aoi_clip_high_resolution_concate_fusin

This model is a fine-tuned version of OFA-Sys/chinese-clip-vit-base-patch16 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.6300
• Accuracy: 0.0309

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: 1e-05
• train_batch_size: 40
• eval_batch_size: 40
• seed: 42
• gradient_accumulation_steps: 5
• total_train_batch_size: 200
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 60.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.4434	5.9923	1866	3.7035	0.0316
2.2689	11.9846	3732	3.9282	0.0312
2.1311	17.9769	5598	4.1890	0.0324
2.0473	23.9692	7464	4.2218	0.0317
2.0065	29.9615	9330	4.1968	0.0317
1.9816	35.9538	11196	4.3277	0.0311
1.9593	41.9461	13062	4.4400	0.0312
1.9448	47.9383	14928	4.4896	0.0311
1.9352	53.9306	16794	4.5710	0.0311
1.9342	59.9229	18660	4.6300	0.0310

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1