Skip to content

[None][fix] DSv4: drop duplicate indexer_k_dtype kwarg on V4 sparse attn #14321

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
lfr-0531 merged 1 commit into from
May 20, 2026
Merged

[None][fix] DSv4: drop duplicate indexer_k_dtype kwarg on V4 sparse attn #14321

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: 0 additions & 2 deletions tensorrt_llm/_torch/model_config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -794,7 +794,6 @@ def update_sparse_attention_indexer_config(pretrained_config, kwargs):
if checkpoint_window_size is None:
checkpoint_window_size = getattr(
pretrained_config, 'sliding_window', None)
indexer_k_dtype = sparse_attention_config.indexer_k_dtype if sparse_attention_config else 'fp8'
if sparse_attention_config:
compress_ratios = sparse_attention_config.compress_ratios
window_size = sparse_attention_config.window_size
Expand Down Expand Up @@ -838,7 +837,6 @@ def update_sparse_attention_indexer_config(pretrained_config, kwargs):
'sparse_attention_config'] = DeepSeekV4SparseAttentionConfig(
compress_ratios=compress_ratios,
window_size=window_size,
indexer_k_dtype=indexer_k_dtype,
**indexer_config)
else:
raise ValueError(
Expand Down
Loading