Add --enable-auto-tool-choice and --tool-call-parser qwen3_xml for tool choice support

vllm-qwen36.sh

@@ -1,9 +1,6 @@
 #!/usr/bin/env bash
 set -euo pipefail
 
-# Qwen3.6-35B-A3B-FP8 vLLM serving script for dankserver
-# CPU expert offloading to fit MoE model on 24GB RTX 3090
-
 MODEL="Qwen/Qwen3.6-35B-A3B-FP8"
 PORT=8001
 HOST="0.0.0.0"
@@ -15,13 +12,13 @@ exec docker run --rm --name vllm-qwen36 \
     -p "${PORT}:${PORT}" \
     -e PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True \
     -v /home/dank/.cache/huggingface:/root/.cache/huggingface \
-    docker.io/vllm/vllm-openai:nightly \
+    docker.io/vllm/vllm-openai:latest \
     "${MODEL}" \
     --port "${PORT}" \
     --host "${HOST}" \
     --kv-cache-dtype fp8 \
     --gpu-memory-utilization 0.95 \
-    --max-num-seqs 3 \
+    --max-num-seqs 4 \
     --max-model-len 16384 \
     --max-num-batched-tokens 4096 \
     --trust-remote-code \
@@ -29,4 +26,6 @@ exec docker run --rm --name vllm-qwen36 \
     --cpu-offload-params "experts" \
     --enable-chunked-prefill \
     --enable-prefix-caching \
-    --served-model-name frank-default-small
+    --served-model-name frank-default-small \
+    --enable-auto-tool-choice \
+    --tool-call-parser qwen3_xml