Skip to content

[Kernel] Use self.kv_cache and forward_context.attn_metadata in Attention.forward #12536

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 5 commits into from
Feb 5, 2025
Merged

[Kernel] Use self.kv_cache and forward_context.attn_metadata in Attention.forward #12536

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
95813d9
fix Attention.forward
heheda12345 Jan 29, 2025
a860063
fix lint error
heheda12345 Jan 29, 2025
40905a3
fix tpu test
heheda12345 Jan 29, 2025
92d3397
fix pre-commit
heheda12345 Jan 29, 2025
e1399a8
update note
heheda12345 Jan 29, 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
28 changes: 22 additions & 6 deletions vllm/attention/layer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -148,9 +148,13 @@ def forward(
kv_cache: torch.Tensor,
attn_metadata: AttentionMetadata,
Copy link
Collaborator

Choose a reason for hiding this comment

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

should this get removed as a parameter then?

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Theoretically we can. But it is a large user-faced API change, we need to wait until we have to make such a change.

) -> torch.Tensor:
if self.calculate_kv_scales and \
attn_metadata.enable_kv_scales_calculation:
self.calc_kv_scales(key, value)
# NOTE: please avoid accessing `kv_cache` and `attn_metadata` arguments
# directly, use `self.kv_cache` and
# `get_forward_context().attn_metadata` instead.
if self.calculate_kv_scales:
ctx_attn_metadata = get_forward_context().attn_metadata
if ctx_attn_metadata.enable_kv_scales_calculation:
self.calc_kv_scales(key, value)
if self.use_output:
output = torch.empty_like(query)
hidden_size = query.size(-1)
Expand All @@ -164,15 +168,27 @@ def forward(
if value is not None:
value = value.view(-1, self.num_kv_heads, self.head_size)
if self.use_direct_call:
unified_attention_with_output(query, key, value, output,
self.layer_name)
forward_context: ForwardContext = get_forward_context()
ctx_attn_metadata = forward_context.attn_metadata
self_kv_cache = self.kv_cache[forward_context.virtual_engine]
self.impl.forward(self,
query,
key,
value,
self_kv_cache,
ctx_attn_metadata,
output=output)
else:
torch.ops.vllm.unified_attention_with_output(
query, key, value, output, self.layer_name)
return output.view(-1, hidden_size)
else:
if self.use_direct_call:
return unified_attention(query, key, value, self.layer_name)
forward_context = get_forward_context()
ctx_attn_metadata = forward_context.attn_metadata
self_kv_cache = self.kv_cache[forward_context.virtual_engine]
return self.impl.forward(self, query, key, value,
self_kv_cache, ctx_attn_metadata)
else:
return torch.ops.vllm.unified_attention(
query, key, value, self.layer_name)
Expand Down