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

Remove prompt string from engine core data structures (#663)

Browse Source 
### What this PR does / why we need it?
vLLM Ascend side followup on:
[Core] Remove prompt string from engine core data structures

df6f3ce883

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
CI passed

Signed-off-by: Yikun Jiang <yikunkero@gmail.com>
This commit is contained in:
Yikun Jiang
2025-04-26 23:15:58 +08:00
committed by GitHub
parent c99c4c8c70
commit 742f679c7d
1 changed files with 28 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
vllm_ascend/worker/model_runner_v1.py
View File

@@ -346,20 +346,34 @@ class NPUModelRunner:
generator.manual_seed(sampling_params.seed)
else:
generator = None
self.requests[req_id] = CachedRequestState(
req_id=req_id,
prompt_token_ids=new_req_data.prompt_token_ids,
prompt=new_req_data.prompt,
mm_inputs=new_req_data.mm_inputs,
mm_positions=new_req_data.mm_positions,
sampling_params=sampling_params,
generator=generator,
block_ids=new_req_data.block_ids,
num_computed_tokens=new_req_data.num_computed_tokens,
output_token_ids=[],
lora_request=new_req_data.lora_request,
)
if vllm_version_is("0.8.4"):
self.requests[req_id] = CachedRequestState(
req_id=req_id,
prompt_token_ids=new_req_data.prompt_token_ids,
prompt=new_req_data.prompt,
mm_inputs=new_req_data.mm_inputs,
mm_positions=new_req_data.mm_positions,
sampling_params=sampling_params,
generator=generator,
block_ids=new_req_data.block_ids,
num_computed_tokens=new_req_data.num_computed_tokens,
output_token_ids=[],
lora_request=new_req_data.lora_request,
)
else:
# the prompt removed by: https://github.com/vllm-project/vllm/pull/17214
self.requests[req_id] = CachedRequestState(
req_id=req_id,
prompt_token_ids=new_req_data.prompt_token_ids,
mm_inputs=new_req_data.mm_inputs,
mm_positions=new_req_data.mm_positions,
sampling_params=sampling_params,
generator=generator,
block_ids=new_req_data.block_ids,
num_computed_tokens=new_req_data.num_computed_tokens,
output_token_ids=[],
lora_request=new_req_data.lora_request,
)
req_ids_to_add.append(req_id)