EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

adapt to main2main for model runner v2 (#7578)

Browse Source 
### What this PR does / why we need it?
This PR aims to adapt to newest commit of vllm main branch for model
runner v2. please refer to
https://github.com/vllm-project/vllm-ascend/issues/5208
### Does this PR introduce _any_ user-facing change?
no

### How was this patch tested?

- vLLM version: v0.18.0
- vLLM main:
ed359c497a

---------

Signed-off-by: Ronald1995 <ronaldautomobile@163.com>
This commit is contained in:
Ronald
2026-03-25 09:08:44 +08:00
committed by GitHub
parent fc3ec100bc
commit d96440924a
16 changed files with 239 additions and 264 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
vllm_ascend/patch/worker/patch_v2/patch_triton.py Normal file
View File

@@ -0,0 +1,12 @@
from vllm.v1.worker.gpu import input_batch
from vllm.v1.worker.gpu.sample import gumbel, logprob, penalties
from vllm_ascend.worker.v2.input_batch import post_update
from vllm_ascend.worker.v2.sample.gumbel import gumbel_sample
from vllm_ascend.worker.v2.sample.logprob import compute_token_logprobs
from vllm_ascend.worker.v2.sample.penalties import apply_penalties
logprob.compute_token_logprobs = compute_token_logprobs
penalties.apply_penalties = apply_penalties
gumbel.gumbel_sample = gumbel_sample
input_batch.post_update = post_update