@@ -4,6 +4,7 @@ This file contains the changelog for the Deeploy project. The changelog is divid
44## Unreleased (Planned Release Target: v0.2.1)
55
66### List of Pull Requests
7+ - FP32 ReduceMean operator improvement [ #137 ] ( https://github.com/pulp-platform/Deeploy/pull/137 )
78- Support for RMSNorm (Pow and Sqrt operators) [ #136 ] ( https://github.com/pulp-platform/Deeploy/pull/136 )
89- Demo TinyViT compatibility with tiled Siracusa [ #124 ] ( https://github.com/pulp-platform/Deeploy/pull/124 )
910- TinyViT on non-tiled Siracusa [ #117 ] ( https://github.com/pulp-platform/Deeploy/pull/117 )
@@ -27,6 +28,10 @@ This file contains the changelog for the Deeploy project. The changelog is divid
2728- Fix bias hoisting in generic GEMM with no bias [ #126 ] ( https://github.com/pulp-platform/Deeploy/pull/126 )
2829
2930### Added
31+ - Support for unknown number of data dimensions in the tiler
32+ - Parallelization support for the FP32 ReduceMean operator on PULPOpen
33+ - Extensive testing for the ReduceMean operator
34+ - Pass to remove ReduceMean operators that don't change data content, but only its shape
3035- Support for RMSNorm operation via operator decomposition.
3136- Added ` Pow ` (Power) and ` Sqrt ` (Square Root) operation support (Parsers, Layers, Bindings, Templates, and FP32 Kernels) for the Generic platform.
3237- Support for input tiling for PULP FP regular and DW conv 2D.
@@ -78,6 +83,8 @@ This file contains the changelog for the Deeploy project. The changelog is divid
7883- annotateNCores method to PULPDeployer that adds an ` n_cores ` key to all PULPClusterEngine templates' operatorRepresentations
7984
8085### Changed
86+ - Structure of Tests subdir for improved ordering
87+ - Structure of .gitignore file for improved ordering
8188- Decreased L1 maximal memory limit for CI pipeline tests where compatible thanks to the implementation of Conv2D input tiling support.
8289- Reduced size of reshape & skip connection test, for non-tiled Siracusa memory compatibility.
8390- Replaced platform-specific tags (` *-amd64 ` , ` *-arm64 ` ) with direct digest references in ` Noelware/docker-manifest-action ` .
@@ -118,6 +125,7 @@ This file contains the changelog for the Deeploy project. The changelog is divid
118125- changed ` _mapNode ` to ` _selectEngine ` which reduces the responsibility of that function to, as the name states, just engine selection
119126
120127### Fixed
128+ - Fixed ReduceMean parallelization and tiling issues described in Issue [ #134 ] ( https://github.com/pulp-platform/Deeploy/issues/134 ) .
121129- Fixed PULP FP32 regular and DW Conv2D, and MatMul tile constraints.
122130- Fixed type casting for tiling code generation.
123131- Fixed bug in buffer name identification in code generation for tests with L3 default memory level.
0 commit comments