[diffusion] CI: rebalance 2-gpu shards and cut the job timeout to 45m (#38239)

.github/workflows/pr-test-multimodal-gen.yml

@@ -292,7 +292,7 @@ jobs:
       ((github.event_name == 'schedule' || inputs.test_parallel_dispatch == 'true') || (inputs.caller_needs_failure != 'true' && !cancelled())) &&
       inputs.multimodal_gen == 'true'
     runs-on: 2-gpu-h100
-    timeout-minutes: 240
+    timeout-minutes: 45
     strategy:
       fail-fast: false
       matrix: ${{ fromJson(needs.compute-diffusion-partitions.outputs.matrix-2gpu) }}
@@ -324,7 +324,7 @@ jobs:
           CUSTOM_BUILD_SGL_KERNEL=${{inputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh diffusion
 
       - name: Run diffusion server tests
-        timeout-minutes: 240
+        timeout-minutes: 45
         env:
           HF_TOKEN: ${{ secrets.SGLANG_DIFFUSION_CI_HF_TOKEN || secrets.HF_TOKEN }}
           HUGGING_FACE_HUB_TOKEN: ${{ secrets.SGLANG_DIFFUSION_CI_HF_TOKEN || secrets.HF_TOKEN }}