jialicheng/unlearn_nlvr2_vilt_salun_10_87

87

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: 0.9299
• Accuracy: 0.7250
• Dt Accuracy: 0.7250
• Df Accuracy: 0.4331
• Unlearn Overall Accuracy: 0.4645
• 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: 87
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Overall Accuracy	Unlearn Overall Accuracy	Time
0.397	1.0	2700	0.5253	0.8723	0.3142	0.3142	None
0.3294	2.0	5400	0.5806	0.7146	0.3785	0.3785	None
0.2695	3.0	8100	0.6696	0.5835	0.4233	0.4233	None
0.2149	4.0	10800	0.7918	0.4999	0.4474	0.4474	None
0.1718	5.0	13500	0.8848	0.4499	0.4606	0.4606	None
0.1478	6.0	16200	0.9299	0.4331	0.4645	0.4645	None

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu118
• Datasets 2.18.0
• Tokenizers 0.15.2