clip-wikiart-raw-v1-10ep

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

• Loss: 0.2893

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: 5e-06
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.4775	0.3262	500	0.4802
0.4113	0.6523	1000	0.4507
0.4131	0.9785	1500	0.4046
0.2668	1.3046	2000	0.3955
0.2665	1.6308	2500	0.3572
0.2629	1.9569	3000	0.3376
0.1465	2.2831	3500	0.3744
0.1474	2.6093	4000	0.3236
0.1227	2.9354	4500	0.3406
0.0905	3.2616	5000	0.3588
0.0956	3.5877	5500	0.3493
0.0867	3.9139	6000	0.3497
0.0733	4.2401	6500	0.3383
0.0676	4.5662	7000	0.3260
0.0651	4.8924	7500	0.3260
0.0484	5.2185	8000	0.3309
0.0595	5.5447	8500	0.3242
0.0542	5.8708	9000	0.3099
0.0443	6.1970	9500	0.3208
0.0386	6.5232	10000	0.3239
0.0489	6.8493	10500	0.3097
0.0291	7.1755	11000	0.3229
0.0318	7.5016	11500	0.3138
0.0262	7.8278	12000	0.3148
0.0232	8.1539	12500	0.3082
0.0247	8.4801	13000	0.3046
0.0278	8.8063	13500	0.3055
0.0255	9.1324	14000	0.2977
0.0217	9.4586	14500	0.2922
0.0199	9.7847	15000	0.2893

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.1+cu124
• Datasets 4.4.1
• Tokenizers 0.19.1