TensorRT-LLMs/cpp/tensorrt_llm/kernels/decoderMaskedMultiheadAttention/decoderXQAImplJIT
Bo Li 1b79041f5d
fix: XQA is not enabled when history_length < kMinHistoryTokensPerBlock. (#4264)
Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com>
2025-06-11 09:38:10 +08:00
..
nvrtcWrapper chore: Mass integration of release/0.20 (#4898) 2025-06-08 23:26:26 +08:00
compileEngine.cpp [feat] Support XQA-based MLA on SM120 (#4858) 2025-06-06 22:32:49 +08:00
compileEngine.h refactor: Clean up CMakeLists.txt (#3479) 2025-04-18 14:39:29 +08:00
cubinObj.cpp fix: rename some terms (#4534) 2025-05-23 23:23:49 +08:00
cubinObj.h Update TensorRT-LLM (#2755) 2025-02-11 03:01:00 +00:00
cubinObjRegistry.h fix: segfault in cudaDriverWrapper (#3017) 2025-04-02 08:55:19 +02:00
decoderXQAImplJIT.cpp fix: XQA is not enabled when history_length < kMinHistoryTokensPerBlock. (#4264) 2025-06-11 09:38:10 +08:00
decoderXQAImplJIT.h Support speculative decoding with Hopper XQA (#3269) 2025-04-07 17:14:34 +08:00
kernelUtils.cpp [feat] Support XQA-based MLA on SM120 (#4858) 2025-06-06 22:32:49 +08:00
kernelUtils.h [feat] Support XQA-based MLA on SM120 (#4858) 2025-06-06 22:32:49 +08:00
serializationUtils.h Update TensorRT-LLM (#1554) 2024-05-07 23:34:28 +08:00