HF RL Explorer

Task: Feeding .npy (NumPy files) into TensorFlow Data Pipeline

Task: Feeding .npy (NumPy files) into TensorFlow Data Pipeline: a task in Terminal-Lego-15k (Harbor dataset). Create a TensorFlow data pipeline that efficiently reads .npy files without loading all data into memory at once. Each .npy file contains 2 different ndarrays as features and a scalar as…

The task

Create a TensorFlow data pipeline that efficiently reads `.npy` files without loading all data into memory at once. Each `.npy` file contains 2 different ndarrays as features and a scalar as their label. The solution must use `tf.data.Dataset` to lazily load data from individual `.npy` files.

Part of PrimeIntellect/Terminal-Lego-15k.