dylanyunlong/project_6
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

[fix] docker callback timeout

Browse Source
This commit is contained in:
root
2026-08-25 03:54:08 +00:00
parent 44314afb70
commit 324fd7f3a9
2 changed files with 21 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
computility-run.yaml
View File

@@ -9,9 +9,9 @@ command:
cp /workspace/qwen3_6_scripts/serving_chat.py "${VLLM_ROOT}/entrypoints/openai/serving_chat.py" 2>/dev/null && echo '[runtime] serving_chat.py redeployed' || echo '[runtime] serving_chat.py redeploy skipped';
cp /workspace/qwen3_6_scripts/multimodal_utils.py "${VLLM_ROOT}/multimodal/utils.py" 2>/dev/null && echo '[runtime] multimodal_utils.py redeployed' || echo '[runtime] multimodal_utils.py redeploy skipped';
cp /workspace/qwen3_6_scripts/envs.py "${VLLM_ROOT}/envs.py" 2>/dev/null && echo '[runtime] envs.py redeployed' || echo '[runtime] envs.py redeploy skipped';
find "${VLLM_ROOT}" -name '*.pyc' -path '*/openai/*' -delete 2>/dev/null;
find "${VLLM_ROOT}" -name '*.pyc' -path '*/multimodal/*' -delete 2>/dev/null;
find "${VLLM_ROOT}" -maxdepth 1 -name 'envs*.pyc' -delete 2>/dev/null;
cp /workspace/qwen3_6_scripts/vendor_overrides/vllm/sampling_params.py "${VLLM_ROOT}/sampling_params.py" 2>/dev/null && echo '[runtime] sampling_params.py redeployed' || echo '[runtime] sampling_params.py redeploy skipped';
find "${VLLM_ROOT}" -name '*.pyc' -delete 2>/dev/null;
find "${VLLM_ROOT}" -name '__pycache__' -type d -exec rm -rf {} + 2>/dev/null;
exec python3 -m vllm.entrypoints.openai.api_server
--model /model
--served-model-name llm

21
qwen3_6_scripts/serving_chat.py
View File

@@ -396,8 +396,23 @@ class OpenAIServingChat(OpenAIServing):
# OpenAI API: max_completion_tokens takes precedence over max_tokens
if request.max_completion_tokens is not None and request.max_tokens is None:
request.max_tokens = request.max_completion_tokens
default_max_tokens = min(self.max_model_len - len(
prompt_inputs["prompt_token_ids"]), 8192)
# [BI100] Adaptive max_tokens based on prompt length.
_prompt_len = len(prompt_inputs["prompt_token_ids"])
if _prompt_len > 65536:
_adaptive_cap = 256
elif _prompt_len > 32768:
_adaptive_cap = 512
elif _prompt_len > 16384:
_adaptive_cap = 1024
else:
_adaptive_cap = 8192
if request.max_tokens is not None:
request.max_tokens = min(request.max_tokens, _adaptive_cap)
default_max_tokens = min(self.max_model_len - _prompt_len,
_adaptive_cap)
if request.use_beam_search:
sampling_params = request.to_beam_search_params(
default_max_tokens)
@@ -412,7 +427,7 @@ class OpenAIServingChat(OpenAIServing):
prompt_adapter_request=prompt_adapter_request)
engine_inputs = TokensPrompt(
prompt_token_ids=prompt_inputs["prompt_token_ids"]), 8192)
prompt_token_ids=prompt_inputs["prompt_token_ids"])
if mm_data is not None:
engine_inputs["multi_modal_data"] = mm_data