Skip to content

[Bugfix][Hardware][Gaudi] Bump vllm_hpu_extension version #11028

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 2 commits into from
Dec 9, 2024
Merged

[Bugfix][Hardware][Gaudi] Bump vllm_hpu_extension version #11028

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
61bbf51
fix vllm-hpu-extension dependency
kzawora-intel Dec 9, 2024
7574c3e
bump version again
kzawora-intel Dec 9, 2024
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 requirements-hpu.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,4 +8,4 @@ pandas
tabulate
setuptools>=61
setuptools-scm>=8
vllm-hpu-extension @ git+https://github.com/HabanaAI/vllm-hpu-extension.git@fd7f2e6
vllm-hpu-extension @ git+https://github.com/HabanaAI/vllm-hpu-extension.git@e096d6f
11 changes: 11 additions & 0 deletions vllm/attention/backends/hpu_attn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -111,8 +111,16 @@ def __init__(
self.matmul_qk = Matmul()
self.softmax = Softmax()
self.matmul_av = Matmul()
self.batch2block_matmul = Matmul()
self.block2batch_matmul = Matmul()
# NOTE(kzawora): Contiguous PA is off until model runner supports it
self.k_cache = VLLMKVCache()
self.k_cache.use_contiguous_pa = False
self.v_cache = VLLMKVCache()
self.v_cache.use_contiguous_pa = False
# NOTE(kzawora): Pipelined PA is off until model runner supports it
ops.pa_impl = ops.pa

self.num_kv_heads = num_heads if num_kv_heads is None else num_kv_heads
self.sliding_window = sliding_window
self.alibi_slopes = alibi_slopes
Expand Down Expand Up @@ -228,9 +236,12 @@ def forward(
block_mapping=attn_metadata.block_mapping,
block_bias=attn_metadata.attn_bias,
block_scales=attn_metadata.block_scales,
block_groups=None,
scale=self.scale,
matmul_qk_op=self.matmul_qk,
matmul_av_op=self.matmul_av,
batch2block_matmul_op=self.batch2block_matmul,
block2batch_matmul_op=self.block2batch_matmul,
keys_fetch_func=self.k_cache.fetch_from_cache,
values_fetch_func=self.v_cache.fetch_from_cache)
# Reshape the output tensor.
Expand Down
Loading