-
glu_python.py
- Implements the GELU (Gaussian Error Linear Unit) activation function in Python.
- Used for verifying the correctness of CUDA implementation and testing input/output behavior.
- Supports input arrays of arbitrary dimensions.
-
glu.cu
- CUDA implementation of the GELU activation function for high-performance computation on GPUs.
- Handles large-scale tensor operations with optimized memory allocation and parallelism.