https://github.com/doppeltilde/fine_tune_vision_model
https://github.com/doppeltilde/fine_tune_vision_model
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/doppeltilde/fine_tune_vision_model
- Owner: doppeltilde
- License: mit
- Created: 2026-04-11T11:10:21.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2026-04-18T08:11:01.000Z (3 months ago)
- Last Synced: 2026-04-18T10:10:03.537Z (3 months ago)
- Language: Python
- Homepage: https://www.kaggle.com/code/doppeltilde/
- Size: 142 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Fine-tune EfficientNet
### 1. Fine-tune a pre-trained **TF-EfficientNet-Lite4** model on a custom image dataset.
> [!NOTE]
> It's recommended to use Python 3.11.x
1. Create a `dataset` folder.
2. Create subfolders (example: dog, cat, fish, bird) and populate them.
3. Run `python src/fine_tune_from_base.py`.
When its finished it will return a checkpoint file.
### 2. Create a tflite model
1. Run `python src/convert_to_tflite.py`.
This will return a tflite and labels.txt file.
### 3. Add metadata to tflite model
1. Run `python src/add_metadata_to_tflite.py`.
> [!TIP]
> Fine-tuned models are used in conjunction with the [Vision Classification](https://github.com/doppeltilde/vision_classification) project, utilizing its Mediapipe Image Classification task.
---
_Notice:_ _This project was initally created to be used in-house, as such the
development is first and foremost aligned with the internal requirements._