colin/nix-files
1
0
Fork 1
Code Issues Pull Requests Projects Releases Wiki Activity

ollama: ~ship kimi-k2~ don't ship it: it needs 3x the RAM any of my machines have :|

Browse Source
This commit is contained in:
Colin
2025-07-26 08:06:45 +00:00
parent 5a6dbada11
commit bfbd5624fc
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
modules/services/ollama.nix
View File

@@ -53,6 +53,7 @@ let
gemma3n-e4b gemma3n-e4b
# glm4-9b # it generates invalid code # glm4-9b # it generates invalid code
# hermes3-8b # FAST, but unwieldy # hermes3-8b # FAST, but unwieldy
# kimi-k2-1026b # MoE with 32B activated parameters, 384 experts (requires 384 GiB RAM)
# llama3-chatqa-8b # it gets stuck # llama3-chatqa-8b # it gets stuck
# llama3_1-70b # generates like 1 word/sec, decent output (comparable to qwen2_5-32b) # llama3_1-70b # generates like 1 word/sec, decent output (comparable to qwen2_5-32b)
# llama3_2-3b # redundant with uncensored llama # llama3_2-3b # redundant with uncensored llama