2 Commits

Author	SHA1	Message	Date
Justine Tunney	918234ce80	Remove flush to zero from bf16 ... After closely analyzing Google Brain codebases, we decided that flushing to zero was the wrong thing to do. Intel and AMD probably designed their microprocessors to always flush to zero for the wrong reasons. It should have been made conditional on FTZ being set in MXCSR like other opcodes. See ggerganov/llama.cpp#7843	2024-07-03 05:39:16 -07:00
Justine Tunney	ede59bb742	Add BF16 support and fix warnings ... This change updates the data type definitions to be the same as the latest source code. Support for the bfloat16 data type is available however it can't interpret the IQ quantization formats yet. Cleanup of compiler warnings and other nits have been fixed, but behavioral changes have been avoided, and no new features are as of yet added.	2024-05-25 22:58:50 -07:00