medsiglip-448-ft-crc100k

This model is a fine-tuned version of google/medsiglip-448 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.2713

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 5
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.3465	0.3556	50	1.3722
1.3101	0.7111	100	1.3356
1.3028	1.064	150	1.2855
1.2389	1.4196	200	1.3003
1.2522	1.7751	250	1.2713

Framework versions

• Transformers 4.55.4
• Pytorch 2.5.1+cu121
• Datasets 4.0.0
• Tokenizers 0.21.0