]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: c7d92e6)
convert.py : advanced option (#2753)
authorKerfuffle <redacted>
Sat, 26 Aug 2023 20:13:36 +0000 (14:13 -0600)
committerGitHub <redacted>
Sat, 26 Aug 2023 20:13:36 +0000 (23:13 +0300)
commit730d9c681e339b76407659344e5a2cd50af7d7d5
tree646309864045445e6d9dca452ab048af8a5401a3tree
parentc7d92e6dfec3f54849f3a0ba373054d29f321ea2commit | diff
convert.py : advanced option (#2753)

* Allow convert.py to convert to q8_0

Fix issue with bounded_parallel_map and greedy consuming iterator

Display elapsed time during conversion

* Add --concurrency option

Minor improvements to help text

Clean up bounded_parallel_map function a bit

* Massive speed improvement thanks to Cebtenzzre

* Refactor types
convert.py diff | blob | history
Packaging of ggml-org/llama.cpp