gemma.cpp/compression
Luca Versari 4c23932289 Improve weight handling.
- Allow scaling of SFP weights
- Allow using uncompressed weights
- Do not try to compress weights in the main model calls
- Reduce code duplication in weight handling with some macros

Co-authored-by: Eugene Kliuchnikov <eustas@google.com>
Co-authored-by: Thomas Fischbacher <tfish@google.com>
Co-authored-by: Zoltan Szabadka <szabadka@google.com>
2024-04-06 11:08:47 +02:00
..
BUILD Fix SFP/NUQ for bf16 rounding in Highway 2024-03-21 19:06:19 -07:00
analyze.h initial commit 2024-02-21 03:31:22 +00:00
blob_store.cc Improve weight handling. 2024-04-06 11:08:47 +02:00
blob_store.h Improve weight handling. 2024-04-06 11:08:47 +02:00
compress-inl.h Improve weight handling. 2024-04-06 11:08:47 +02:00
compress.h Improve weight handling. 2024-04-06 11:08:47 +02:00
distortion.h Merge pull request #65 from enum-class:narrowing-issues 2024-03-03 18:51:59 -08:00
nuq-inl.h Fix SFP/NUQ for bf16 rounding in Highway 2024-03-21 19:06:19 -07:00
nuq.h initial commit 2024-02-21 03:31:22 +00:00
nuq_test.cc Towards building tests without GUnit Refs #29 2024-03-21 19:28:02 -07:00
sfp-inl.h Fix SFP/NUQ for bf16 rounding in Highway 2024-03-21 19:06:19 -07:00
sfp.h initial commit 2024-02-21 03:31:22 +00:00
sfp_test.cc Towards building tests without GUnit Refs #29 2024-03-21 19:28:02 -07:00
stats.cc Warning fix: sign cast 2024-02-26 22:31:39 -08:00
stats.h Code update 2024-02-22 09:19:47 -08:00
test_util.h Fix SFP/NUQ for bf16 rounding in Highway 2024-03-21 19:06:19 -07:00