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

[Bugfix] Fix KeyError: 1 When Using LoRA adapters (#5164)

Browse Source
This commit is contained in:
Bla_ckB 2024-06-10 06:23:14 +07:00 committed by GitHub
parent 5467ac3196
commit 45f92c00cf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/core/scheduler.py
View File

@ -423,7 +423,9 @@ class Scheduler:
num_running_seqs = seq_group.get_max_num_running_seqs() num_running_seqs = seq_group.get_max_num_running_seqs()
budget.subtract_num_seqs(seq_group.request_id, budget.subtract_num_seqs(seq_group.request_id,
num_running_seqs) num_running_seqs)
if curr_loras is not None and seq_group.lora_int_id > 0:
if (curr_loras is not None and seq_group.lora_int_id > 0
and seq_group.lora_int_id in curr_loras):
curr_loras.remove(seq_group.lora_int_id) curr_loras.remove(seq_group.lora_int_id)
if running_queue: if running_queue: