Added qwen3.5-122b

create_new_image.sh

@@ -1,7 +1,7 @@
 #!/bin/bash
 export BASE_IMAGE="nvcr.io/nvidia/cuda:13.2.0-cudnn-devel-ubuntu24.04"
 
-docker run --name ik_llama_dw_$(date +%Y-%m-%d) --gpus=all $BASE_IMAGE /bin/bash -c \
+docker run --name ik-llama-dw-$(date +%Y-%m-%d) --gpus=all $BASE_IMAGE /bin/bash -c \
 'apt-get update
 apt-get --yes upgrade
 apt-get --yes install build-essential git libcurl4-openssl-dev curl libgomp1 cmake
@@ -9,11 +9,11 @@ git clone https://github.com/ikawrakow/ik_llama.cpp
 cd ik_llama.cpp
 cmake -B build -DGGML_NATIVE=ON -DGGML_CUDA=ON
 cmake --build build --config Release -j$(nproc)'
-docker commit ik_llama_dw_$(date +%Y-%m-%d) ik_llama_dw:$(date +%Y-%m-%d)
-docker rm ik_llama_dw_$(date +%Y-%m-%d)
+docker commit ik-llama-dw-$(date +%Y-%m-%d) ik-llama-dw:$(date +%Y-%m-%d)
+docker rm ik-llama-dw-$(date +%Y-%m-%d)
 
 #Updated env for other scripts
-IMAGE=ik_llama_dw:$(date +%Y-%m-%d)
+IMAGE=ik-llama-dw:$(date +%Y-%m-%d)
 
 cat > env.sh << EOF
 #!/bin/bash

env.sh

@@ -1,3 +1,3 @@
 #!/bin/bash
 #updated from create_new_image.sh
-export IK_IMAGE=ik_llama_dw:2026-03-18
+export IK_IMAGE=ik-llama-dw:2026-03-25

qwen3.5-122b.sh

@@ -0,0 +1,23 @@
+#!/bin/bash
+#grab container image name
+MODEL_NAME="Qwen3.5-122B-A10B"
+MODEL="Qwen3.5-122B-A10B-IQ4_KSS.gguf"
+MODEL_FOLDER="/sam4t/ENC/LLM/Qwen3.5"
+
+source ./env.sh
+echo $IK_IMAGE
+docker container create --name ik-$MODEL_NAME --network llms -p 10987:10987 -v $MODEL_FOLDER:/model --gpus=all --restart on-failure:3 \
+	--health-cmd "curl -f http://ik-$MODEL_NAME:10987/v1/models || exit 1" \
+	--health-interval 5s \
+	--health-timeout 2s \
+	--health-start-period 15s \
+	$IK_IMAGE \
+	/ik_llama.cpp/build/bin/llama-server \
+	--model /model/$MODEL \
+	--alias Qwen3.5-122B-A10B \
+	-ctk q8_0 -ctv q8_0 \
+	-ger -ngl 99 \
+	-ub 4096 -b 4096 \
+	--parallel 1 --threads 1 \
+	--host ik-$MODEL_NAME \
+	--port 10987 --jinja --no-mmap

qwen3.5-coder-next.sh

@@ -6,18 +6,18 @@ MODEL_FOLDER="/sam4t/ENC/LLM/Qwen3-Coder-Next"
 
 source ./env.sh
 echo $IK_IMAGE
-docker run --name ik_$MODEL_NAME -d --network llms -p 8654:8654 -v $MODEL_FOLDER:/model --gpus=all \
-	--health-cmd "curl -f http://ik_$MODEL_NAME:8654/v1/models || exit 1" \
-	--health-interval 15s \
-	--health-timeout 5s \
+docker container create --name ik-$MODEL_NAME --network llms -p 10987:10987 -v $MODEL_FOLDER:/model --gpus=all --restart on-failure:3 \
+	--health-cmd "curl -f http://ik-$MODEL_NAME:10987/v1/models || exit 1" \
+	--health-interval 5s \
+	--health-timeout 2s \
 	--health-start-period 15s \
 	$IK_IMAGE \
 	/ik_llama.cpp/build/bin/llama-server \
 	--model /model/$MODEL \
 	-ctk q8_0 -ctv q8_0 \
 	--parallel 1 --threads 32 \
-	--host ik_$MODEL_NAME \
-	--port 8654 -ngl 99 \
+	--host ik-$MODEL_NAME \
+	--port 10987 -ngl 99 \
 	--alias $MODEL_NAME \
 	--no-mmap --jinja \
 	-b 4096 -ub 4096 \