Stabilize streaming runtime completion

This commit is contained in:
2026-05-21 23:51:23 +08:00
parent 9e5d5bbd5b
commit e6b82f0376
5 changed files with 261 additions and 441 deletions
+1 -1
View File
@@ -8,7 +8,7 @@ DUCK_CTX_SIZE=4096
DUCK_N_GPU_LAYERS=20
DUCK_PARALLEL=1
DUCK_LLAMA_DEVICE=Vulkan0
DUCK_LLAMA_EXTRA_ARGS="--reasoning-budget 512 --cache-ram 0"
DUCK_LLAMA_EXTRA_ARGS="--reasoning off --cache-ram 0"
DUCK_HOST=127.0.0.1
DUCK_API_HOST=127.0.0.1