noonghunnaandClaude Opus 4.8
3fc26857f2
Replace the separate `image-studio` / `video-studio` / `comfyui` scenes with
a single `ai-studio` scene: ComfyUI holds both GPUs always, and image / video /
audio are picked as *lanes* in OWUI rather than gpu-mode switches. Removes the
gemma-12b chat brain from the studio bundle (kept on disk, just unwired) and
points OWUI's DEFAULT_MODELS at the qwen director.
- gpu-mode.sh: `mode_ai_studio` replaces the two studio modes; drop the gemma
start/stop + the GPU0-pin split; preflight now reads the shared manifest.
- scripts/lib/studio-models.tsv: single source-of-truth model manifest (modality,
label, root, rel_path, size, installer) consumed by both gpu-mode preflight and
the c3 cockpit, so the two can't drift.
- download_{director,video,audio,ace_step,stable_audio,studio}_models.sh: idempotent
fetchers; download_studio_models.sh orchestrates "grab everything".
- setup-video-studio.sh + repointed setup-image-studio.sh → `gpu-mode ai-studio`.
- litellm: drop the gemma-4-12b :8069 route; comfyui compose/entrypoint GPU-pin
comments corrected (no scene sets the pin now).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EfF565T9eSLaqGzidyJ1Pm
148 lines
7.8 KiB
Bash
Executable File
148 lines
7.8 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
# One-shot setup for the club-3090 IMAGE-STUDIO bundle:
|
|
# ComfyUI (Ideogram-4 image gen) + Open WebUI front-end + gemma-4-12b chat,
|
|
# coexisting on a 2-GPU box (ComfyUI -> GPU0, chat -> GPU1).
|
|
#
|
|
# Usage:
|
|
# bash scripts/setup-image-studio.sh # build + download + bring up (asks to confirm)
|
|
# bash scripts/setup-image-studio.sh --yes # skip the confirmation prompt
|
|
#
|
|
# Env knobs:
|
|
# SKIP_BUILD=1 skip the ComfyUI image build (already built)
|
|
# SKIP_DOWNLOAD=1 skip the ~27 GB Ideogram-4 weight pull (already on disk)
|
|
# ASSUME_YES=1 same as --yes (also auto-yes when not a TTY / under CI)
|
|
# LANIP=<ip> host IP shown in the final URLs (auto-detected otherwise)
|
|
#
|
|
# Idempotent: re-running rebuilds/re-pulls only what changed, then brings the
|
|
# stack up via `gpu-mode image-studio`.
|
|
set -euo pipefail
|
|
|
|
REPO_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
|
|
COMFYUI_DIR="$REPO_DIR/services/comfyui"
|
|
LANIP="${LANIP:-$(hostname -I 2>/dev/null | tr ' ' '\n' | grep -E '^(192\.168|10\.|172\.)' | head -1)}"
|
|
LANIP="${LANIP:-<host-ip>}"
|
|
|
|
ASSUME_YES="${ASSUME_YES:-}"
|
|
case "${1:-}" in
|
|
-y|--yes) ASSUME_YES=1 ;;
|
|
-h|--help) sed -n '2,18p' "$0" | sed 's/^# \{0,1\}//'; exit 0 ;;
|
|
esac
|
|
# Auto-yes when non-interactive (piped / nohup / CI) so we never hang on the prompt.
|
|
[ -t 0 ] || ASSUME_YES=1
|
|
[ -n "${CI:-}" ] && ASSUME_YES=1
|
|
|
|
say() { echo -e "\033[0;36m$*\033[0m"; }
|
|
warn() { echo -e "\033[1;33m$*\033[0m"; }
|
|
ok() { echo -e "\033[0;32m$*\033[0m"; }
|
|
|
|
# --- 0. Preflight + plan ----------------------------------------------------
|
|
# Reuse the repo's preflight library (docker / gpu / disk), then add image-studio
|
|
# specifics. Hard-fail before the long build/download if something's missing.
|
|
# shellcheck disable=SC1091
|
|
. "$REPO_DIR/scripts/preflight.sh" 2>/dev/null || true
|
|
MODEL_DIR_RESOLVED="$(grep -E '^MODEL_DIR=' "$REPO_DIR/.env" 2>/dev/null | tail -1 | cut -d= -f2-)"
|
|
MODEL_DIR_RESOLVED="${MODEL_DIR_RESOLVED:-/mnt/models/huggingface}"
|
|
if declare -f preflight_docker >/dev/null 2>&1; then
|
|
preflight_docker || exit 1 # docker + compose v2 + daemon
|
|
preflight_gpu 1 || exit 1 # >=1 GPU (coexist needs 2 — warned below)
|
|
[ -z "${SKIP_BUILD:-}" ] && { preflight_disk / 32 || exit 1; } # comfyui-local image (~27 GB) + base pull
|
|
[ -z "${SKIP_DOWNLOAD:-}" ] && { preflight_disk /mnt/models/comfyui 30 || exit 1; } # Ideogram-4 set (~27 GB)
|
|
preflight_gpu_idle || true # soft warn if VRAM already in use
|
|
else
|
|
command -v docker >/dev/null 2>&1 || { echo "ERROR: docker not found." >&2; exit 1; }
|
|
fi
|
|
# Image-studio specifics (not in preflight.sh):
|
|
if [ -z "${SKIP_DOWNLOAD:-}" ] && ! command -v hf >/dev/null 2>&1; then
|
|
echo "[preflight] ERROR: 'hf' (huggingface_hub CLI) not found — needed for the weight download." >&2
|
|
echo " Fix: pip install -U huggingface_hub (or SKIP_DOWNLOAD=1 if weights are present)." >&2
|
|
exit 1
|
|
fi
|
|
GEMMA_GGUF="$MODEL_DIR_RESOLVED/gemma-4-12b-gguf/unsloth-ud-q8kxl/gemma-4-12b-it-UD-Q8_K_XL.gguf"
|
|
GEMMA_MISSING=""
|
|
if [ ! -f "$GEMMA_GGUF" ]; then
|
|
GEMMA_MISSING=1
|
|
echo "[preflight] WARN: gemma-4-12b GGUF not found at $GEMMA_GGUF" >&2
|
|
echo " The chat model won't start until it's present (image gen still works). Fetch it:" >&2
|
|
echo " hf download unsloth/gemma-4-12b-it-GGUF gemma-4-12b-it-UD-Q8_K_XL.gguf \\" >&2
|
|
echo " --local-dir \"$(dirname "$GEMMA_GGUF")\"" >&2
|
|
fi
|
|
NGPU=$(nvidia-smi -L 2>/dev/null | wc -l)
|
|
|
|
say "═══ club-3090 image-studio setup ═══"
|
|
echo " repo: $REPO_DIR"
|
|
echo " GPUs: $NGPU"
|
|
echo ""
|
|
echo " This will:"
|
|
[ -z "${SKIP_BUILD:-}" ] && echo " • build the ComfyUI image (comfyui-local) — pulls a ~9 GB CUDA base (one-time, slow)" \
|
|
|| echo " • (skip build — SKIP_BUILD set)"
|
|
[ -z "${SKIP_DOWNLOAD:-}" ] && echo " • download the Ideogram-4 model set (~27 GB) into the ComfyUI models tree" \
|
|
|| echo " • (skip download — SKIP_DOWNLOAD set)"
|
|
echo " • start the bundle via 'gpu-mode image-studio':"
|
|
echo " - ComfyUI / Ideogram-4 → GPU 0, port 8188"
|
|
echo " - gemma-4-12b chat → GPU 1, port 8069"
|
|
echo " - Open WebUI + LiteLLM + SearXNG (always-on)"
|
|
echo ""
|
|
if [ "${NGPU:-0}" -lt 2 ]; then
|
|
warn " ⚠ <2 GPUs — image gen and a local chat model can't run at once (GPU-mutex)."
|
|
warn " The bundle will run ComfyUI image gen only; for chat use 'gpu-mode chat'."
|
|
fi
|
|
if [ -z "$ASSUME_YES" ]; then
|
|
printf " Proceed? [y/N] "
|
|
read -r reply
|
|
case "$reply" in [yY]|[yY][eE][sS]) ;; *) echo " aborted."; exit 0 ;; esac
|
|
fi
|
|
|
|
# --- 1. Build the ComfyUI image (clones a pinned ComfyUI + nodes on first boot) ---
|
|
if [ -z "${SKIP_BUILD:-}" ]; then
|
|
say "── [1/3] Building ComfyUI image (comfyui-local:latest) ──"
|
|
(cd "$COMFYUI_DIR" && sudo docker compose build)
|
|
else
|
|
echo " (SKIP_BUILD set — skipping image build)"
|
|
fi
|
|
|
|
# --- 2. Download the model sets (Ideogram-4 + director + audio lanes) --------
|
|
if [ -z "${SKIP_DOWNLOAD:-}" ]; then
|
|
say "── [2/3] Downloading model sets (~43 GB; skip with SKIP_DOWNLOAD=1) ──"
|
|
echo " • Ideogram-4 image model set (~27 GB)"
|
|
bash "$COMFYUI_DIR/download_ideogram4.sh"
|
|
# The studio DIRECTOR (qwen3.5-4b, GPU0) crafts the prompt behind the 🖼️ image
|
|
# button — without it studio-director boots but has no model, so the button
|
|
# fails. GGUF + vision mmproj (~2.7 GB) → MODEL_DIR/qwen3.5-4b-gguf/… (where the
|
|
# enhancer compose's -m / --mmproj defaults point).
|
|
echo " • Studio director GGUF (Qwen3.5-4B-Uncensored, ~2.7 GB + mmproj)"
|
|
hf download HauhauCS/Qwen3.5-4B-Uncensored-HauhauCS-Aggressive \
|
|
Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf \
|
|
mmproj-Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf \
|
|
--local-dir "$MODEL_DIR_RESOLVED/qwen3.5-4b-gguf/hauhaucs-uncensored-q4km"
|
|
# Audio lanes (🎵 music · 🔊 SFX · 🗣️ narration) — shared across the ai-studio scene.
|
|
echo " • Studio audio models (🎵 music · 🔊 SFX · 🗣️ narration; ~13 GB)"
|
|
bash "$COMFYUI_DIR/download_audio_models.sh"
|
|
else
|
|
echo " (SKIP_DOWNLOAD set — skipping weight download)"
|
|
fi
|
|
|
|
# --- 3. Bring the stack up via gpu-mode -------------------------------------
|
|
say "── [3/3] Starting the bundle (gpu-mode ai-studio) ──"
|
|
bash "$REPO_DIR/scripts/gpu-mode.sh" ai-studio
|
|
|
|
# --- Done — onboarding -------------------------------------------------------
|
|
echo ""
|
|
ok "═══ Image-studio ready ═══"
|
|
echo " Open WebUI: http://$LANIP:8080 ← start here (chat + image)"
|
|
echo " ComfyUI: http://$LANIP:8188 ← optional: full node-graph control"
|
|
echo ""
|
|
say " Get started:"
|
|
echo " 1. Open the Open WebUI URL and SIGN UP — the FIRST account you create"
|
|
echo " becomes the admin. (No credentials are pre-set; you choose them here.)"
|
|
echo " 2. Pick the chat model 'gemma-4-12b…' in the top selector and chat normally."
|
|
echo " 3. Generate an image: send a prompt, then click the 🖼️ image icon on the"
|
|
echo " reply — it renders via Ideogram-4 on ComfyUI. (Image gen is pre-wired."
|
|
echo " Toggle/inspect it under Admin → Settings → Images.)"
|
|
echo ""
|
|
warn " Notes:"
|
|
warn " • First image after a cold ComfyUI is slow (~2 min, loads ~20 GB); warm ~70 s."
|
|
warn " • 🖼️ button missing / image gen unconfigured? You reused an existing Open WebUI"
|
|
warn " volume — the image-gen env only auto-applies on a FRESH volume. Set it in"
|
|
warn " Admin → Settings → Images (Engine=ComfyUI, Base URL=http://host.docker.internal:8188)."
|
|
warn " • 2048² fits but is tight (~22 GB, batch 1); prefer 1024² + upscale for routine high-res."
|