[Deprecation] Deprecate items scheduled for 0.29 (#55353)

Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Wentao Ye
2026-09-11 08:17:02 -07:00
committed by GitHub
parent 1e1060f998
commit 5fe77aecfc
28 changed files with 62 additions and 305 deletions
@@ -409,10 +409,6 @@ def _build_attention_metadata(
seq_lens_cpu = torch.tensor(kv_lens, dtype=torch.int32)
seq_lens_gpu = seq_lens_cpu.to(device)
# Build num_computed_tokens (context length for each request)
context_lens = [kv_len - q_len for q_len, kv_len in zip(q_lens, kv_lens)]
num_computed_tokens_cpu = torch.tensor(context_lens, dtype=torch.int32)
# Build block table
num_blocks_per_req = [(kv + block_size - 1) // block_size for kv in kv_lens]
max_num_blocks = max(num_blocks_per_req)
@@ -454,8 +450,6 @@ def _build_attention_metadata(
query_start_loc_cpu=q_start_cpu,
seq_lens=seq_lens_gpu,
seq_lens_cpu_upper_bound=seq_lens_cpu,
_seq_lens_cpu=seq_lens_cpu,
_num_computed_tokens_cpu=num_computed_tokens_cpu,
slot_mapping=slot_mapping,
block_table_tensor=block_table_gpu,
dcp_local_seq_lens=None,