404eb6ad0e
feat: add placeholder input image for CogVideoX I2V workflow
...
- Create 1024x1024 white placeholder with 'Input Frame' text
- Allows workflow validation without external image upload
- Will be replaced by API input in production use
2025-11-23 09:43:38 +01:00
47824ab987
fix: completely rebuild CogVideoX I2V workflow with correct configurations
...
Major fixes:
- Replace DualCLIPLoader with CLIPLoader using t5xxl_fp16.safetensors
- Fix CogVideoSampler parameter order: [num_frames, steps, cfg, seed, control, scheduler, denoise]
- Fix CogVideoImageEncode input: 'image' -> 'start_image'
- Remove CogVideoXVAELoader, use VAE directly from DownloadAndLoadCogVideoModel
- Add CogVideoTextEncode strength and force_offload parameters
- Simplify to 8 nodes (removed node 10)
- All nodes properly connected with correct link IDs
Version: 1.2.0
Tested against: ComfyUI-CogVideoXWrapper example workflows
2025-11-23 09:41:01 +01:00
5cd9237d82
fix: add h264-mp4 format parameters to VHS_VideoCombine
...
- Add required format-specific parameters: pix_fmt, crf, save_metadata, trim_to_audio
- Values: [8, 0, 'cogvideox_output', 'video/h264-mp4', 'yuv420p', 19, true, false]
- Fixes red node error in ComfyUI UI
2025-11-23 09:09:35 +01:00
6fab6386d7
feat: complete CogVideoX I2V workflow with proper node connections
...
- Add all necessary nodes: DualCLIPLoader, CogVideoImageEncode, CogVideoXVAELoader
- Add negative prompt support (node 8)
- Properly connect all nodes with links array (11 connections)
- Workflow now fully functional for image-to-video generation
Node flow:
1. LoadImage -> CogVideoImageEncode
2. DownloadAndLoadCogVideoModel -> CogVideoSampler (model)
3. DownloadAndLoadCogVideoModel -> CogVideoImageEncode (vae)
4. DualCLIPLoader -> CogVideoTextEncode (positive & negative)
5. CogVideoTextEncode (pos/neg) -> CogVideoSampler
6. CogVideoImageEncode -> CogVideoSampler (image conditioning)
7. CogVideoSampler -> CogVideoDecode
8. CogVideoXVAELoader -> CogVideoDecode
9. CogVideoDecode -> VHS_VideoCombine
Version: 1.1.0
2025-11-23 09:07:36 +01:00
a9c26861a4
fix: correct CogVideoX node types for I2V workflow
...
- Change CogVideoXSampler -> CogVideoSampler
- Change DiffusersLoader -> DownloadAndLoadCogVideoModel
- Change CLIPTextEncode -> CogVideoTextEncode
- Change VAEDecode -> CogVideoDecode
- Update model path to THUDM/CogVideoX-5b-I2V
- Fix sampler parameters: [seed, scheduler, num_frames, steps, cfg]
- Add CogVideoDecode tiling parameters
Note: Workflow still needs proper node connections (links array is empty)
2025-11-23 09:04:05 +01:00
897dcb175a
refactor: reorganize directory structure and remove hardcoded paths
...
Move comfyui and vllm out of models/ directory to top level for better
organization. Replace all hardcoded /workspace paths with relative paths
to make the configuration portable across different environments.
Changes:
- Move models/comfyui/ → comfyui/
- Move models/vllm/ → vllm/
- Remove models/ directory (empty)
- Update arty.yml: replace /workspace with environment variables
- Update supervisord.conf: use relative paths from /workspace/ai
- Update all script references to use new paths
- Maintain TQDM_DISABLE=1 to fix BrokenPipeError
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com >
2025-11-22 20:49:27 +01:00
