
0.0125	3.0	45	0.0037	0.0037	0.0610	0.0514	0.0429	0.5167
0.0098	4.0	60	0.0026	0.0026	0.0508	0.0407	0.3377	0.6333
0.0105	5.0	75	0.0029	0.0029	0.0543	0.0413	0.2439	0.7
0.0083	6.0	90	0.0029	0.0029	0.0535	0.0392	0.2644	0.7
0.0052	7.0	105	0.0026	0.0026	0.0512	0.0432	0.3280	0.7
0.0044	8.0	120	0.0023	0.0023	0.0482	0.0376	0.4025	0.7667
0.0045	9.0	135	0.0037	0.0037	0.0605	0.0519	0.0601	0.4333
0.0033	10.0	150	0.0028	0.0028	0.0525	0.0367	0.2909	0.7
0.0035	11.0	165	0.0024	0.0024	0.0487	0.0403	0.3918	0.7667
0.003	12.0	180	0.0022	0.0022	0.0471	0.0367	0.4305	0.75
0.0023	13.0	195	0.0022	0.0022	0.0464	0.0364	0.4468	0.8
0.0022	14.0	210	0.0024	0.0024	0.0489	0.0387	0.3863	0.7167
0.0025	15.0	225	0.0024	0.0024	0.0492	0.0379	0.3792	0.75
0.0021	16.0	240	0.0025	0.0025	0.0497	0.0358	0.3667	0.7833
0.0021	17.0	255	0.0024	0.0024	0.0486	0.0366	0.3943	0.7667
0.0017	18.0	270	0.0024	0.0024	0.0485	0.0354	0.3950	0.7833
0.0017	19.0	285	0.0028	0.0028	0.0529	0.0388	0.2819	0.75
0.0018	20.0	300	0.0025	0.0025	0.0496	0.0363	0.3676	0.7667

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2