Training/Validation Data structure [Prio 2]

[MehdiGhorb]

• Define how many Datasets should be used
• Define how many videos each dataset should contain
• Define the structure of the Dataset (Text and video tensors to be stored together or separately?)
• IMPORTANT: How to automate and enhance this process? (Shall we start just with small Dataset to see if the dataset quality is good?)

[MehdiGhorb]

The automation precess has been finished!
We only need to figure out the training dataset size