frank-large: roll back to gufo Qwen3.8-27B Q4_K_XL, one session x 262144 f16 KV

frank-small: llama.cpp CUDA (master, sm_86) on MiMo Q6_K GGUF, 2 x 262144, asymmetric KV (f16 K + q8_0 V, flash-attn pinned for the V quant)
2026-10-01 22:40:59 -07:00
parent e452a1afbf
commit 8551bb006b
11 changed files with 729 additions and 816 deletions
@@ -71,7 +71,7 @@ fi
 ASI_ARGS="$(ssh "$HOST" 'podman inspect frank-large-mimo \
    --format "{{join .Config.Cmd \" \"}}" 2>/dev/null' | tr -d "\n" | head -c 400 || true)"
 if [ -z "$ASI_ARGS" ]; then
-    ASI_ARGS="$(ssh "$HOST" "tr '\\0' ' ' < /proc/\\$(pgrep -x llama-server | head -1)/cmdline 2>/dev/null" | head -c 400 || true)"
+    ASI_ARGS="$(ssh "$HOST" "sh -c 'tr \"\\0\" \" \" < /proc/\\$(pgrep -x llama-server | head -1)/cmdline 2>/dev/null || true'" | head -c 400 || true)"
 fi
 echo "ASI engine_args=${ASI_ARGS:0:400}"