Model Card
Training Details
Adam Optimizer with a constant learning rate 1e-5 for 4000 steps training (batch_size=128). Only the vision encoder is fine-tuned.
Evaluation Results
Validation set accuracy:
- pre-trained: 0.6803921461105347
- fine-tuned: 0.8980392217636108
Test set accuracy:
- pre-trained: 0.6644982695579529
- fine-tuned: 0.8855098485946655
- Downloads last month
- 533
Model tree for tanganke/clip-vit-base-patch32_oxford_flowers102
Base model
openai/clip-vit-base-patch32