Update qwen35-35b-a3b-fp8.yaml

--max_num_batched_tokens is a default variable now, which can be overriden via the CLI

recipes/qwen35-35b-a3b-fp8.yaml

@@ -25,23 +25,25 @@ defaults:
   tensor_parallel: 2
   gpu_memory_utilization: 0.7
   max_model_len: 131072
+  max_num_batched_tokens: 16384
 
 # Environment variables
-env: {}
+env: 
+  VLLM_MARLIN_USE_ATOMIC_ADD: 1
 
 # The vLLM serve command template
 command: |
   vllm serve Qwen/Qwen3.5-35B-A3B-FP8 \
-    --max-num-batched-tokens 16384 \
-    --enable-auto-tool-choice \
-    --tool-call-parser qwen3_coder \
-    --gpu-memory-utilization {gpu_memory_utilization} \
     --host {host} \
     --port {port} \
+    --max-model-len {max_model_len} \
+    --max-num-batched-tokens {max_num_batched_tokens} \
+    --gpu-memory-utilization {gpu_memory_utilization} \
+    --enable-auto-tool-choice \
+    --tool-call-parser qwen3_coder \
     --kv-cache-dtype fp8 \
     --load-format fastsafetensors \
     --attention-backend flashinfer \
     --enable-prefix-caching \
-    --max-model-len {max_model_len} \
     -tp {tensor_parallel} \
     --distributed-executor-backend ray