ginnoir/homelabstack
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(llm): add --jinja so gpt-oss harmony template returns content

Browse Source 
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
ginnoir
2026-06-27 00:09:09 -05:00
co-authored by Claude Opus 4.8
parent f20712a5d8
commit 6cef600d25
1 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
stacks/llm/llama-swap-config.yaml
+3 -1
View File
@@ -15,7 +15,9 @@ healthCheckTimeout: 300
logLevel: info logLevel: info
macros: macros:
common: "-ngl 99 --parallel 1 --ctx-size 65536 --flash-attn on --cache-type-k q8_0 --cache-type-v q8_0" # --jinja applies each model's embedded chat template (REQUIRED for gpt-oss'
# harmony format, else content comes back empty; harmless/correct for gemma4).
common: "-ngl 99 --parallel 1 --ctx-size 65536 --flash-attn on --cache-type-k q8_0 --cache-type-v q8_0 --jinja"
models: models:
"gpt-oss-20b": "gpt-oss-20b":