Update benchmarks.yml

Author: juliagmt-google

.github/workflows/benchmarks.yml

@@ -129,9 +129,9 @@ jobs:
         run: mkdir results
       - name: Configure XLA for GPU backend
         run: ./configure.py --backend CUDA --nccl
-      - name: Set TF_CPP_MAX_VLOG_LEVEL
-        env:
-          TF_CPP_MAX_VLOG_LEVEL: 1
+      # - name: Set TF_CPP_MAX_VLOG_LEVEL
+      #   env:
+      #     TF_CPP_MAX_VLOG_LEVEL: 1
       - name: Build hlo_runner_main
         run: bazel build -c opt --config=cuda --dynamic_mode=off //xla/tools/multihost_hlo_runner:hlo_runner_main
       # TODO(juliagmt): Add more performance-criticalHLOs to benchmark.