VIT_Captioning

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.0461
• Rouge1: 0.4850
• Rouge2: 0.2566
• Rougel: 0.3589
• Rougelsum: 0.3595

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: 5e-05
• train_batch_size: 8
• 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
• lr_scheduler_warmup_steps: 1024
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
2.4042	1.0	1828	1.7451	0.4622	0.1906	0.3370	0.3422
1.5875	2.0	3656	1.5933	0.4599	0.2060	0.3451	0.3472
1.3882	3.0	5484	1.5322	0.4606	0.2082	0.3422	0.3442
1.2415	4.0	7312	1.5130	0.4687	0.2208	0.3458	0.3476
1.1113	5.0	9140	1.5186	0.4630	0.2146	0.3398	0.3402
0.9671	6.0	10968	1.5683	0.4720	0.2290	0.3517	0.3520
0.8528	7.0	12796	1.6352	0.4704	0.2281	0.3491	0.3496
0.7555	8.0	14624	1.7122	0.4725	0.2305	0.3477	0.3481
0.6567	9.0	16452	1.7814	0.4763	0.2389	0.3537	0.3543
0.5612	10.0	18280	1.8528	0.4777	0.2410	0.3515	0.3516
0.4953	11.0	20108	1.9072	0.4799	0.2487	0.3562	0.3565
0.4445	12.0	21936	1.9503	0.4829	0.2514	0.3571	0.3574
0.3976	13.0	23764	1.9928	0.4834	0.2543	0.3569	0.3573
0.3643	14.0	25592	2.0249	0.4820	0.2520	0.3575	0.3581
0.3263	15.0	27420	2.0461	0.4850	0.2566	0.3589	0.3595

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3