metadata
library_name: peft
license: gemma
base_model: google/paligemma-3b-pt-224
tags:
- generated_from_trainer
model-index:
- name: paligemma_vqav2
results: []
paligemma_vqav2
This model is a fine-tuned version of google/paligemma-3b-pt-224 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2568
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-05
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_HF with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.2845 | 0.3137 | 100 | 1.7300 |
| 0.9773 | 0.6275 | 200 | 0.6286 |
| 0.5181 | 0.9412 | 300 | 0.4802 |
| 0.4126 | 1.2549 | 400 | 0.4208 |
| 0.365 | 1.5686 | 500 | 0.3774 |
| 0.3369 | 1.8824 | 600 | 0.3645 |
| 0.3037 | 2.1961 | 700 | 0.3299 |
| 0.2854 | 2.5098 | 800 | 0.3164 |
| 0.2939 | 2.8235 | 900 | 0.3093 |
| 0.2547 | 3.1373 | 1000 | 0.2961 |
| 0.2275 | 3.4510 | 1100 | 0.2943 |
| 0.2456 | 3.7647 | 1200 | 0.2824 |
| 0.2368 | 4.0784 | 1300 | 0.2723 |
| 0.2148 | 4.3922 | 1400 | 0.2733 |
| 0.2118 | 4.7059 | 1500 | 0.2737 |
| 0.1991 | 5.0196 | 1600 | 0.2715 |
| 0.1879 | 5.3333 | 1700 | 0.2657 |
| 0.1841 | 5.6471 | 1800 | 0.2746 |
| 0.1912 | 5.9608 | 1900 | 0.2642 |
| 0.1509 | 6.2745 | 2000 | 0.2964 |
| 0.1818 | 6.5882 | 2100 | 0.2607 |
| 0.1736 | 6.9020 | 2200 | 0.2644 |
| 0.1618 | 7.2157 | 2300 | 0.2663 |
| 0.1563 | 7.5294 | 2400 | 0.2637 |
| 0.159 | 7.8431 | 2500 | 0.2561 |
| 0.1488 | 8.1569 | 2600 | 0.2554 |
| 0.1417 | 8.4706 | 2700 | 0.2589 |
| 0.1329 | 8.7843 | 2800 | 0.2599 |
| 0.1455 | 9.0980 | 2900 | 0.2589 |
| 0.1521 | 9.4118 | 3000 | 0.2576 |
| 0.1334 | 9.7255 | 3100 | 0.2568 |
Framework versions
- PEFT 0.13.2
- Transformers 4.46.2
- Pytorch 2.4.0.post301
- Datasets 3.1.0
- Tokenizers 0.20.3