convnext-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: 0.6620
• Accuracy: 86.07
• Bleu-1: 0.8490
• Bleu-2: 0.7595
• Bleu-3: 0.6866
• Bleu-4: 0.6275
• Meteor: 0.7640
• Rouge-l: 0.7516
• Cider: 1.7522

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
0.9173	1.0	2215	0.8352	84.14	0.6448	0.4677	0.3542	0.2779	0.5099	0.4842	0.4886
0.7188	2.0	4430	0.6649	85.64	0.8234	0.7204	0.6402	0.5760	0.7224	0.7116	1.6629
0.6309	3.0	6645	0.6222	85.03	0.8261	0.7273	0.6482	0.5841	0.7503	0.7312	1.6781
0.5804	4.0	8860	0.6049	85.5	0.8578	0.7753	0.7055	0.6474	0.7733	0.7608	1.7797
0.5463	5.0	11075	0.6012	85.94	0.8624	0.7756	0.7048	0.6465	0.7729	0.7618	1.7969
0.5173	6.0	13290	0.6060	85.8	0.8549	0.7668	0.6951	0.6368	0.7857	0.7651	1.8258
0.5068	7.0	15505	0.6140	85.94	0.8612	0.7775	0.7090	0.6526	0.7882	0.7719	1.8286
0.4875	8.0	17720	0.6223	86.14	0.8648	0.7816	0.7142	0.6588	0.7806	0.7693	1.8333
0.4685	9.0	19935	0.6283	86.12	0.8630	0.7765	0.7036	0.6440	0.7890	0.7705	1.8139
0.4531	10.0	22150	0.6428	86.16	0.8582	0.7712	0.7014	0.6448	0.7713	0.7630	1.8270
0.4381	11.0	24365	0.6506	86.23	0.8631	0.7771	0.7078	0.6512	0.7870	0.7701	1.8459
0.4241	12.0	26580	0.6620	86.07	0.8490	0.7595	0.6866	0.6275	0.7640	0.7516	1.7522

Framework versions

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