biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V1][TPU] Apply the ragged paged attention kernel fix and remove the padding. (#14846)

Browse Source 
Signed-off-by: Xiongfei Wei <isaacwxf23@gmail.com>
This commit is contained in:
iefgnoix
2025-03-17 01:48:28 -07:00
committed by GitHub
parent 69698f257e
commit b4ad56c1bd
2 changed files with 8 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
vllm/v1/worker/tpu_model_runner.py
View File

@@ -23,8 +23,7 @@ from vllm.multimodal.utils import group_mm_inputs_by_modality
from vllm.sampling_params import SamplingType
from vllm.sequence import IntermediateTensors
from vllm.utils import LayerBlockType, cdiv, is_pin_memory_available
from vllm.v1.attention.backends.pallas import (NUM_KV_PAGES_PER_BLOCK,
PallasAttentionBackend,
from vllm.v1.attention.backends.pallas import (PallasAttentionBackend,
PallasMetadata)
from vllm.v1.core.encoder_cache_manager import compute_encoder_budget
from vllm.v1.kv_cache_interface import (FullAttentionSpec, KVCacheConfig,
@@ -139,10 +138,8 @@ class TPUModelRunner:
device="cpu")
self.slot_mapping_np = self.slot_mapping_cpu.numpy()
padded_max_num_blocks_per_req = _get_padded_number(
self.max_num_blocks_per_req, NUM_KV_PAGES_PER_BLOCK)
self.block_table_cpu = torch.zeros(
(self.max_num_tokens, padded_max_num_blocks_per_req),
(self.max_num_tokens, self.max_num_blocks_per_req),
dtype=self.input_batch.block_table.get_cpu_tensor().dtype,
device="cpu")