jdelony/ik_llama_docker
Template
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Increased start time, removed cache

Browse Source
This commit is contained in:
Jared Delony 2026-05-18 13:32:17 -05:00
parent 3591a33d37
commit 441b1fa7a0
Signed by: jdelony
SSH Key Fingerprint: SHA256:W5F7KMb6FLjlHNhQhtrDl4NvR5fCfinaDDXSQnA2sF0
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
qwen3.5-397B-A17B-IQ4_KSS.sh
View File

@ -19,7 +19,7 @@ docker container create --name $CONT_NAME --network llms -p $PORT:$PORT -v $MODE
--health-cmd "curl -f http://$CONT_NAME:$PORT/v1/models || exit 1" \ --health-cmd "curl -f http://$CONT_NAME:$PORT/v1/models || exit 1" \
--health-interval 5s \ --health-interval 5s \
--health-timeout 5s \ --health-timeout 5s \
--health-start-period 20s \ --health-start-period 50s \
$IMAGE \ $IMAGE \
/ik_llama.cpp/build/bin/llama-server \ /ik_llama.cpp/build/bin/llama-server \
--model /model/$MODEL \ --model /model/$MODEL \
@ -28,7 +28,7 @@ docker container create --name $CONT_NAME --network llms -p $PORT:$PORT -v $MODE
-ngl 999 \ -ngl 999 \
--n-cpu-moe 38 \ --n-cpu-moe 38 \
--parallel 2 --threads 32 \ --parallel 2 --threads 32 \
-cram 32768 -cram-n-min 128 -crs .9 --ctx-checkpoints 512 \ --cache-ram 0 \
--host $CONT_NAME \ --host $CONT_NAME \
--port $PORT --jinja --no-mmap --port $PORT --jinja --no-mmap