videomae-base-finetuned-weightlifting-subset

El siguiente hiperparámetros fueron utilizados durante el entrenamiento:

- learning_rate: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam con betas=(0.9,0.999) y epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 64

Versiones del Framework utilizadas:

• Transformers 4.27.3
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2

Resultados de entrenamiento por epochs:

Epoch 0:
- Training Loss: 0.7656
- Step: 17
- Validation Loss: 0.6768
- Accuracy: 0.6162

Epoch 1:
- Training Loss: 0.6427
- Step: 34
- Validation Loss: 0.6639
- Accuracy: 0.6061

Epoch 2:
- Training Loss: 0.6179
- Step: 51
- Validation Loss: 0.6410
- Accuracy: 0.6128

Epoch 3:
- Training Loss: 0.4667
- Step: 64
- Validation Loss: 0.6279
- Accuracy: 0.6869