distilbert-base-uncased-finetuned-lora-qnli

This model is a fine-tuned version of distilbert-base-uncased on the glue dataset. It achieves the following results on the evaluation set:

• Accuracy: 0.8807
• trainable model parameters: 887042
• all model parameters: 67842052
• percentage of trainable model parameters: 1.31%

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-04
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• weight_decay: 0.01
• rank: 16
• lora_alpha: 32
• lora_dropout: 0.1
• num_epochs: 4