distil_vanilla

This model is a fine-tuned version of distilbert/distilbert-base-uncased on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 0.9608
• Accuracy: 0.795

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: 64
• 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	Accuracy
No log	1.0	25	0.5529	0.72
No log	2.0	50	0.5398	0.745
No log	3.0	75	0.4987	0.7675
No log	4.0	100	0.5019	0.79
No log	5.0	125	0.5674	0.7875
No log	6.0	150	0.6207	0.8025
No log	7.0	175	0.6747	0.7975
No log	8.0	200	0.6935	0.795
No log	9.0	225	0.7399	0.7975
No log	10.0	250	0.7689	0.81
No log	11.0	275	0.8257	0.79
No log	12.0	300	0.8732	0.785
No log	13.0	325	0.8766	0.79
No log	14.0	350	0.8985	0.8
No log	15.0	375	0.9504	0.78
No log	16.0	400	0.9379	0.79
No log	17.0	425	0.9633	0.7925
No log	18.0	450	0.9582	0.7925
No log	19.0	475	0.9602	0.795
0.1096	20.0	500	0.9608	0.795

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2