810aef8c39
Root cause analysis of Sub508 (41.2% score): 1. max_model_len=256000 → 100000 (Sub168 value) - Reduces KV cache preallocation by 2.56x - d01: should drop from 95.87s to ~8-10s - Frees GPU memory for stable inference 2. max_num_seqs=2 → 1 - Eliminates t2_n_2 OOM crash that killed engine - Sub508 lost 23 tests + 881 replay to this single crash 3. gpu_memory_utilization=0.95 → 0.9 (matches Sub168 docker log) 4. serving_chat.py content fallback improved for d07
54 lines
1.4 KiB
YAML
54 lines
1.4 KiB
YAML
concurrency: 1
|
|
command:
|
|
- python3
|
|
- -m
|
|
- vllm.entrypoints.openai.api_server
|
|
- --model
|
|
- /model
|
|
- --served-model-name
|
|
- llm
|
|
- --max-model-len
|
|
- '100000'
|
|
- --gpu-memory-utilization
|
|
- '0.9'
|
|
- --trust-remote-code
|
|
- -tp
|
|
- '4'
|
|
- --max-num-seqs
|
|
- '1'
|
|
- --max-num-batched-tokens
|
|
- '4096'
|
|
- --disable-log-requests
|
|
- --disable-frontend-multiprocessing
|
|
- --enforce-eager
|
|
- --enable-auto-tool-choice
|
|
- --tool-call-parser
|
|
- qwen3_coder
|
|
- --reasoning-parser
|
|
- qwen3
|
|
- --enable-prefix-caching
|
|
- --enable-chunked-prefill
|
|
- --dtype
|
|
- half
|
|
env:
|
|
- name: VLLM_ENGINE_ITERATION_TIMEOUT_S
|
|
value: '3600'
|
|
- name: VLLM_ATTENTION_BACKEND
|
|
value: XFORMERS
|
|
- name: ENABLE_CUSTOM_IPC
|
|
value: '1'
|
|
- name: PYTHONPATH
|
|
value: /usr/local/corex/lib/python3/dist-packages:/usr/local/corex/lib64/python3/dist-packages
|
|
- name: LD_LIBRARY_PATH
|
|
value: /usr/local/corex/lib64:/usr/local/openmpi/lib
|
|
- name: VLLM_COREX_FA2_LIBRARY
|
|
value: /usr/local/corex/lib64/libcorex_fa2.so
|
|
- name: VLLM_COREX_GDN_LIBRARY
|
|
value: /usr/local/corex/lib64/libcorex_gdn.so
|
|
- name: VLLM_COREX_MOE_LIBRARY
|
|
value: /usr/local/corex/lib64/libcorex_moe.so
|
|
- name: VLLM_REQUEST_METRICS_FILE
|
|
value: /tmp/vllm-request-metrics.jsonl
|
|
- name: VLLM_CACHE_BLOCK_SIZE
|
|
value: '16'
|