Revert "Rebase 2025.01.21 (#714)"

This reverts commit c9db39b.

Author: kzawora-intel

.buildkite/nightly-benchmarks/scripts/nightly-annotate.sh

@@ -43,7 +43,7 @@ main() {
 
 
 
-    # The figures should be generated by a separate process outside the CI/CD pipeline
+    # The figures should be genereated by a separate process outside the CI/CD pipeline
 
     # # generate figures
     # python3 -m pip install tabulate pandas matplotlib

.buildkite/test-pipeline.yaml

@@ -52,6 +52,7 @@ steps:
   - tests/worker
   - tests/standalone_tests/lazy_torch_compile.py
   commands:
+  - pip install git+https://github.com/Isotr0py/DeepSeek-VL2.git  # Used by multimoda processing test
   - python3 standalone_tests/lazy_torch_compile.py
   - pytest -v -s mq_llm_engine # MQLLMEngine
   - pytest -v -s async_engine # AsyncLLMEngine
@@ -477,9 +478,7 @@ steps:
   - pytest models/encoder_decoder/language/test_bart.py -v -s -m 'distributed(num_gpus=2)'
   - pytest models/encoder_decoder/vision_language/test_broadcast.py -v -s -m 'distributed(num_gpus=2)'
   - pytest models/decoder_only/vision_language/test_models.py -v -s -m 'distributed(num_gpus=2)'
-  # this test fails consistently.
-  # TODO: investigate and fix
-  # - pytest -v -s spec_decode/e2e/test_integration_dist_tp2.py
+  - pytest -v -s spec_decode/e2e/test_integration_dist_tp2.py
   - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s test_sharded_state_loader.py
   - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s kv_transfer/disagg_test.py
 
@@ -517,9 +516,7 @@ steps:
   - vllm/engine
   - tests/multi_step
   commands:
-  # this test is quite flaky
-  # TODO: investigate and fix.
-  # - pytest -v -s multi_step/test_correctness_async_llm.py
+  - pytest -v -s multi_step/test_correctness_async_llm.py
   - pytest -v -s multi_step/test_correctness_llm.py
 
 - label: Pipeline Parallelism Test # 45min