RobertaLr1e-8Wd0.02E20

This model is a fine-tuned version of deepset/roberta-base-squad2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.2435

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-08
• 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
3.9511	1.0	72	3.9642
2.6869	2.0	144	3.8774
3.0261	3.0	216	3.7963
2.659	4.0	288	3.7220
5.0333	5.0	360	3.6540
4.4433	6.0	432	3.5911
4.8192	7.0	504	3.5341
3.7868	8.0	576	3.4826
2.4543	9.0	648	3.4379
2.4081	10.0	720	3.3981
4.04	11.0	792	3.3631
3.5419	12.0	864	3.3344
2.8603	13.0	936	3.3092
4.191	14.0	1008	3.2884
3.1615	15.0	1080	3.2721
2.8836	16.0	1152	3.2593
4.573	17.0	1224	3.2509
2.5001	18.0	1296	3.2459
3.3777	19.0	1368	3.2438
2.5643	20.0	1440	3.2435

Framework versions

• Transformers 4.41.2
• Pytorch 2.4.0
• Datasets 2.20.0
• Tokenizers 0.19.1