roberta-petco-fullemailbody-ctr

This model is a fine-tuned version of yimiwang/bert-petco-emailbody-ctr on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0032
• Mse: 0.0032
• Rmse: 0.0568
• Mae: 0.0421
• R2: 0.3701

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: 1e-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
0.0073	1.0	15	0.0044	0.0044	0.0660	0.0490	0.1516
0.0064	2.0	30	0.0042	0.0042	0.0647	0.0511	0.1826
0.0056	3.0	45	0.0046	0.0046	0.0681	0.0497	0.0951
0.0041	4.0	60	0.0037	0.0037	0.0611	0.0452	0.2723
0.0043	5.0	75	0.0039	0.0039	0.0625	0.0460	0.2390
0.0034	6.0	90	0.0035	0.0035	0.0592	0.0449	0.3155
0.0031	7.0	105	0.0036	0.0036	0.0597	0.0476	0.3040
0.003	8.0	120	0.0035	0.0035	0.0588	0.0470	0.3255
0.0029	9.0	135	0.0038	0.0038	0.0617	0.0452	0.2582
0.0023	10.0	150	0.0037	0.0037	0.0609	0.0469	0.2767
0.0023	11.0	165	0.0035	0.0035	0.0593	0.0444	0.3135
0.0021	12.0	180	0.0040	0.0040	0.0633	0.0467	0.2179
0.0024	13.0	195	0.0034	0.0034	0.0585	0.0449	0.3327
0.0024	14.0	210	0.0033	0.0033	0.0572	0.0434	0.3620
0.002	15.0	225	0.0032	0.0032	0.0568	0.0421	0.3701
0.0019	16.0	240	0.0034	0.0034	0.0581	0.0426	0.3425
0.002	17.0	255	0.0035	0.0035	0.0594	0.0439	0.3123
0.0016	18.0	270	0.0033	0.0033	0.0573	0.0426	0.3609
0.0018	19.0	285	0.0033	0.0033	0.0573	0.0424	0.3603
0.0017	20.0	300	0.0032	0.0032	0.0570	0.0423	0.3674

Framework versions

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