nodes/slahiri-ComfyUI-AI-Photography-Toolkit
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
134 Commits 13 Branches 8 Tags
dev/docs
Commit Graph
6 Commits
Author SHA1 Message Date
slahiriandClaude Opus 4.5 9d134da3e8 v4.2.0: Consolidate to 3 unified nodes 
Major refactor consolidating 8+ nodes into 3 unified nodes:

NEW NODES:
- SID_LLM_API: Unified cloud provider (15+ APIs including Anthropic, OpenAI, Gemini, Grok, Ollama, LM Studio)
- SID_LLM_Local: Unified local models (Florence-2, Moondream2, SmolVLM, Phi-3.5, QwenVL)
- SID_ZImagePromptGenerator: Unified prompt generator with auto Single-Shot/Agentic pipeline selection

DEPRECATED (removed):
- SID_Anthropic_LLM → use SID_LLM_API
- SID_OpenAI_Compatible_LLM → use SID_LLM_API
- SID_Grok_LLM → use SID_LLM_API
- SID_GGUF_LLM → use SID_LLM_Local
- SID_QwenVL_LLM → use SID_LLM_Local
- SID_ZImagePromptGenerator_Advanced_V2 → use SID_ZImagePromptGenerator

SPEED OPTIMIZATIONS (SID_LLM_Local):
- CUDA optimizations (cudnn.benchmark, TF32)
- BF16 auto-selection for Ampere+ GPUs
- Flash Attention 2 / SDPA auto-selection
- KV cache enabled for all models
- Memory cleanup on model switch

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

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-11 03:43:05 +05:30
slahiriandClaude Opus 4.5 af467b27cf Remove GGUF LLM provider node 
- Delete SID_GGUF_LLM node and related files
- Remove llm_providers/sid_gguf_llm.py
- Remove prompt_templates/local_gguf_template.py
- Remove sample_workflows/sample_workflow_gguf.json
- Remove GGUF provider from generator nodes
- Simplify template_factory.py to use Claude template for all providers
- Update pyproject.toml keywords
- QwenVL is now the recommended local LLM provider

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

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-11 02:10:16 +05:30
slahiriandClaude Opus 4.5 70457c5828 feat: Enhance SID_QwenVL_LLM with advanced options from QwenVL node 
- Add 12 model variants: Instruct, Thinking, FP8, and 32B models
- Add VRAM auto-downgrade with 20% safety margin
- Add attention_mode selection (auto, flash_attention_2, sdpa, eager)
- Add generation parameters: top_p, repetition_penalty, num_beams
- Thinking models now report supports_reasoning=True for agentic mode
- Pre-quantized FP8 models for fast loading
- Update generator _get_client() to pass new QwenVL parameters
- Bump version to 4.2.0

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

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-11 00:38:28 +05:30
slahiriandClaude Opus 4.5 10244097e0 feat: Add SID_QwenVL_LLM - local vision model using HuggingFace transformers 
Adds new QwenVL provider as alternative to GGUF:
- Uses HuggingFace transformers with BitsAndBytes quantization
- No llama-cpp-python required (easier cross-platform installation)
- Supports Qwen3-VL (2B/4B/8B) and Qwen2.5-VL (3B/7B) models
- Auto-downloads models from HuggingFace on first use
- Model caching option to keep loaded between runs

Changes:
- New llm_providers/sid_qwenvl_llm.py with QwenVLClient class
- Mark SID_GGUF_LLM as deprecated in favor of QwenVL
- Add qwenvl provider routing in generator _get_client
- Update template_factory.py for qwenvl provider
- Add transformers and accelerate to dependencies
- bitsandbytes as optional dependency for quantization

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

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-11 00:22:18 +05:30
slahiri 793b344a86 Add exclude patterns to pyproject.toml for registry publish 2025-12-10 20:14:01 +05:30
slahiriandClaude Opus 4.5 43fbb497d7 Add pyproject.toml for ComfyUI Registry 
Prepare repository for ComfyUI Registry publishing with:
- Project metadata (name, version, license, dependencies)
- Publisher ID and display name
- Icon URL

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

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-10 20:01:35 +05:30