vh 4a4a6fc803 llama-swap: artemis-31b-v1i — add --reasoning on + deepseek format ...

Worldtree-dev's --reasoning-format gemma suggestion isn't supported in
the deployed llama.cpp build (accepts only none|deepseek|deepseek-legacy).
Falling back to deepseek, which also populates reasoning_content — the
field Worldtree's GemmaProvider fallback path checks. Verified via test
inference: 608 reasoning_content deltas + 199 content deltas + 0 raw
<|channel> marker leaks.

2026-05-27 00:06:54 -07:00

..

config.yaml

llama-swap: artemis-31b-v1i — add --reasoning on + deepseek format

2026-05-27 00:06:54 -07:00