js-fake-bach-epochs20

This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5800
• Accuracy: 0.0015

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: 0.0006058454513356471
• train_batch_size: 16
• eval_batch_size: 32
• seed: 1
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.01
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.2488	1.25	315	0.8243	0.0009
0.8134	2.51	630	0.7738	0.0010
0.7677	3.76	945	0.7396	0.0002
0.7314	5.02	1260	0.7088	0.0006
0.692	6.27	1575	0.6734	0.0009
0.6545	7.53	1890	0.6414	0.0010
0.6175	8.78	2205	0.6071	0.0008
0.5782	10.04	2520	0.5945	0.0017
0.5385	11.29	2835	0.5838	0.0009
0.5026	12.55	3150	0.5722	0.0013
0.4694	13.8	3465	0.5676	0.0011
0.4389	15.06	3780	0.5713	0.0011
0.4083	16.31	4095	0.5761	0.0015
0.389	17.57	4410	0.5790	0.0015
0.3771	18.82	4725	0.5800	0.0015

Framework versions

• Transformers 4.29.1
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3