Question_Generation_ComQ_onT5base_withDataGen10

This model is a fine-tuned version of t5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3923

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss
0.7546	0.23	1000	0.5504
0.6013	0.47	2000	0.5134
0.557	0.7	3000	0.4834
0.5217	0.94	4000	0.4513
0.449	1.17	5000	0.4519
0.4441	1.41	6000	0.4389
0.439	1.64	7000	0.4322
0.4355	1.88	8000	0.4153
0.3996	2.11	9000	0.4263
0.388	2.35	10000	0.4183
0.3856	2.58	11000	0.4129
0.3782	2.82	12000	0.4101
0.3719	3.05	13000	0.4091
0.3395	3.29	14000	0.4091
0.3453	3.52	15000	0.4053
0.3538	3.76	16000	0.3933
0.3468	3.99	17000	0.3897
0.3128	4.22	18000	0.4035
0.3191	4.46	19000	0.4005
0.322	4.69	20000	0.3944
0.3204	4.93	21000	0.3881
0.302	5.16	22000	0.3951
0.2947	5.4	23000	0.3948
0.3011	5.63	24000	0.3932
0.303	5.87	25000	0.3873
0.2902	6.1	26000	0.3916
0.2777	6.34	27000	0.3940
0.2811	6.57	28000	0.3937
0.2815	6.81	29000	0.3923

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0