npc-dialogue-generator-gpt-2

This model is a fine-tuned version of openai-community/gpt2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0356
• Precision: 0.8420
• Recall: 0.8635
• F1: 0.8526

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: 0.0005
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 2000
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
15.3736	0.1970	1000	2.4567	0.8188	0.8493	0.8336
2.5231	0.3939	2000	2.2844	0.8238	0.8527	0.8379
2.3796	0.5909	3000	2.1923	0.8357	0.8572	0.8463
2.3096	0.7879	4000	2.1543	0.8382	0.8594	0.8486
2.2672	0.9848	5000	2.1211	0.8389	0.8606	0.8496
2.2388	1.1818	6000	2.0999	0.8400	0.8615	0.8506
2.2202	1.3788	7000	2.0815	0.8407	0.8619	0.8511
2.2023	1.5757	8000	2.0684	0.8413	0.8623	0.8516
2.1897	1.7727	9000	2.0558	0.8415	0.8624	0.8518
2.1808	1.9697	10000	2.0511	0.8411	0.8628	0.8518
2.1736	2.1666	11000	2.0444	0.8416	0.8632	0.8522
2.1646	2.3636	12000	2.0399	0.8418	0.8634	0.8525
2.1646	2.5606	13000	2.0373	0.8420	0.8634	0.8526
2.1613	2.7575	14000	2.0364	0.8419	0.8635	0.8525
2.1593	2.9545	15000	2.0356	0.8420	0.8635	0.8526

Framework versions

• PEFT 0.13.2
• Transformers 4.45.2
• Pytorch 2.5.0+cu124
• Datasets 3.0.2
• Tokenizers 0.20.1