[bugfix][mm] change get_num_encoder_tokens to get_num_encoder_embeds in recompute_schedule.py (#5132)

### What this PR does / why we need it?
adapt to: https://github.com/vllm-project/vllm/pull/30475.

just change get_num_encoder_tokens() to get_num_encoder_embeds() in
recompute_schedule.py, which seems that it is currently not in use. The
get_num_encoder_tokens() function in VLLM no longer exists.


- vLLM version: v0.13.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9

Signed-off-by: 01267596 <xiongkai123@cmbchina.com>
Co-authored-by: 01267596 <xiongkai123@cmbchina.com>
This commit is contained in:
kxand01267596 authored and GitHub committed 2026-01-21 09:13:52 +08:00
1 parent b399117e89
commit 936d81a258
1 file changed
+3 -3
+3 -3
View File
@@ -235,10 +235,10 @@ class RecomputeScheduler(Scheduler):
if preempted_encoder_inputs:
# Restore encoder compute budget if the preempted
# request had encoder inputs scheduled in this step.
num_tokens_to_restore = sum(
preempted_req.get_num_encoder_tokens(i)
num_embeds_to_restore = sum(
preempted_req.get_num_encoder_embeds(i)
for i in preempted_encoder_inputs)
encoder_compute_budget += num_tokens_to_restore
encoder_compute_budget += num_embeds_to_restore
req_index -= 1
else:
preempted_req = self.running.pop()