Adjust B200 k8s CI job coverage

2026-04-30 02:06:40 -07:00
2 changed files with 3 additions and 2 deletions
@@ -39,6 +39,7 @@ steps:
  timeout_in_minutes: 120
  device: b200
  optional: true
+  num_devices: 2
  source_file_dependencies:
  - csrc/
  - vllm/model_executor/layers/quantization
@@ -21,7 +21,7 @@ steps:
    - vllm/v1/worker/gpu/spec_decode/
    - tests/v1/e2e/spec_decode/
  commands:
-    - pytest -v -s v1/e2e/spec_decode -k "eagle_correctness"
+    - pytest -v -s v1/e2e/spec_decode -k "eagle_correctness and not deepseek_eagle"

 - label: Spec Decode Speculators + MTP
  timeout_in_minutes: 30
@@ -44,7 +44,7 @@ steps:
    - vllm/transformers_utils/configs/speculators/
    - tests/v1/e2e/spec_decode/
  commands:
-    - pytest -v -s v1/e2e/spec_decode -k "speculators or mtp_correctness"
+    - pytest -v -s v1/e2e/spec_decode -k "(speculators or mtp_correctness) and not deepseek"
  
 - label: Spec Decode Ngram + Suffix
  timeout_in_minutes: 30