[Bugfix][Mamba] Fix Multistep on Mamba-like models (vllm-project#10705)

Signed-off-by: mzusman <mor.zusmann@gmail.com>

Author: mzusman

tests/models/decoder_only/language/test_jamba.py

@@ -275,6 +275,44 @@ def test_state_cleanup(
                     "could be related to finished_requests_ids")
 
 
+@pytest.mark.parametrize("model", MODELS)
+@pytest.mark.parametrize("dtype", ["float"])
+def test_multistep(
+    vllm_runner,
+    model: str,
+    dtype: str,
+    example_prompts,
+) -> None:
+    # This test is verifying that multistep works correctly
+    #on mamba-like models
+    with vllm_runner(model, num_scheduler_steps=8,
+                     max_num_seqs=2) as vllm_model:
+        vllm_model.generate_greedy([example_prompts[0]] * 10, 1)
+
+
+@pytest.mark.parametrize("model", MODELS)
+@pytest.mark.parametrize("dtype", ["float"])
+@pytest.mark.parametrize("max_tokens", [64])
+def test_multistep_correctness(vllm_runner, model: str, dtype: str,
+                               max_tokens: int, example_prompts) -> None:
+    with vllm_runner(model, num_scheduler_steps=8,
+                     max_num_seqs=2) as vllm_model:
+        vllm_outputs_multistep = vllm_model.generate_greedy(
+            example_prompts, max_tokens)
+
+    with vllm_runner(model, num_scheduler_steps=1,
+                     max_num_seqs=2) as vllm_model:
+        vllm_outputs_single_step = vllm_model.generate_greedy(
+            example_prompts, max_tokens)
+
+    check_outputs_equal(
+        outputs_0_lst=vllm_outputs_multistep,
+        outputs_1_lst=vllm_outputs_single_step,
+        name_0="vllm_outputs_multistep",
+        name_1="vllm_outputs_single_step",
+    )
+
+
 @multi_gpu_test(num_gpus=2)
 @pytest.mark.parametrize("model", MODELS)
 @pytest.mark.parametrize("dtype", ["float"])

tests/models/decoder_only/language/test_mamba.py

@@ -283,3 +283,39 @@ def test_state_cleanup(
     except ValueError:
         pytest.fail("Mamba inner state wasn't cleaned up between states, "
                     "could be related to finished_requests_ids")
+
+
+@pytest.mark.parametrize("model", MODELS)
+@pytest.mark.parametrize("dtype", ["float"])
+def test_multistep(
+    vllm_runner,
+    model: str,
+    dtype: str,
+    example_prompts,
+) -> None:
+    with vllm_runner(model, num_scheduler_steps=8,
+                     max_num_seqs=2) as vllm_model:
+        vllm_model.generate_greedy([example_prompts[0]] * 10, 1)
+
+
+@pytest.mark.parametrize("model", MODELS)
+@pytest.mark.parametrize("dtype", ["float"])
+@pytest.mark.parametrize("max_tokens", [64])
+def test_multistep_correctness(vllm_runner, model: str, dtype: str,
+                               max_tokens: int, example_prompts) -> None:
+    with vllm_runner(model, num_scheduler_steps=8,
+                     max_num_seqs=2) as vllm_model:
+        vllm_outputs_multistep = vllm_model.generate_greedy(
+            example_prompts, max_tokens)
+
+    with vllm_runner(model, num_scheduler_steps=1,
+                     max_num_seqs=2) as vllm_model:
+        vllm_outputs_single_step = vllm_model.generate_greedy(
+            example_prompts, max_tokens)
+
+    check_outputs_equal(
+        outputs_0_lst=vllm_outputs_multistep,
+        outputs_1_lst=vllm_outputs_single_step,
+        name_0="vllm_outputs_multistep",
+        name_1="vllm_outputs_single_step",
+    )

vllm/engine/async_llm_engine.py

@@ -301,6 +301,9 @@ async def step_async(
             ctx.seq_group_metadata_list = seq_group_metadata_list
             ctx.scheduler_outputs = scheduler_outputs
 
+            finished_requests_ids = self.scheduler[
+                virtual_engine].get_and_reset_finished_requests_ids()
+
             # Maybe switch from async mode to sync mode
             if not allow_async_output_proc and len(ctx.output_queue) > 0:
                 self._process_model_outputs(ctx=ctx)
@@ -312,13 +315,13 @@ async def step_async(
                 self._cache_scheduler_outputs_for_multi_step(
                     virtual_engine, seq_group_metadata_list, scheduler_outputs,
                     allow_async_output_proc)
+        else:
+            finished_requests_ids = list()
 
         assert seq_group_metadata_list is not None
         assert scheduler_outputs is not None
 
         if not scheduler_outputs.is_empty():
-            finished_requests_ids = self.scheduler[
-                virtual_engine].get_and_reset_finished_requests_ids()
 
             # Check if we have a cached last_output from the previous iteration.
             # For supporting PP this is probably the best way to pass the

vllm/engine/llm_engine.py

@@ -1403,6 +1403,9 @@ def step(self) -> List[Union[RequestOutput, EmbeddingRequestOutput]]:
             ctx.seq_group_metadata_list = seq_group_metadata_list
             ctx.scheduler_outputs = scheduler_outputs
 
+            finished_requests_ids = self.scheduler[
+                virtual_engine].get_and_reset_finished_requests_ids()
+
             # Maybe switch from async mode to sync mode
             if not allow_async_output_proc and len(ctx.output_queue) > 0:
                 self._process_model_outputs(ctx=ctx)
@@ -1414,13 +1417,13 @@ def step(self) -> List[Union[RequestOutput, EmbeddingRequestOutput]]:
                 self._cache_scheduler_outputs_for_multi_step(
                     virtual_engine, seq_group_metadata_list, scheduler_outputs,
                     allow_async_output_proc)
+        else:
+            finished_requests_ids = list()
 
         assert seq_group_metadata_list is not None
         assert scheduler_outputs is not None
 
         if not scheduler_outputs.is_empty():
-            finished_requests_ids = self.scheduler[
-                virtual_engine].get_and_reset_finished_requests_ids()
 
             # Check if we have a cached last_output from the previous iteration.
             # For supporting PP this is probably the best way to pass the