How to create a tflite file with custom dataset?

TensorFlow
1

Greetings to all!
I begin to understand something about deep learning to create a custom tflite file.
So I used LabelImg to create a dataset of images and file xml associated.
Now I shoul like to know how create a tflite file from my directories.
I have 2 directory called ‘train’ and ‘verified’ where inside I put the same file (as I saw in a video tutorial for Gooble Colab), I installed Jupiter Notebook but I have not found a script pre-configurated to give to Jupiter to create my personal tflite file in local.
Can someone help me please? :pray:
Sorry for my english, but I hope I have explained myself. :grin:
Thanks to all to support.
Roberto

2

Hi @inVoce_Italia, Could you please let me know if my understanding is correct or not, like you want to know how to create a tflite file for directories containing images. If that is the case you cannot create a tflite file for the dataset. You can create a tflite file for the model that was trained using the dataset which you created. Once the model is trained on the dataset which you created you can create a tflite file for the model using

# Convert the model.
converter = tf.lite.TFLiteConverter.from_keras_model(model)
tflite_model = converter.convert()

# Save the model.
with open('model.tflite', 'wb') as f:
  f.write(tflite_model)

Thank You.