iHore/vLLM-4x-RTX-3090
8
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

chore: Reset vLLM logging interval back to default

Browse Source 
Co-Authored-By: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
This commit is contained in:
MOVZX
2026-10-15 06:37:23 +00:00
co-authored by Hinata
parent 7e80cc67ee
commit 5cc9ee2505
4 changed files with 0 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.env.example
-2
View File
@@ -31,8 +31,6 @@ TOPO_TIMEOUT=20
GUARD=warn GUARD=warn
# 1 adds the two logging-only counters (exact-restart counters /30 s, FP8 KV clips /300 s). # 1 adds the two logging-only counters (exact-restart counters /30 s, FP8 KV clips /300 s).
COUNTERS=0 COUNTERS=0
# Seconds between engine stats lines (throughput, KV usage, queue depth).
VLLM_LOG_STATS_INTERVAL=1
# Extracted from orcarouter/Qwen3.8-Flash-Next-Uncensored, hidden 2560 # Extracted from orcarouter/Qwen3.8-Flash-Next-Uncensored, hidden 2560
QWEN38_ABLITERATION=Qwen3.8-Flash-Next-Uncensored QWEN38_ABLITERATION=Qwen3.8-Flash-Next-Uncensored
README.md
-1
View File
@@ -252,5 +252,4 @@ docker run -d --name vllm --gpus all --stop-timeout 20 --shm-size=16g \
| `CACHE_HOST_DIR` | `.vllm-cache` in this directory | Where the compile caches live. The host path uses the same folder. | | `CACHE_HOST_DIR` | `.vllm-cache` in this directory | Where the compile caches live. The host path uses the same folder. |
| `CUDA_VISIBLE_DEVICES` | `0,1,2,3` | The GPU lane, in bus order. | | `CUDA_VISIBLE_DEVICES` | `0,1,2,3` | The GPU lane, in bus order. |
| `PLE_HOME` | `0` | The card that hosts the PLE offload. That card gets no CUDA context for it. | | `PLE_HOME` | `0` | The card that hosts the PLE offload. That card gets no CUDA context for it. |
| `VLLM_LOG_STATS_INTERVAL` | `1` | Seconds between engine stats lines (throughput, KV use, queue). |
| `MAX_JOBS`, `ARCH` | `32`, `8.6` | Compile jobs and GPU architecture, for the build only. | | `MAX_JOBS`, `ARCH` | `32`, `8.6` | Compile jobs and GPU architecture, for the build only. |
docker-compose.yml
-1
View File
@@ -40,7 +40,6 @@ services:
TOPO_TIMEOUT: ${TOPO_TIMEOUT:-20} TOPO_TIMEOUT: ${TOPO_TIMEOUT:-20}
GUARD: ${GUARD:-warn} GUARD: ${GUARD:-warn}
COUNTERS: ${COUNTERS:-0} COUNTERS: ${COUNTERS:-0}
VLLM_LOG_STATS_INTERVAL: ${VLLM_LOG_STATS_INTERVAL:-1}
QWEN38_ABLITERATION: ${QWEN38_ABLITERATION:-Qwen3.8-Flash-Next-Uncensored} QWEN38_ABLITERATION: ${QWEN38_ABLITERATION:-Qwen3.8-Flash-Next-Uncensored}
volumes: volumes:
- "${MODEL_DIR:?Set MODEL_DIR in .env}:${MODEL_MOUNT_PATH:-/models/Qwen3.8-Flash-Next}:ro" - "${MODEL_DIR:?Set MODEL_DIR in .env}:${MODEL_MOUNT_PATH:-/models/Qwen3.8-Flash-Next}:ro"
start.sh
-2
View File
@@ -22,7 +22,6 @@
# PLE_HOME GPU that hosts the PLE offload home (0) # PLE_HOME GPU that hosts the PLE offload home (0)
# GUARD bounds-guard telemetry off|warn|raise (warn) # GUARD bounds-guard telemetry off|warn|raise (warn)
# COUNTERS=1 logging-only counters (top-k exact restarts /30s, FP8 KV clips /300s) # COUNTERS=1 logging-only counters (top-k exact restarts /30s, FP8 KV clips /300s)
# VLLM_LOG_STATS_INTERVAL seconds between engine stats lines (unset = vLLM default 1 s)
# CUDA_HOME toolkit for first-serve JIT; default: the venv's CUDA 13.0 wheels # CUDA_HOME toolkit for first-serve JIT; default: the venv's CUDA 13.0 wheels
# AUTO_TOPO 0 (default, fixed shipped settings) | 1 | strict -> scripts/topo_select.py # AUTO_TOPO 0 (default, fixed shipped settings) | 1 | strict -> scripts/topo_select.py
# AUTO_TOPO_NCCL=default AUTO_TOPO_BIND=1 LANE_GPUS=auto LANE_CPUS=... TOPO_TIMEOUT=20 # AUTO_TOPO_NCCL=default AUTO_TOPO_BIND=1 LANE_GPUS=auto LANE_CPUS=... TOPO_TIMEOUT=20
@@ -204,7 +203,6 @@ export VLLM_HOST_EMBED_TABLE=1
export VLLM_USE_BREAKABLE_CUDAGRAPH=0 export VLLM_USE_BREAKABLE_CUDAGRAPH=0
export VLLM_QSA_KV_SCALES="${SCALES:-$CKPT/qsa_kv_scales_swift.json}" export VLLM_QSA_KV_SCALES="${SCALES:-$CKPT/qsa_kv_scales_swift.json}"
# engine stats line (throughput, KV usage, queue). Set empty to keep vLLM's default. # engine stats line (throughput, KV usage, queue). Set empty to keep vLLM's default.
if [ -n "${VLLM_LOG_STATS_INTERVAL:-}" ]; then export VLLM_LOG_STATS_INTERVAL; fi
export HF_HUB_OFFLINE=1 HF_DATASETS_OFFLINE=1 export HF_HUB_OFFLINE=1 HF_DATASETS_OFFLINE=1
# the kernel toggles (the shipped values). These names are the tree's own: envs.py in the # the kernel toggles (the shipped values). These names are the tree's own: envs.py in the
# serving tree registers MERLIN_* for the kernel toggles and VLLM_E2_* for the runtime knobs. # serving tree registers MERLIN_* for the kernel toggles and VLLM_E2_* for the runtime knobs.