swin-Mistral-NWPU-without-captioning

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

• Loss: 0.7019
• Accuracy: 86.37
• Bleu-1: 0.8660
• Bleu-2: 0.7823
• Bleu-3: 0.7136
• Bleu-4: 0.6568
• Meteor: 0.7892
• Rouge-l: 0.7758
• Cider: 1.8386

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: 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
0.6905	1.0	2215	0.6479	86.1	0.8331	0.7379	0.6596	0.5963	0.7332	0.7218	1.6610
0.6036	2.0	4430	0.6193	86.22	0.8629	0.7787	0.7093	0.6520	0.7742	0.7620	1.8648
0.5564	3.0	6645	0.6088	85.62	0.8439	0.7552	0.6853	0.6285	0.7813	0.7607	1.7890
0.5217	4.0	8860	0.6091	86.25	0.8761	0.7970	0.7298	0.6739	0.7949	0.7797	1.8825
0.4930	5.0	11075	0.6110	86.36	0.8715	0.7926	0.7273	0.6730	0.7907	0.7807	1.8995
0.4640	6.0	13290	0.6223	86.29	0.8741	0.7943	0.7280	0.6737	0.8039	0.7862	1.9263
0.4537	7.0	15505	0.6356	86.34	0.8717	0.7911	0.7246	0.6700	0.7985	0.7852	1.8791
0.4302	8.0	17720	0.6526	86.22	0.8683	0.7872	0.7207	0.6654	0.7897	0.7760	1.8573
0.4071	9.0	19935	0.6605	86.38	0.8687	0.7864	0.7184	0.6621	0.7895	0.7777	1.8321
0.3863	10.0	22150	0.6838	86.41	0.8635	0.7800	0.7127	0.6580	0.7855	0.7717	1.8612
0.3677	11.0	24365	0.7019	86.37	0.8660	0.7823	0.7136	0.6568	0.7892	0.7758	1.8386

Framework versions

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