Recently, I’ve made some revisions one of my previous project (Hybrid EfficientNet Swin-Transformer), hosted on kaggle; to the overall implementation and training strategies.
You may try
(Recommend kaggle environment for efficient and smooth UX.)
|
|
|
|
|
Some highlights
- Enhanced code quality.
- Applied augmentation from
KerasCV
(ported due totf 2.6
). - Applied
Jax
based augmentation intf.data
API. - Applied Gradient Accumulation techniques, cosine annealed warm -restart learning schedulers, etc.
- Illustrate saving and reloading in
SaveModel
format. - Conversion in
TF-Lite
- Code demonstration on Kaggle, Colab, and Deepnote.
- Quick Inference on Gradio (Huggingface Space).