bart_base_qgen

This model is a fine-tuned version of facebook/bart-base on the learn3r/squad_with_test dataset. It achieves the following results on the evaluation set:

• Loss: 1.4211

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.7237	1.0	171	1.5064
1.5995	2.0	342	1.4698
1.5289	3.0	513	1.4482
1.5082	4.0	684	1.4363
1.4782	5.0	855	1.4286
1.4084	6.0	1026	1.4265
1.4229	6.99	1197	1.4239
1.4	8.0	1369	1.4211
1.3865	9.0	1540	1.4215
1.3871	9.99	1710	1.4219

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.8.0
• Tokenizers 0.12.1