jialicheng/unlearn_nlvr2_vilt_bad_teaching_6_42

42

This model is a fine-tuned version of dandelin/vilt-b32-finetuned-nlvr2 on the nlvr2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5263
• Accuracy: 0.7355
• Dt Accuracy: 0.7355
• Df Accuracy: 0.8759
• Unlearn Overall Accuracy: 0.6218
• Unlearn Time: None

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• 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
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Overall Accuracy	Unlearn Overall Accuracy	Time
No log	1.0	324	0.5474	0.9126	0.5881	0.5881	None
0.4885	2.0	648	0.5345	0.9118	0.5892	0.5892	None
0.4885	3.0	972	0.5297	0.8911	0.6089	0.6089	None
0.3043	4.0	1296	0.5268	0.8842	0.6147	0.6147	None
0.2354	5.0	1620	0.5263	0.8759	0.6218	0.6218	None

Framework versions

• Transformers 4.48.0
• Pytorch 2.3.0+cu121
• Datasets 2.18.0
• Tokenizers 0.21.0