Changes:
- Added Stable Diffusion v1.5 model to comfyui_models.yaml
- Required for AnimateDiff motion modules (mm_sd_v15_v2.ckpt)
- Size: 4GB, VRAM: 8GB
- Updated AnimateDiff workflow to use SD 1.5 instead of SDXL
- Changed checkpoint from sd_xl_base_1.0.safetensors to v1-5-pruned-emaonly.safetensors
- Updated VRAM requirement from 18GB to 12GB
- Updated model requirements in metadata
AnimateDiff v15 motion modules are not compatible with SDXL.
This resolves: "Motion module 'mm_sd_v15_v2.ckpt' is intended for SD1.5 models"
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Fixed widget_values order for KSamplerAdvanced node:
- Changed sampler from dpmpp_2m to euler (valid sampler)
- Changed scheduler from karras to normal (valid scheduler)
- Fixed end_at_step from 10000 to 25 (matching steps count)
This resolves validation errors:
- scheduler '0' not in list
- cfg could not convert 'dpmpp_2m' to float
- sampler_name 'karras' not in list
- end_at_step 'disable' invalid literal for int
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Change CheckpointLoaderSimple to use 'sd_xl_base_1.0.safetensors' instead of
'diffusers/stable-diffusion-xl-base-1.0' to match the actual checkpoint file
linked in /root/ComfyUI/models/checkpoints/
- Replace deprecated AnimateDiffLoaderV1 with ADE_LoadAnimateDiffModel
- Add ADE_ApplyAnimateDiffModelSimple to create M_MODELS
- Add ADE_UseEvolvedSampling to inject motion into base model
- Replace AnimateDiffSampler with KSamplerAdvanced
- Add complete node links for all connections
- Update node IDs and execution order
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Replace ADE_AnimateDiffSampler with KSamplerAdvanced
- Adjust widget values for KSamplerAdvanced compatibility
- Add proper node name mappings in fix_workflows.py
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
ComfyUI-CogVideoXWrapper requires diffusers>=0.31.0 to function properly. Pin the version requirement to ensure the correct version is installed during setup.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add scikit-image for image processing
- Add piexif for EXIF metadata handling
- Add segment-anything for SAM model support
These dependencies are required for ComfyUI-Impact-Pack to load correctly.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Remove flux/musicgen standalone implementations in favor of ComfyUI:
- Delete models/flux/ and models/musicgen/ directories
- Remove redundant scripts (install.sh, download-models.sh, prepare-template.sh)
- Update README.md to reference Ansible playbook commands
- Update playbook.yml to remove flux/musicgen service definitions
- Add COMFYUI_MODELS.md with comprehensive model installation guide
- Update stop-all.sh to only manage orchestrator and vLLM services
All model downloads and dependency management now handled via
Ansible playbook tags (base, python, vllm, comfyui, comfyui-essential).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add ComfyUI installation to Ansible playbook with 'comfyui' tag
- Create ComfyUI requirements.txt and start.sh script
- Clone ComfyUI from official GitHub repository
- Symlink HuggingFace cache for Flux model access
- ComfyUI runs on port 8188 with CORS enabled
- Add ComfyUI to services list in playbook
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Set enforce_eager=True to disable CUDA graphs which were batching outputs
- Add disable_log_stats=True for better streaming performance
- This ensures AsyncLLMEngine yields tokens incrementally instead of returning complete response
- Track previous_text to calculate deltas instead of sending full accumulated text
- Fixes WebUI streaming issue where responses appeared empty
- Only send new tokens in each SSE chunk delta
- Resolves OpenAI API compatibility for streaming chat completions
