michael/AI-Profile-Router
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Validate Pure quant for 256K Ultra profile

Browse Source
This commit is contained in:
Mikei386
2026-08-22 11:01:13 +02:00
parent f52cf14069
commit 977f8f5c76
9 changed files with 85 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
compose.yaml
+1 -1
View File
@@ -222,7 +222,7 @@ services:
- --spec-draft-type-v
- q4_0
# Text-only long-context profile. This exact IQ4-MIX / 256K / 80:20
# Text-only long-context profile. This exact IQ4_XS-pure / 256K / 80:20
# combination completed the 220K fill test on RTX 5080 + RTX 3060.
# Deliberately no vision projector: Ultra prioritizes maximum usable context.
llama-ultra: