17 Commits
Author SHA1 Message Date
dseditor c1c57dc5c9 adding frame node 2026-02-09 00:11:36 +08:00
dseditor 8382b7247e 重構 PhotoMagazine 節點:模板化架構
- 提示詞注入器:從 prompts/ 資料夾讀取模板並注入參數
- JSON 解析器:改為 forceInput,只接受 STRING 連接
- 移除多行文字框,改用節點連接方式
- 風格改為完全自由輸入
- 新增 prompts/photomagazine_json_output.md 模板檔案
2026-01-04 15:21:46 +08:00
dseditor 8143bac1f5 i18nandnodeupdte 2026-01-03 15:58:40 +08:00
dseditor 0fff1e5ba5 S3C 2026-01-01 14:25:39 +08:00
dseditorandClaude Sonnet 4.5 3f55d57bb7 Add auto-install feature for BatchToPSD and update documentation
- Implement automatic psd-tools package detection and installation
- Display installation status message and restart prompt
- Add comprehensive BatchToPSD documentation (English and Chinese)
- Include usage examples, troubleshooting guide, and technical details
- Prevent import failures when psd-tools is not installed

功能更新:
- BatchToPSD 新增套件自動安裝機制
- 首次使用自動安裝 psd-tools 並提示重啟
- 完整的中英文使用說明文件
- 包含使用範例、疑難排解和技術細節

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
2025-12-26 19:23:43 +08:00
dseditorandClaude Sonnet 4.5 ed027a16ef Add BatchToPSD node for converting batch images to multi-layer PSD
- Converts batch images to PSD file with layers
- Skips first image in batch (merged result)
- Supports RGBA transparency preservation
- Auto-detects alpha channels and uses appropriate PSD mode
- Includes layer order reversal option

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
2025-12-26 19:20:32 +08:00
dseditor 85e7d0482c simplewildcard 2025-12-05 22:26:24 +08:00
dseditor 7e73ef1db1 simplewildcard 2025-12-05 22:16:28 +08:00
dseditor 7e6f2430f7 v113-GGUF 2025-12-05 12:27:52 +08:00
dseditor 8433b3fd5d TELLM Update 2025-12-04 15:44:32 +08:00
dseditor 4bd28f74c6 qwen gpu node 2025-12-04 15:30:58 +08:00
dseditorandClaude a90f71066c Add Qwen GPU Inference node with intelligent memory management
- Add QwenGPUInference node for AI photo prompt optimization
- Implement smart GPU memory management with automatic detection
- Support CPU offload when GPU memory is insufficient
- Auto-download model config files from HuggingFace
- Remove <think> tags from model output
- Add bilingual (Chinese/English) support
- Remove deprecated GGUF inference node and related files
- Update README with comprehensive documentation

Features:
- Automatic model detection (qwen_3_4b.safetensors)
- Three loading strategies: Full GPU / CPU Offload / CPU-only
- Memory conflict prevention with ComfyUI models
- Professional photography prompt generation
- Default max_tokens: 2048 for detailed prompts
- Custom system prompt for photo optimization

Performance:
- Full GPU: ~26-30 tokens/second
- CPU Offload: ~1-2 tokens/second (reliable fallback)
- First load: 7-130 seconds depending on hardware
- Subsequent loads: Near-instant (model cached)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-04 11:15:12 +08:00
dseditor df9f3ac0a7 remove ffmpeg req 2025-09-04 08:24:07 +08:00
dseditor 434cc205c8 add llm nodes 2025-09-02 22:50:25 +08:00
dseditor 7be3ba33d7 change nodes 2025-07-07 08:13:40 +08:00
dseditor 25eb1ad213 Adding Split Fade 2025-07-06 01:15:44 +08:00
dseditor 026ffa3b47 Initial commit for ComfyUI-ListHelper 2025-06-21 16:47:32 +08:00