Skip to content

[nvbugs/5369799] fix: Update disaggregation handling in sampler #5762

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

Merged
merged 3 commits into from
Jul 18, 2025
Merged

[nvbugs/5369799] fix: Update disaggregation handling in sampler #5762

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
87343b3
fix: Update disaggregation handling in sampler and executor
stnie Jul 4, 2025
04c9ac8
[fix] Fix bug 5369799: Beam search with disaggregated generation.
stnie Jul 14, 2025
e97a76e
chore: removed unused code lines.
stnie Jul 17, 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
5 changes: 4 additions & 1 deletion cpp/tensorrt_llm/batch_manager/createNewDecoderRequests.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,10 @@ void copySequenceLengths(RequestVector const& contextRequests, DecoderInputBuffe
SizeType32 batchIdx{0};
for (auto const& llmReq : contextRequests)
{
auto const currentSequenceLen = llmReq->mPromptLen + llmReq->getMaxNumGeneratedTokens();
auto const disaggFirstGenTokenSize
= llmReq->getContextPhaseParams() ? llmReq->getContextPhaseParams().value().getFirstGenTokens().size() : 0;
auto const currentSequenceLen
= llmReq->mPromptLen + llmReq->getMaxNumGeneratedTokens() + disaggFirstGenTokenSize;
// Get position of the current sequence in the decoder
auto const seqSlot = llmReq->mSeqSlot.value();
batchSlotsRange[batchIdx] = seqSlot;
Expand Down
1 change: 0 additions & 1 deletion tensorrt_llm/_torch/pyexecutor/_util.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -520,7 +520,6 @@ def create_py_executor_instance(
cache_transceiver_config = executor_config.cache_transceiver_config
kv_cache_transceiver = create_kv_cache_transceiver(
mapping, kv_cache_manager, attention_type, cache_transceiver_config)

return PyExecutor(
resource_manager,
scheduler,
Expand Down
4 changes: 1 addition & 3 deletions tensorrt_llm/_torch/pyexecutor/sampler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -751,8 +751,7 @@ def update_requests_single_beam_single_step(self, state: SampleStateTRTLLM):

reqs_with_new_tokens = [
r for r in reqs
if (sequence_lengths_host_data[r.py_seq_slot] > r.get_num_tokens(0)
or self.is_trt_overlap)
if (sequence_lengths_host_data[r.py_seq_slot] > r.get_num_tokens(0))
]

# Add new tokens
Expand Down Expand Up @@ -821,7 +820,6 @@ def update_requests_multiple_beams_or_drafting(self,
for beam in range(beam_width):
seq_len = sequence_lengths_host_data[seq_slot * beam_width +
beam]
seq_len = seq_len + 1 if self.is_trt_overlap else seq_len
num_new_tokens[beam] = min(
num_generated_tokens,
seq_len - request.get_num_tokens(beam))
Expand Down