michael/AI-Profile-Router
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add Qwen GSQ-RCO Beta 1 profile

Browse Source
This commit is contained in:
Mikei386
2026-09-04 14:36:50 +02:00
parent 744a207e5a
commit f553108912
11 changed files with 172 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
install.sh
+9 -1
View File
@@ -41,6 +41,7 @@ source "$CONFIG"
required=(AI_HOSTNAME ADMIN_USER MODEL_DIR FAST_MODEL_FILE
FAST_MODEL_URL FAST_MODEL_SHA256 MEDIUM_MODEL_FILE MEDIUM_MODEL_URL
MEDIUM_MODEL_SHA256 LARGE_MODEL_FILE LARGE_MODEL_URL LARGE_MODEL_SHA256
BETA1_MODEL_FILE BETA1_MODEL_URL BETA1_MODEL_SHA256
ULTRA_MODEL_FILE ULTRA_MODEL_URL ULTRA_MODEL_SHA256
UNCENSORED_MODEL_FILE UNCENSORED_MODEL_URL UNCENSORED_MODEL_SHA256
UNCENSORED_PROJECTOR_FILE UNCENSORED_PROJECTOR_URL
@@ -319,6 +320,7 @@ XTTS_GPU_DEVICE=${XTTS_GPU_DEVICE:-GPU-4834d9d7-5b61-3004-1fb3-4ae49d482d4b}
AI_DNS=${WG_DNS:-1.1.1.1}
FAST_MODEL_FILE=$FAST_MODEL_FILE
MEDIUM_MODEL_FILE=$MEDIUM_MODEL_FILE
BETA1_MODEL_FILE=$BETA1_MODEL_FILE
LARGE_MODEL_FILE=$LARGE_MODEL_FILE
ULTRA_MODEL_FILE=$ULTRA_MODEL_FILE
UNCENSORED_MODEL_FILE=$UNCENSORED_MODEL_FILE
@@ -332,6 +334,10 @@ MEDIUM_CONTEXT=${MEDIUM_CONTEXT:-160000}
MEDIUM_BATCH_SIZE=${MEDIUM_BATCH_SIZE:-2048}
MEDIUM_UBATCH_SIZE=${MEDIUM_UBATCH_SIZE:-128}
MEDIUM_PARALLEL_SLOTS=${MEDIUM_PARALLEL_SLOTS:-1}
BETA1_CONTEXT=${BETA1_CONTEXT:-192000}
BETA1_BATCH_SIZE=${BETA1_BATCH_SIZE:-2048}
BETA1_UBATCH_SIZE=${BETA1_UBATCH_SIZE:-128}
BETA1_PARALLEL_SLOTS=${BETA1_PARALLEL_SLOTS:-1}
LARGE_CONTEXT=${LARGE_CONTEXT:-192000}
LARGE_BATCH_SIZE=${LARGE_BATCH_SIZE:-2048}
LARGE_UBATCH_SIZE=${LARGE_UBATCH_SIZE:-128}
@@ -347,6 +353,7 @@ UNCENSORED_PARALLEL_SLOTS=${UNCENSORED_PARALLEL_SLOTS:-1}
FAST_GPU_DEVICES=${TEXT_GPU_DEVICES:-0}${SECONDARY_GPU_DEVICES:+,$SECONDARY_GPU_DEVICES}
MEDIUM_GPU_DEVICES=${TEXT_GPU_DEVICES:-0}${SECONDARY_GPU_DEVICES:+,$SECONDARY_GPU_DEVICES}
MEDIUM_TENSOR_SPLIT=${MEDIUM_TENSOR_SPLIT:-90,10}
BETA1_GPU_DEVICES=${TEXT_GPU_DEVICES:-0}${SECONDARY_GPU_DEVICES:+,$SECONDARY_GPU_DEVICES}
LARGE_GPU_DEVICES=${TEXT_GPU_DEVICES:-0}${SECONDARY_GPU_DEVICES:+,$SECONDARY_GPU_DEVICES}
LARGE_TENSOR_SPLIT=${LARGE_TENSOR_SPLIT:-86,14}
ULTRA_GPU_DEVICES=${TEXT_GPU_DEVICES:-0}${SECONDARY_GPU_DEVICES:+,$SECONDARY_GPU_DEVICES}
@@ -394,6 +401,7 @@ download_models() {
done <<EOF
$FAST_MODEL_FILE|$FAST_MODEL_URL|$FAST_MODEL_SHA256
$MEDIUM_MODEL_FILE|$MEDIUM_MODEL_URL|$MEDIUM_MODEL_SHA256
$BETA1_MODEL_FILE|$BETA1_MODEL_URL|$BETA1_MODEL_SHA256
$LARGE_MODEL_FILE|$LARGE_MODEL_URL|$LARGE_MODEL_SHA256
$ULTRA_MODEL_FILE|$ULTRA_MODEL_URL|$ULTRA_MODEL_SHA256
$UNCENSORED_MODEL_FILE|$UNCENSORED_MODEL_URL|$UNCENSORED_MODEL_SHA256
@@ -488,7 +496,7 @@ build_and_start() {
# Portable MCPs and Hermes live on Unraid and are restored through Appdata.
"$STACK_DIR/platform/mcp/install-tools.sh"
docker compose --env-file "$SECRETS_DIR/stack.env" --profile inference create \
llama-fast llama-medium llama-large llama-ultra llama-uncensored
llama-fast llama-medium llama-beta1 llama-large llama-ultra llama-uncensored
docker compose --env-file "$SECRETS_DIR/stack.env" --profile image create image-worker
docker volume inspect portainer_data >/dev/null 2>&1 || docker volume create portainer_data >/dev/null
docker compose --env-file "$SECRETS_DIR/stack.env" stop llama-dashboard portainer