multiple_answer_QA

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

• Loss: 5.7334
• Accuracy: 0.2675

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	175	1.3886	0.23
No log	2.0	350	1.3892	0.255
1.3936	3.0	525	1.3863	0.2375
1.3936	4.0	700	1.3868	0.255
1.3936	5.0	875	1.4511	0.26
1.2973	6.0	1050	1.6738	0.24
1.2973	7.0	1225	1.9684	0.265
1.2973	8.0	1400	2.5002	0.28
0.6916	9.0	1575	2.9872	0.28
0.6916	10.0	1750	3.2354	0.28
0.6916	11.0	1925	3.6618	0.3025
0.2428	12.0	2100	4.1750	0.275
0.2428	13.0	2275	4.1384	0.275
0.2428	14.0	2450	4.5173	0.26
0.1118	15.0	2625	4.6013	0.275
0.1118	16.0	2800	4.2549	0.2525
0.1118	17.0	2975	5.2751	0.275
0.0482	18.0	3150	4.9489	0.275
0.0482	19.0	3325	5.6077	0.28
0.0261	20.0	3500	5.3054	0.2625
0.0261	21.0	3675	5.1955	0.2625
0.0261	22.0	3850	5.6210	0.2575
0.0175	23.0	4025	5.6576	0.25
0.0175	24.0	4200	5.6687	0.2725
0.0175	25.0	4375	5.6082	0.255
0.0133	26.0	4550	5.5999	0.27
0.0133	27.0	4725	5.7972	0.26
0.0133	28.0	4900	5.6427	0.2575
0.0089	29.0	5075	5.7253	0.2675
0.0089	30.0	5250	5.7334	0.2675

Framework versions

• Transformers 4.27.4
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3