Andrew Kwangwoong Park
Andrew Kwangwoong Park
### Details: - Disable TCs(_SUPPORTED_CONFIG_KEYS/METRICS) for OVClassHeteroExecutableNetworkGetMetricTest same as GPU and GNA plugin - This is the same change with https://github.com/openvinotoolkit/openvino/pull/12433 ### Tickets: - 88667
### Details: - Divide InferRequest into InferRequest with API 2.0 and InferRequest with legacy - Remove cldnn::layout depedency from create_device_blob - Add reinterpret_device_blob and ov::PartialShape constructor in layout temporarily
### Details: - Enable collecting profiling data only at iterations with requested range - For example for dump profiling data only when iteration is from 10 to 20, you can...
### Details: - Currently `indirect Gemm` is implemented with an assumption of `KV Cache`'s gather axis is 0 - Update `KV Cache` and `indirect Gemm` to support non-batch gather axis...
### Details: - Apply dynamic padding for onednn gemm - Update memory descriptor to handle the padded dims and strides ### Tickets: - 140516
### Details: - Convert QKV FC + StridedSlices pattern to QKV FC + VariadicSplit for potential optimization regarding crop runtime buffer fusing ### Tickets: - 144394
### Details: - Apply in place crop for dynamic shape at runtime ### Tickets: - 143096
### Details: - Enabled RoPE op fusion for GLM-4-9B Hugging Face model on GPU to improve the performance - New config for RoPE is added to distinguish between mode that...
### Details: - Backport https://github.com/openvinotoolkit/openvino/pull/30858