distilgpt2-finetuned-wellness

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

• Loss: 1.6526

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.0101	1.0	158	2.1220
2.0886	2.0	316	1.9541
1.9941	3.0	474	1.8785
1.8701	4.0	632	1.8243
1.8238	5.0	790	1.7894
1.743	6.0	948	1.7593
1.699	7.0	1106	1.7376
1.6606	8.0	1264	1.7188
1.6345	9.0	1422	1.7027
1.6208	10.0	1580	1.6914
1.5896	11.0	1738	1.6830
1.5694	12.0	1896	1.6722
1.5468	13.0	2054	1.6671
1.5311	14.0	2212	1.6663
1.5172	15.0	2370	1.6602
1.51	16.0	2528	1.6544
1.4916	17.0	2686	1.6555
1.4782	18.0	2844	1.6529
1.4829	19.0	3002	1.6532
1.4639	20.0	3160	1.6526

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu124
• Datasets 3.0.1
• Tokenizers 0.20.0