xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-09-09 03:07:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix openvino on GPU (#10793)

Browse Source
This commit is contained in:
Jani Monoses 2024-12-02 20:52:19 +02:00 committed by GitHub
parent 519cc6ca12
commit 9b14d978aa
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/worker/openvino_worker.py
View File

@ -489,7 +489,7 @@ class OpenVINOWorker(LoraNotSupportedWorkerBase):
block_size = cache_config.block_size block_size = cache_config.block_size
seq_num_blocks = (seq_len + block_size - 1) // block_size seq_num_blocks = (seq_len + block_size - 1) // block_size
seq_data, dummy_multi_modal_data = input_registry \ dummy_data = input_registry \
.dummy_data_for_profiling(model_config, .dummy_data_for_profiling(model_config,
seq_len, seq_len,
mm_registry) mm_registry)
@ -498,11 +498,11 @@ class OpenVINOWorker(LoraNotSupportedWorkerBase):
seq = SequenceGroupMetadata( seq = SequenceGroupMetadata(
request_id=str(group_id), request_id=str(group_id),
is_prompt=True, is_prompt=True,
seq_data={group_id: seq_data}, seq_data={group_id: dummy_data.seq_data},
sampling_params=sampling_params, sampling_params=sampling_params,
block_tables=block_tables, block_tables=block_tables,
lora_request=None, lora_request=None,
multi_modal_data=dummy_multi_modal_data) multi_modal_data=dummy_data.multi_modal_data)
seqs.append(seq) seqs.append(seq)
self.model_runner.block_size = tmp_cache_config.block_size self.model_runner.block_size = tmp_cache_config.block_size