roberta-petco-filtered_annotated-ctr

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

• Loss: 0.0023
• Mse: 0.0023
• Rmse: 0.0477
• Mae: 0.0361
• R2: 0.4149
• Accuracy: 0.75

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: 10
• eval_batch_size: 10
• 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
0.0137	1.0	24	0.0049	0.0049	0.0702	0.0533	-0.2657	0.55
0.0079	2.0	48	0.0042	0.0042	0.0647	0.0533	-0.0751	0.5
0.0074	3.0	72	0.0026	0.0026	0.0505	0.0388	0.3441	0.6833
0.006	4.0	96	0.0041	0.0041	0.0638	0.0544	-0.0467	0.5167
0.0061	5.0	120	0.0027	0.0027	0.0519	0.0409	0.3082	0.7
0.0054	6.0	144	0.0025	0.0025	0.0503	0.0399	0.3498	0.7
0.0052	7.0	168	0.0038	0.0038	0.0615	0.0469	0.0298	0.5833
0.0074	8.0	192	0.0027	0.0027	0.0522	0.0412	0.3000	0.65
0.0049	9.0	216	0.0028	0.0028	0.0530	0.0392	0.2781	0.7333
0.0052	10.0	240	0.0028	0.0028	0.0526	0.0401	0.2885	0.7
0.0035	11.0	264	0.0033	0.0033	0.0572	0.0438	0.1587	0.7
0.0039	12.0	288	0.0034	0.0034	0.0581	0.0455	0.1340	0.65
0.0031	13.0	312	0.0026	0.0026	0.0512	0.0375	0.3267	0.75
0.0043	14.0	336	0.0023	0.0023	0.0477	0.0361	0.4149	0.75
0.0044	15.0	360	0.0027	0.0027	0.0524	0.0397	0.2944	0.7333
0.0033	16.0	384	0.0024	0.0024	0.0485	0.0356	0.3948	0.7833
0.0031	17.0	408	0.0033	0.0033	0.0575	0.0437	0.1517	0.6667
0.0033	18.0	432	0.0026	0.0026	0.0508	0.0373	0.3370	0.7667
0.0031	19.0	456	0.0033	0.0033	0.0571	0.0447	0.1624	0.6667
0.0035	20.0	480	0.0029	0.0029	0.0538	0.0410	0.2562	0.6667

Framework versions

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