https://github.com/jsflo/emotionrectraining

Python script using Keras + TensorFlow to train a custom machine learning model to recognize faces using a dataset with human faces and labeled with emotions.
https://github.com/jsflo/emotionrectraining

keras tensorflow tensorflow-examples tensorflow-serving

Last synced: 2 months ago
JSON representation

Python script using Keras + TensorFlow to train a custom machine learning model to recognize faces using a dataset with human faces and labeled with emotions.

• Host: GitHub
• URL: https://github.com/jsflo/emotionrectraining
• Owner: JsFlo
• Created: 2018-05-20T17:55:17.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-09-03T16:52:30.000Z (over 6 years ago)
• Last Synced: 2024-09-22T08:30:32.852Z (3 months ago)
• Topics: keras, tensorflow, tensorflow-examples, tensorflow-serving
• Language: Jupyter Notebook
• Homepage:
• Size: 2.21 MB
• Stars: 33
• Watchers: 4
• Forks: 19
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# EmotionRecTraining

Python script using Keras + TensorFlow to train a custom machine learning model

to recognize faces using a dataset with human faces and labeled with emotions.

I wrote an article explaining this script and model architecture and that can be

found here: https://medium.com/@jsflo.dev/training-a-tensorflow-model-to-recognize-emotions-a20c3bcd6468

## Training

Script: emotionRecTrain.py

Required args:

  * `csv_file` - path to the fer2013.csv file

  * `export_path` - path to save the trained model artifacts

Optional args:

  * `batch_size` - batch size during training

  * `n_epochs` - number of training epochs

  * `debug` - Will override script arguments for batch_size and n_epocs to 10 and 1

Running

```

# required

python3 emotionRecTrain.py --csv_file=data/fer2013.csv --export_path=model_out/

python3 emotionRecTrain.py --csv_file=data/fer2013.csv --export_path=model_out/ --batch_size=50

python3 emotionRecTrain.py --csv_file=data/fer2013.csv --export_path=model_out/ --n_epochs=5000

python3 emotionRecTrain.py --csv_file=data/fer2013.csv --export_path=model_out/ --debug

```