Hugging Face's logo

statking
/
paligemma_vqa_lower

Image-Text-to-Text
Transformers
TensorBoard
Safetensors
paligemma
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Metrics Community
paligemma_vqa_lower / README.md
statking's picture
statking
Model save
6a68286 verified
preview code
|
raw
history blame
1.97 kB
metadata
license: gemma
base_model: google/paligemma-3b-pt-224
tags:
  - generated_from_trainer
datasets:
  - vq_av2
model-index:
  - name: paligemma_vqa_lower
    results: []

Visualize in Weights & Biases

paligemma_vqa_lower

This model is a fine-tuned version of google/paligemma-3b-pt-224 on the vq_av2 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0131

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: 16
  • eval_batch_size: 16
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 128
  • total_eval_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 1200
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
5.5759 0.1471 500 4.3404
0.2128 0.2943 1000 0.1315
0.0267 0.4414 1500 0.0261
0.0151 0.5886 2000 0.0157
0.0143 0.7357 2500 0.0136
0.0129 0.8829 3000 0.0131

Framework versions

  • Transformers 4.41.0
  • Pytorch 2.2.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1