LuqP2
2a39183ffa
feat: add GPU/performance metrics collection and tracking
...
- Add collect_gpu_metrics() for VRAM/GPU auto-detection (CUDA/MPS/CPU)
- Add collect_version_info() for ComfyUI/PyTorch/Python versions
- Add optional override inputs (vram_peak_mb, gpu_device_override, generation_time_override)
- Add auto-timing wrapper with time.time() in save_images()
- Calculate derived metrics (generation_time_ms, steps_per_second)
- Update build_imh_metadata() analytics section with Tier 1/2/3 metrics
- Silent failures - never interrupts generation
- Reset CUDA peak memory stats after collection
2025-12-23 14:22:59 -03:00
LuqP2
238e458de6
feat: Implement MetaHub Save Image node for ComfyUI
...
Complete implementation of advanced image saving node with dual metadata:
- A1111/Civitai compatible metadata (tEXt chunk "parameters")
- Image MetaHub metadata (iTXt chunk "imagemetahub_data")
- Auto-detection of LoRAs from workflow (UI and API formats)
- SHA256 hash calculation (AutoV2/Civitai format) with caching
- Graceful degradation (never interrupts generation)
- IMH Pro fields support (user_tags, notes, project_name)
- Auto-increment filenames with customizable output paths
Files added:
- metahub_save_node.py: Main node class with INPUT_TYPES
- metadata_utils.py: Hash calculation, metadata formatting, PNG injection
- __init__.py: Node registration for ComfyUI
- requirements.txt: Dependencies (Pillow, numpy)
- README.md: Installation and usage documentation
100% compatible with Image MetaHub comfyUIParser.ts specification.
2025-12-19 20:59:47 -03:00
LuqP2