EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

[EAGLE] Fix draft kv cache layout for fa3 and topk > 1 (#7239)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-06-16 05:47:51 -07:00
committed by GitHub
parent 8321f8e45e
commit 7ddf8e83d2
1 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
python/sglang/srt/layers/attention/flashattention_backend.py
View File

@@ -406,9 +406,10 @@ class FlashAttentionBackend(AttentionBackend):
dtype=torch.int32,
device=device,
)
# shape: [bs, num_steps, topk] -> [bs x topk, num_steps]
cache_loc = forward_batch.out_cache_loc.view(
self.speculative_num_steps, -1
).T.contiguous()
-1, self.speculative_num_steps
)
metadata_expand.page_table = (
cache_loc[:, :decode_length].contiguous().to(torch.int32)
)
@@ -1636,9 +1637,8 @@ class FlashAttentionBackend(AttentionBackend):
# 2. The second half of metadata for draft tokens (per_batch_num_tokens = topk)
metadata_expand = self.draft_decode_metadata_topk_expand[bs]
decode_length = self.speculative_step_id + 1
cache_loc = out_cache_loc.view(
self.speculative_num_steps, -1
).T.contiguous()
# shape: [bs, num_steps, topk] -> [bs x topk, num_steps]
cache_loc = out_cache_loc.view(-1, self.speculative_num_steps)
metadata_expand.page_table[: cache_loc.shape[0]].copy_(
cache_loc[:, :decode_length]
)