Spaces:
Sleeping
Sleeping
Commit History
-
Fix zero-config demo: default profile groq_cloud + Gemini image keyless fallback
- profiles.py: change default INFERENCE_PROFILE from gemini_cloud to groq_cloud (Pollinations images need no API key, so a fresh Space deploy produces real images + local depth + mock audio/video with no secrets configured). - image_generator.py: Gemini image backend now falls back to free Pollinations image generation when GEMINI_API_KEY is missing, instead of crashing the whole pipeline.ec1346a -
Make demo deployable on Hugging Face Spaces + add local/Groq/demo onboarding
- Fix pipeline.py docstring syntax error (unterminated triple-quoted string) - config.py: default MOCK_INFERENCE=False so a clean deploy runs the real hybrid pipeline - PromptExpander: gracefully degrade (local template) when no cloud LLM key; add 'local' backend branch for local_gpu profile - ImageGenerator: Pollinations failure falls back to a mock image so depth/video still work - main.py: SPA catch-all + /static mount so /studio/:id deep links survive refresh - profiles.py: replace removed runwayml/stable-diffusion-v1-5 with stable-diffusion-v1-5/stable-diffusion-v1-5 - KeySetupModal: 3-option chooser (Run locally / Use Groq / Just demo) - README: add mode decision tree + 'Run locally (full models)' section - index.html title + backend/__init__.py + gitignore (frontend/dist)29e37f2 -
feat: BYOK modal - visitors enter their own Groq key, stored in sessionStorage, auto-deleted on tab close
9c34cbf -
feat: add groq_cloud profile with Groq LLM + Pollinations image generation (fully free, no credits)
7a825c2 -
feat: add profiles.py control panel and Gemini backend for prompt expansion and image generation
b9589ac -
fix: restore FLUX.1-schnell and Qwen2.5 LLM configurations in local config
40b3d24 -
fix: update to supported models (zephyr-7b-beta and stable-diffusion-v1-5)
02818c9 -
fix: update model IDs to avoid credit depletion and chat model errors
637853d -
feat: initial release of Multiverse AI Studio
4aed5f4
