Skip to content

[Bugfix] Padded Eagle Specdec with Chunked Prefill #26263

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
+21 −12
Open

[Bugfix] Padded Eagle Specdec with Chunked Prefill #26263

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4dda0af
Fix padded specdec when chunked prefill
Flechman Oct 5, 2025
b43e38b
Add test
Flechman Oct 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 18 additions & 5 deletions tests/v1/e2e/test_spec_decode.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,9 +124,9 @@ def test_ngram_correctness(


@pytest.mark.parametrize(
["model_setup", "mm_enabled"],
["model_setup", "mm_enabled", "chunked_prefill_enabled"],
[
(("eagle3", "Qwen/Qwen3-8B", "AngelSlim/Qwen3-8B_eagle3", 1), False),
(("eagle3", "Qwen/Qwen3-8B", "AngelSlim/Qwen3-8B_eagle3", 1), False, False),
pytest.param(
(
"eagle3",
Expand All @@ -135,6 +135,7 @@ def test_ngram_correctness(
1,
),
False,
False,
marks=pytest.mark.skip(
reason="Skipping due to its head_dim not being a a multiple of 32"
),
Expand All @@ -147,6 +148,7 @@ def test_ngram_correctness(
1,
),
False,
True,
),
(
(
Expand All @@ -156,6 +158,7 @@ def test_ngram_correctness(
1,
),
False,
False,
),
pytest.param(
(
Expand All @@ -165,6 +168,7 @@ def test_ngram_correctness(
4,
),
False,
False,
marks=large_gpu_mark(min_gb=80),
), # works on 4x H100
pytest.param(
Expand All @@ -175,6 +179,7 @@ def test_ngram_correctness(
4,
),
True,
True,
marks=large_gpu_mark(min_gb=80),
), # works on 4x H100
(
Expand All @@ -185,6 +190,7 @@ def test_ngram_correctness(
1,
),
False,
False,
),
],
ids=[
Expand All @@ -203,6 +209,7 @@ def test_eagle_correctness(
sampling_config: SamplingParams,
model_setup: tuple[str, str, str, int],
mm_enabled: bool,
chunked_prefill_enabled: bool,
attn_backend: str,
):
if attn_backend == "TREE_ATTN":
Expand Down Expand Up @@ -239,9 +246,13 @@ def test_eagle_correctness(
m.setenv("VLLM_ROCM_USE_AITER", "1")

method, model_name, spec_model_name, tp_size = model_setup
max_model_len = 2048
max_num_batched_tokens = max_model_len
if chunked_prefill_enabled:
max_num_batched_tokens = 128

ref_llm = LLM(
model=model_name, max_model_len=2048, tensor_parallel_size=tp_size
model=model_name, max_model_len=max_model_len, tensor_parallel_size=tp_size
)
ref_outputs = ref_llm.chat(test_prompts, sampling_config)
del ref_llm
Expand All @@ -256,9 +267,11 @@ def test_eagle_correctness(
"method": method,
"model": spec_model_name,
"num_speculative_tokens": 3,
"max_model_len": 2048,
"max_model_len": max_num_batched_tokens,
},
max_model_len=2048,
max_model_len=max_model_len,
max_num_batched_tokens=max_num_batched_tokens,
enable_chunked_prefill=chunked_prefill_enabled,
)
spec_outputs = spec_llm.chat(test_prompts, sampling_config)
matches = 0
Expand Down
10 changes: 3 additions & 7 deletions vllm/v1/spec_decode/eagle.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -522,13 +522,9 @@ def prepare_next_token_ids_padded(
)

# Generate a mask for all valid tokens within those requests
max_gen_len = sampled_token_ids.shape[-1]
if max_gen_len == 1:
valid_mask = torch.ones_like(valid_sampled_token_ids_gpu, dtype=torch.bool)
else:
valid_mask = (valid_sampled_token_ids_gpu != -1) & (
valid_sampled_token_ids_gpu < gpu_input_batch.vocab_size
)
valid_mask = (valid_sampled_token_ids_gpu != -1) & (
valid_sampled_token_ids_gpu < gpu_input_batch.vocab_size
)
Comment on lines +525 to +527
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

This simplification is a great improvement. By removing the special case for max_gen_len == 1, the code is now more robust. The previous logic didn't account for discarded requests when max_gen_len == 1, which could lead to using an invalid token ID of -1. This unified approach correctly handles all cases.


# Count the number of valid tokens in each request
valid_sampled_tokens_count = valid_mask.sum(dim=1)
Expand Down