aoi_clip_high_resolution_concate_fusin_gpt

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: 4.2052
• Accuracy: 0.0731

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: 20
• 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: 100.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.3923	9.9872	3110	3.0972	0.0765
2.1272	19.9743	6220	3.5858	0.0782
1.9949	29.9615	9330	3.6033	0.0785
1.957	39.9486	12440	3.7208	0.0769
1.9313	49.9358	15550	3.8174	0.0759
1.9255	59.9229	18660	3.9145	0.0748
1.9179	69.9101	21770	4.0367	0.0746
1.9133	79.8972	24880	4.0690	0.0740
1.9102	89.8844	27990	4.1227	0.0737
1.9084	99.8715	31100	4.2052	0.0734

Framework versions

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