vit-convnext-Mistral-RSICD-without-captioning

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

• Loss: 1.6517
• Accuracy: 81.25
• Bleu-1: 0.5750
• Bleu-2: 0.4399
• Bleu-3: 0.3481
• Bleu-4: 0.2828
• Meteor: 0.5351
• Rouge-l: 0.4949
• Cider: 1.5140

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: 0.0001
• train_batch_size: 64
• eval_batch_size: 64
• seed: 50
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 0.1
• num_epochs: 128
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Bleu-1	Bleu-2	Bleu-3	Bleu-4	Meteor	Rouge-l	Cider
No log	1.0	768	1.2861	80.42	0.4668	0.3027	0.2115	0.1537	0.4086	0.3976	0.4800
1.5977	2.0	1536	1.2253	79.94	0.6235	0.4835	0.3845	0.3098	0.6194	0.5658	1.5422
0.8479	3.0	2304	1.2063	81.5	0.6909	0.5700	0.4807	0.4098	0.6876	0.6386	2.2467
0.6153	4.0	3072	1.3530	81.26	0.6439	0.5129	0.4201	0.3515	0.6172	0.5724	1.8687
0.6153	5.0	3840	1.4601	80.32	0.6330	0.4999	0.4064	0.3368	0.6227	0.5726	1.7900
0.4260	6.0	4608	1.5262	81.7	0.6488	0.5174	0.4254	0.3566	0.6168	0.5763	1.8769
0.3484	7.0	5376	1.6301	81.62	0.6390	0.5053	0.4111	0.3410	0.6136	0.5643	1.7938
0.3114	8.0	6144	1.6517	81.25	0.5750	0.4399	0.3481	0.2828	0.5351	0.4949	1.5140

Framework versions

• Transformers 5.12.1
• Pytorch 2.12.1+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2