EfficientNet-B0 fine-tuned on iNaturalist 2019

This model is a fine-tuned EfficientNet-B0 on the iNaturalist 2019 dataset, trained using PyTorch.

Training Results

  • Final Validation Accuracy: 73.1929
  • Total Runtime: 5221.21 seconds

Hyperparameters

  • Batch Size: 256
  • Gradient Accumulation Steps: 4
  • Total Epochs: 10
    • Head-only Epochs: 0
    • Full-body Epochs: 10
  • Weight Decay: 0.01
  • Learning Rates:
    • Head Phase: 0.0
    • Body Phase (Head Params): 0.01
    • Body Phase (Body Params): 0.002

Training was tracked using Weights & Biases.

Downloads last month

-

Downloads are not tracked for this model. How to track
Safetensors
Model size
5.34M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support