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

Remove Beta 1 and Piper fallback

Browse Source
This commit is contained in:
Mikei386
2026-09-10 13:17:57 +02:00
parent 52482627be
commit 08ff3d7c4e
38 changed files with 94 additions and 484 deletions
Download Patch File Download Diff File Expand all files Collapse all files
install.sh
+2 -12
View File
@@ -41,7 +41,6 @@ source "$CONFIG"
required=(AI_HOSTNAME ADMIN_USER MODEL_DIR FAST_MODEL_FILE
FAST_MODEL_URL FAST_MODEL_SHA256 MEDIUM_MODEL_FILE MEDIUM_MODEL_URL
MEDIUM_MODEL_SHA256 LARGE_MODEL_FILE LARGE_MODEL_URL LARGE_MODEL_SHA256
BETA1_MODEL_FILE BETA1_MODEL_URL BETA1_MODEL_SHA256
ULTRA_MODEL_FILE ULTRA_MODEL_URL ULTRA_MODEL_SHA256
UNCENSORED_MODEL_FILE UNCENSORED_MODEL_URL UNCENSORED_MODEL_SHA256
UNCENSORED_PROJECTOR_FILE UNCENSORED_PROJECTOR_URL
@@ -314,8 +313,6 @@ MODEL_DIR=$MODEL_DIR
WIREGUARD_CONFIG_FILE=${WIREGUARD_CONFIG_FILE:-/etc/mike-ai/wireguard/fritz-athena.conf}
ROUTER_API_KEY=$(<$SECRETS_DIR/router-api-key)
CONTROLLER_TOKEN=$(<$SECRETS_DIR/controller-token)
PIPER_TTS_VERSION=${PIPER_TTS_VERSION:-1.6.0}
PIPER_VOICE=${PIPER_VOICE:-de_DE-thorsten-high}
QWEN3_TTS_IMAGE=${QWEN3_TTS_IMAGE:-ghcr.io/malaiwah/qwen3-tts-server:latest@sha256:b363a01d08b1bbecbfc3ca6f585368fae2cfdc591f9ecca6643738369f9a9d98}
QWEN3_TTS_CACHE_DIR=$QWEN3_TTS_CACHE_DIR
QWEN3_TTS_VOICES_DIR=$QWEN3_TTS_VOICES_DIR
@@ -323,7 +320,6 @@ QWEN3_TTS_GPU_DEVICE=${QWEN3_TTS_GPU_DEVICE:-GPU-4834d9d7-5b61-3004-1fb3-4ae49d4
AI_DNS=${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
@@ -337,10 +333,6 @@ 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=${BETA1_CONTEXT:-112000}
BETA1_BATCH_SIZE=${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=${LARGE_UBATCH_SIZE:-128}
@@ -356,7 +348,6 @@ UNCENSORED_PARALLEL_SLOTS=${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}
@@ -405,7 +396,6 @@ 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|$UNCENSORED_MODEL_SHA256
@@ -514,11 +504,11 @@ build_and_start() {
# Portable MCPs and Hermes live on Unraid and are restored through Appdata.
"$STACK_DIR/platform/mcp/install-tools.sh"
docker compose --env-file "$SECRETS_DIR/stack.env" --profile inference create \
llama-fast llama-medium llama-beta1 llama-large llama-ultra llama-uncensored
llama-fast llama-medium llama-large llama-ultra llama-uncensored
docker compose --env-file "$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" up -d --build \
wireguard-gateway qwen3-tts piper tts-gateway profile-controller router llama-dashboard portainer backup
wireguard-gateway qwen3-tts tts-gateway profile-controller router llama-dashboard portainer backup
if [[ ${WIREGUARD_MODE:-container} == container ]]; then
systemctl restart mike-ai-container-vpn-guard.service
fi