diff --git a/apps/llama/configs/config.yaml b/apps/llama/configs/config.yaml index ff9a6b2..6f5fb40 100644 --- a/apps/llama/configs/config.yaml +++ b/apps/llama/configs/config.yaml @@ -4,7 +4,7 @@ logToStdout: "both" # proxy and upstream macros: base_args: "--no-warmup --port ${PORT}" - common_args: "--fit-target 1024 --no-warmup --port ${PORT}" + common_args: "--fit-target 1536 --no-warmup --port ${PORT}" gemma3_ctx_128k: "--ctx-size 131072" qwen35_ctx_256k: "--ctx-size 262144" gemma_sampling: "--prio 2 --temp 1.0 --repeat-penalty 1.0 --min-p 0.00 --top-k 64 --top-p 0.95"