CFDistilGPT

This model is a fine-tuned version of distilgpt2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.9080

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	140	3.2381
No log	2.0	280	3.1353
No log	3.0	420	3.0809
3.2346	4.0	560	3.0432
3.2346	5.0	700	3.0202
3.2346	6.0	840	2.9960
3.2346	7.0	980	2.9784
2.9216	8.0	1120	2.9666
2.9216	9.0	1260	2.9535
2.9216	10.0	1400	2.9435
2.7856	11.0	1540	2.9330
2.7856	12.0	1680	2.9253
2.7856	13.0	1820	2.9221
2.7856	14.0	1960	2.9184
2.6979	15.0	2100	2.9159
2.6979	16.0	2240	2.9133
2.6979	17.0	2380	2.9096
2.6479	18.0	2520	2.9102
2.6479	19.0	2660	2.9080
2.6479	20.0	2800	2.9080

Framework versions

• Transformers 4.46.1
• Pytorch 2.4.0+cu121
• Datasets 3.0.1
• Tokenizers 0.20.1