finetuned-indian-food

This model is a fine-tuned version of Salesforce/blip-image-captioning-base on the indian_food_images dataset. It achieves the following results on the evaluation set:

• Loss: 2.1894
• Accuracy: 0.3454

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• 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: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.9029	0.3003	100	2.9256	0.1031
2.8658	0.6006	200	2.7789	0.0967
2.678	0.9009	300	2.6917	0.1838
2.7905	1.2012	400	2.6983	0.1498
2.5964	1.5015	500	2.4903	0.2168
2.4471	1.8018	600	2.5496	0.1987
2.3428	2.1021	700	2.4333	0.2540
2.3367	2.4024	800	2.3813	0.2763
2.2419	2.7027	900	2.3520	0.2965
2.2023	3.0030	1000	2.2766	0.3050
2.2717	3.3033	1100	2.2615	0.3071
2.2311	3.6036	1200	2.2066	0.3284
2.0541	3.9039	1300	2.1894	0.3454

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3