]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
Fix recurrent state serialization for partial reads and writes (llama/22362)
authorGaurav Garg <redacted>
Sun, 26 Apr 2026 11:34:40 +0000 (17:04 +0530)
committerGeorgi Gerganov <redacted>
Thu, 30 Apr 2026 08:29:20 +0000 (11:29 +0300)
commit7296b9c7faec4df1e683d0ef652c3ed4c79ac6ff
treeed9cca225fd2ccc464cf31254d515ff656674e0c
parent9bf6c3c8602b976b79139d908fd63ffe048749ba
Fix recurrent state serialization for partial reads and writes (llama/22362)

The previous code worked only for full tensor reads and writes and was hitting `GGML_ASSERT(size == ggml_nbytes(tensor)); ` assert when tested with llama-server.
ggml/src/ggml-backend-meta.cpp