All notable changes to this project will be documented in this file.
- Support scatter operations for MArray (#32)
- Support GCNConv layer accepting graph input (#31)
- Compatible with Julia v1.4 while not support before v1.3
- Not support old version CuArrays, CUDAnative and CUDAapi
- Improve performance of scatter operations for CPU and new benchmark (#29)
- Scatters support almost all Real numbers except Bool on CPU
- Add benchmark for scatter operations
- Implement TopKPool layer (#22)
- Improve performance of scatter operations in both CPU/CUDA version
- Add benchmark result
- Add multihead GAT on graph support
- Move
pool_dim_checktoDimsconstructor
- Available on Julia v1.2 and v1.3
- Convolution layers works with CUDA
- Provide scatter add, sub, mul, div, max, min, mean for CPU and CUDA
- Provide pool add, sub, mul, div, max, min, mean for CPU and CUDA
- Provide gradient of scatter add, sub, mul, div, max, min, mean for CPU and CUDA
- Provide gradient of pool add, sub, mul, div, max, min, mean for CPU and CUDA
- Provide gather
- Provide good abstract for graph network block
- Integrate message passing scheme and graph network block
- Add logo
- Add docs
- Add layer docs and Base.show
- Provide dynamically change graph in runtime
- Provide GlobalPool layer