gmra_model_distilbert-base-uncased-distilled-squad_17082023T171709

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

• Loss: 0.4068
• Accuracy: 93.67%
• F1: 0.9524

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	284	0.3408	87.61%	0.7439
0.4678	2.0	568	0.2731	92.27%	0.9301
0.4678	3.0	852	0.2912	92.62%	0.9418
0.1052	4.0	1137	0.3316	93.5%	0.9521
0.1052	5.0	1421	0.3360	93.67%	0.9532
0.0395	6.0	1705	0.3762	93.23%	0.9497
0.0395	7.0	1989	0.3841	93.67%	0.9534
0.0145	8.0	2274	0.4092	92.97%	0.9481
0.0081	9.0	2558	0.4093	93.5%	0.9512
0.0081	9.99	2840	0.4068	93.67%	0.9524

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1
• Datasets 2.14.4
• Tokenizers 0.13.2