Skip to content

[None][fix] Fix stale sparse attention kwargs #15460

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
bobboli merged 1 commit into from
Jun 18, 2026
Merged

[None][fix] Fix stale sparse attention kwargs #15460

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 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 @@ -1301,7 +1301,7 @@ def _create_cross_kv_cache_manager(
max_seq_len=max_seq_len,
max_batch_size=self._max_batch_size,
spec_config=None,
sparse_attn_config=None,
sparse_attention_config=None,
max_num_tokens=self._max_num_tokens,
max_beam_width=1,
kv_connector_manager=None,
Expand Down
6 changes: 5 additions & 1 deletion tensorrt_llm/_torch/pyexecutor/model_engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5234,6 +5234,10 @@ def _prepare_tp_inputs_encoder(
# Build a fresh, no-cache attention metadata for the encoder
# pass. We do not reuse ``self.attn_metadata`` because that
# object is bound to the decoder's KV-cache manager.
sparse_metadata_params = (
self.sparse_attention_config.to_sparse_metadata_params(
pretrained_config=self.model.model_config.pretrained_config)
if self.sparse_attention_config is not None else None)
encoder_attn_metadata = self.attn_backend.Metadata(
max_num_requests=self.batch_size,
max_num_tokens=self.max_num_tokens,
Expand All @@ -5244,7 +5248,7 @@ def _prepare_tp_inputs_encoder(
enable_flash_mla=self.model.model_config.enable_flash_mla,
enable_context_mla_with_cached_kv=False,
cache_indirection=None,
sparse_attention_config=self.sparse_attention_config,
sparse_metadata_params=sparse_metadata_params,
num_heads_per_kv=1,
)
assert isinstance(
Expand Down
Loading