From de93790d08905a497b29741f593d7c5f1f98df73 Mon Sep 17 00:00:00 2001
From: drslark <96540755+drslark@users.noreply.github.com>
Date: Thu, 12 Mar 2026 18:38:50 +0800
Subject: [PATCH] [main][bugfix] Fixed the problem of drafter crashed in FULL
 mode (#7158)

### What this PR does / why we need it?

The merged graph of draft in `FULL` mode is broken now.

This pr solves it.

Also, `actual_seq_lengths_q` in `model_runner` is found redundant, so,
it is removed.

It depends on https://github.com/vllm-project/vllm-ascend/pull/7144 and
https://github.com/vllm-project/vllm-ascend/pull/7148.

### Does this PR introduce _any_ user-facing change?

N/A

### How was this patch tested?

Test code is shown as below:

```python
prompts = [
    "1.Who are you?",
    "2. Who are you?",
]

sampling_params = SamplingParams(temperature=0.0, top_p=0.95, top_k=40, max_tokens=200)
llm = LLM(
    model="/home/some-model/Meta-Llama-3.1-8B-Instruct",
    tensor_parallel_size=1,
    max_num_seqs=32,
    # enforce_eager=True,
    disable_log_stats=False,
    distributed_executor_backend="mp",
    gpu_memory_utilization=0.7,
    async_scheduling=True,

    speculative_config={
        "enforce_eager": True,
        "model": "/home/some-model/EAGLE3-LLaMA3.1-Instruct-8B",
        "disable_padded_drafter_batch": False,
        "method": "eagle3",
        "num_speculative_tokens": 3,
    },

    compilation_config={
        "cudagraph_mode": "FULL",
        "cudagraph_num_of_warmups": 1,
    },

    max_model_len=4096,
    enable_prefix_caching=False,
)

outputs = llm.generate(prompts, sampling_params)
```

The result before:

```text
   File "/vllm-workspace/vllm-ascend/vllm_ascend/attention/attention_v1.py", line 575, in full_graph_fia
     graph_params.events[num_tokens].append(event)
     ~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^
 KeyError: 132
```

The result after:

```text
--------------------------------------------------
total_num_output_tokens: 400
num_drafts: 242
num_draft_tokens: 726
num_accepted_tokens: 156
mean acceptance length: 1.64
--------------------------------------------------
acceptance at token 0: 0.42
acceptance at token 1: 0.16
acceptance at token 2: 0.07
```

We also test `FULL_DECODE_ONLY` mode.

The result is:

```text
--------------------------------------------------
total_num_output_tokens: 400
num_drafts: 244
num_draft_tokens: 732
num_accepted_tokens: 155
mean acceptance length: 1.64
--------------------------------------------------
acceptance at token 0: 0.42
acceptance at token 1: 0.16
acceptance at token 2: 0.06
```

- vLLM version: v0.16.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/4034c3d32e30d01639459edd3ab486f56993876d

Signed-off-by: drslark <slarksblood@qq.com>
---
 .../singlecard/spec_decode/test_mtp_eagle_correctness.py   | 4 ++--
 vllm_ascend/spec_decode/eagle_proposer.py                  | 7 ++++---
 vllm_ascend/worker/model_runner_v1.py                      | 3 ---
 3 files changed, 6 insertions(+), 8 deletions(-)

diff --git a/tests/e2e/singlecard/spec_decode/test_mtp_eagle_correctness.py b/tests/e2e/singlecard/spec_decode/test_mtp_eagle_correctness.py
index 77447e8f..ef89d328 100644
--- a/tests/e2e/singlecard/spec_decode/test_mtp_eagle_correctness.py
+++ b/tests/e2e/singlecard/spec_decode/test_mtp_eagle_correctness.py
@@ -170,8 +170,8 @@ def test_llama_qwen3_eagle_correctness(
                         "max_model_len": 128,
                     },
                     compilation_config=CompilationConfig(
-                        cudagraph_mode="FULL_DECODE_ONLY",
-                        cudagraph_capture_sizes=[12])) as llm:
+                        cudagraph_mode="FULL",
+                        cudagraph_capture_sizes=[5, 12])) as llm:
         spec_outputs = llm.generate(example_prompts, sampling_params)
         cleanup_dist_env_and_memory()
         del llm
diff --git a/vllm_ascend/spec_decode/eagle_proposer.py b/vllm_ascend/spec_decode/eagle_proposer.py
index b8e0c2f9..b9d4e3bd 100644
--- a/vllm_ascend/spec_decode/eagle_proposer.py
+++ b/vllm_ascend/spec_decode/eagle_proposer.py
@@ -315,10 +315,11 @@ class AscendEagleProposer(EagleProposer):
             aclgraph_runtime_mode = CUDAGraphMode.NONE
         if aclgraph_runtime_mode == CUDAGraphMode.FULL and len(self.runner.attn_groups) > 0:
             num_computed_tokens_cpu = self.runner.input_batch.num_computed_tokens_cpu_tensor[:num_reqs]
-            self.query_start_loc.cpu[: num_reqs + 1] = torch.tensor(
-                [0] + self.runner.actual_seq_lengths_q[:num_reqs], device="cpu", dtype=torch.int32
-            )
+
+            # num_reqs is already the padded version
+            self.query_start_loc.cpu[: num_reqs + 1].copy_(self.runner.query_start_loc.cpu[: num_reqs + 1])
             self.query_start_loc.copy_to_gpu()
+
             common_attn_metadata = AscendCommonAttentionMetadata(
                 query_start_loc=self.query_start_loc.gpu[: num_reqs + 1],
                 query_start_loc_cpu=self.query_start_loc.cpu[: num_reqs + 1],
diff --git a/vllm_ascend/worker/model_runner_v1.py b/vllm_ascend/worker/model_runner_v1.py
index 8000dbca..a102d0b1 100644
--- a/vllm_ascend/worker/model_runner_v1.py
+++ b/vllm_ascend/worker/model_runner_v1.py
@@ -420,9 +420,6 @@ class NPUModelRunner(GPUModelRunner):
                     assert isinstance(self.drafter, AscendEagleProposer)
                     self.use_aux_hidden_state_outputs = self.drafter.eagle3_use_aux_hidden_state
                 self.rejection_sampler = RejectionSampler(self.sampler)
-            self.actual_seq_lengths_q = list(
-                range(self.decode_token_per_req, self.max_num_tokens + 1, self.decode_token_per_req)
-            )
         self.discard_request_indices = self._make_buffer(self.max_num_reqs, dtype=torch.int64)
         self.num_discarded_requests = 0