Eric Hiss
00c692782d
Major Release - Complete Advanced Node Redesign ✨ New Features: - Advanced node: 4 operation modes (expand/refine/modify_style/add_details) - Advanced node: Clear detail levels (concise/moderate/detailed/exhaustive) - Advanced node: Optional image input with Qwen3-VL integration - Video nodes: Emphasis syntax preservation (keyword:1.5) - Video nodes: Alternation syntax support {option1|option2} 🔧 Changes: - Replaced confusing tier system with intuitive detail levels - Added operation modes for modifying existing prompts - Removed manual mode selection (auto-detected from image presence) - Added Qwen3-VL backend for image analysis - Backward compatible tier name mapping in expansion_engine 📚 Documentation: - README.md completely rewritten - docs/ADVANCED_NODE_REDESIGN.md - Complete redesign guide - docs/BUGFIX_ADVANCED_NODE.md - Syntax preservation details - Organized old docs into docs/archive/ 🐛 Fixes: - Emphasis syntax now properly preserved through LLM processing - Alternation syntax for random element selection - Image-to-video mode now functional with Qwen3-VL ⚡ Breaking Changes: None (fully backward compatible) See RELEASE_NOTES_v1.8.0.md for complete details.
27 lines
744 B
Plaintext
27 lines
744 B
Plaintext
# Requirements for Eric's Prompt Enhancers for ComfyUI
|
|
# Most dependencies are already provided by ComfyUI
|
|
|
|
# Core dependency for LLM API calls
|
|
requests>=2.25.0
|
|
|
|
# Image processing (usually included with ComfyUI)
|
|
# Pillow>=8.0.0
|
|
|
|
# Tensor operations (usually included with ComfyUI)
|
|
# torch>=1.10.0
|
|
# numpy>=1.19.0
|
|
|
|
# Note: PyTorch, NumPy, and Pillow are typically already installed
|
|
# with ComfyUI. Only install if you encounter import errors.
|
|
|
|
# Optional: LLM Backends (install separately)
|
|
# LM Studio: https://lmstudio.ai/ (recommended)
|
|
# Ollama: https://ollama.ai/
|
|
|
|
# Optional: Local Qwen3-VL backend support
|
|
# transformers>=4.42.0
|
|
# accelerate>=0.30.0
|
|
# huggingface_hub>=0.23.0
|
|
# bitsandbytes>=0.43.0 # required for 4bit/8bit quantization
|
|
|