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
1.9 KiB
1.9 KiB
| 1 | # Read by: c3 (services.py STUDIO_MODELS loader) · gpu-mode.sh (preflight_studio_models) · | |||||
|---|---|---|---|---|---|---|
| 2 | # the c3 download modal. Edit HERE only — every surface parses this file (no canary drift). | |||||
| 3 | # Columns (TAB-separated): modality<TAB>label<TAB>root<TAB>rel_path<TAB>size_gb<TAB>installer | |||||
| 4 | # modality : director | image | video | audio (director is the hard requirement; rest warn) | |||||
| 5 | # root : weights (HF weights root, MODEL_DIR) | comfy (ComfyUI models tree, COMFYUI_MODELS_DIR) | |||||
| 6 | # rel_path : the canary file under <root> whose presence == the model is installed | |||||
| 7 | # installer: repo-relative download script (`bash <installer>`) that fetches it | |||||
| 8 | # Roster note: the uncensored picks (Krea / Z-Image / Wan) get appended after the bake-off. | |||||
| 9 | director | Director (Qwen3.5-4B uncensored) | weights | qwen3.5-4b-gguf/hauhaucs-uncensored-q4km/Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf | 2.5 | services/comfyui/download_director.sh |
| 10 | image | Ideogram-4 (design / logo / text) | comfy | diffusion_models/ideogram4_fp8_scaled.safetensors | 9.0 | services/comfyui/download_ideogram4.sh |
| 11 | video | LTX-2.3 (video + audio) | comfy | unet/ltx2.3/distilled-1.1/ltx-2.3-22b-distilled-1.1-Q8_0.gguf | 22.8 | services/comfyui/download_video_models.sh |
| 12 | video | Sulphur (uncensored video) | comfy | unet/sulphur-2/sulphur_dev-Q8_0.gguf | 22.8 | services/comfyui/download_video_models.sh |
| 13 | video | 10Eros (uncensored video) | comfy | unet/10eros/10Eros_v1-Q8_0.gguf | 22.8 | services/comfyui/download_video_models.sh |
| 14 | audio | ACE-Step (music) | comfy | checkpoints/ace-step-1.5/all_in_one/ace_step_v1_3.5b.safetensors | 7.7 | services/comfyui/download_ace_step.sh |
| 15 | audio | Stable Audio (SFX) | comfy | checkpoints/stable-audio-open-1.0.safetensors | 4.85 | services/comfyui/download_stable_audio.sh |
| 16 | audio | Kokoro (narration) | comfy | tts/kokoro/kokoro-v1.0.onnx | 0.3 | services/comfyui/download_kokoro.sh |