fix: frank-default-small ctx-size 786432 (3 slots x 262144 each)

llama.cpp splits --ctx-size across --parallel slots, so 262144 total gave 87552 per stream. Non-negotiable is 262144 per stream: total pool must be 786432.
2026-08-11 01:05:24 -07:00
parent f93096f2bf
commit 6a86a9df93
1 changed files with 2 additions and 1 deletions
@@ -4,7 +4,8 @@ MODEL_DIR=/home/dank/models
 ALIAS=frank-default-small
 HOST=0.0.0.0
 PORT=7770
-CTX_SIZE=262144
+# 3 parallel slots x 262144 ctx each = 786432 total context pool.
+CTX_SIZE=786432
 PARALLEL=3
 SPEC_TYPE=draft-dflash
 SPEC_N_MAX=7