paligemma_vqav2
This model is a fine-tuned version of google/paligemma-3b-pt-224 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9150
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: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.2238 | 0.2502 | 100 | 2.5060 |
| 2.3649 | 0.5003 | 200 | 1.6552 |
| 1.7686 | 0.7505 | 300 | 1.3643 |
| 1.6699 | 1.0006 | 400 | 1.2258 |
| 1.469 | 1.2508 | 500 | 1.1457 |
| 1.4727 | 1.5009 | 600 | 1.0778 |
| 1.447 | 1.7511 | 700 | 1.0340 |
| 1.3456 | 2.0013 | 800 | 1.0025 |
| 1.2264 | 2.2514 | 900 | 0.9677 |
| 1.264 | 2.5016 | 1000 | 0.9387 |
| 1.2483 | 2.7517 | 1100 | 0.9150 |
Framework versions
- PEFT 0.13.0
- Transformers 4.46.0.dev0
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0
- Downloads last month
- 78
Model tree for RoyRoyRpy/paligemma_vqav2
Base model
google/paligemma-3b-pt-224