vit-Mistral-RSICD-captioning

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

• Loss: 2.0912
• Accuracy: 78.96
• Bleu-1: 0.6450
• Bleu-2: 0.4698
• Bleu-3: 0.3611
• Bleu-4: 0.2892
• Meteor: 0.4753
• Rouge-l: 0.4802
• Cider: 0.8008

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: 8
• 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: 1024
• 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	1.2484	78.47	0.6265	0.4531	0.3491	0.2784	0.4676	0.4716	0.7468
1.1026	2.0	1536	1.3858	78.48	0.6507	0.4778	0.3684	0.2941	0.4930	0.4924	0.8439
0.5912	3.0	2304	1.5799	78.64	0.6491	0.4738	0.3641	0.2916	0.4808	0.4832	0.8186
0.3975	4.0	3072	1.7235	78.97	0.6534	0.4784	0.3695	0.2971	0.4851	0.4892	0.8348
0.3975	5.0	3840	1.8366	78.59	0.6325	0.4578	0.3497	0.2784	0.4712	0.4764	0.7838
0.3025	6.0	4608	1.9921	78.87	0.6385	0.4646	0.3601	0.2908	0.4716	0.4752	0.8024
0.2665	7.0	5376	2.0912	78.96	0.6450	0.4698	0.3611	0.2892	0.4753	0.4802	0.8008

Framework versions

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