ft_paligemma2_cwv

This model is a fine-tuned version of google/paligemma2-3b-pt-224 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0617

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: 0.0002
• train_batch_size: 7
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 28
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2092	0.2356	100	0.0749
0.0615	0.4711	200	0.0742
0.0557	0.7067	300	0.0625
0.051	0.9423	400	0.0669
0.0418	1.1767	500	0.0567
0.0384	1.4122	600	0.0690
0.0309	1.6478	700	0.0470
0.0271	1.8834	800	0.0559
0.0201	2.1178	900	0.0498
0.0154	2.3534	1000	0.0617

Framework versions

• PEFT 0.15.1
• Transformers 4.48.0
• Pytorch 2.5.1
• Datasets 3.5.0
• Tokenizers 0.21.1