distilbert_rand_50_v2

This model is a fine-tuned version of on the Hartunka/processed_wikitext-103-raw-v1-rand-50_v2 dataset. It achieves the following results on the evaluation set:

• Loss: 10.1505
• Accuracy: 0.1530

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: 0.0001
• train_batch_size: 96
• eval_batch_size: 96
• seed: 10
• 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
• lr_scheduler_warmup_steps: 10000
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
9.8946	4.1982	10000	10.1297	0.1506
9.0265	8.3963	20000	10.6675	0.1519
7.5938	12.5945	30000	12.1523	0.1529
6.6753	16.7926	40000	13.5060	0.1516
6.3865	20.9908	50000	14.1213	0.1517

Framework versions

• Transformers 4.50.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.21.1