clip-roberta-finetuned_text_short

This model is a fine-tuned version of openai/clip-vit-base-patch32 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

• Loss: 2.9825
• Accuracy: 0.0829

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-07
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.1601	4.0	500	4.1445	0.0017
3.9445	8.0	1000	3.6581	0.0451
3.4566	12.0	1500	3.3360	0.0569
3.263	16.0	2000	3.1920	0.0722
3.1373	20.0	2500	3.0826	0.0738
3.0426	24.0	3000	3.0141	0.0812
2.9973	28.0	3500	2.9825	0.0829

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1