Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
KewkLW authored Jun 21, 2024
1 parent 5757b61 commit 50290d8
Showing 1 changed file with 41 additions and 15 deletions.
56 changes: 41 additions & 15 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,21 +3,47 @@ Dumb shit to make life a lil more comfy

## Tools

### FormattedPromptNode
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/7bf50966-6e6b-43cd-92ba-42aae291be8f)

### Example pasting unformatted text
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/18a8cee5-3b13-47c5-87b1-6171fe7cd093)

### Example scheduling of Interrogator output
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/27b54da6-6b96-4adc-a601-508e4d3179c5)


### TensorDebugPlus
Edit of https://github.com/cubiq/ComfyUI_essentials
All I did was add a string output so I don't have to rely on the console.
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/36012f09-d46a-4ec7-9373-5ce596657606)

1. text_append_node.py
Appends text to existing text, line by line.
- Takes two inputs: current_text and append_text
- Appends each line of append_text to each line of current_text
- Useful for generating combinations of text or expanding existing text in a structured way
- Can be used in ComfyUI workflows for text preprocessing or generation tasks
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/c9774759-f1d5-4f38-bc44-4639d7c6762b)
2. vramdebugplus.py
Monitors and manages VRAM usage.
- Optionally clears CUDA cache and performs garbage collection
- Can unload all models to free up VRAM
- Provides detailed or simple reports on VRAM usage
- Shows free VRAM before and after operations
- Offers insights into system RAM usage
- Useful for optimizing VRAM usage in complex ComfyUI workflows

3. tensordebugplus.py
This script provides detailed information about tensors and neural network models.
- Analyzes the structure of tensors, including nested tensors and PyTorch modules
- Reports shape, data type, device, and memory usage of tensors
- Optionally includes gradient information and statistical measures (min, max, mean, std)
- Provides a summary of total parameters and memory usage
- Useful for debugging and optimizing machine learning models in ComfyUI
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/36012f09-d46a-4ec7-9373-5ce596657606)

4. animation_schedule_output.py
Designed to format text input for animation keyframes.
- Takes unformatted prompts and formats them with keyframe numbers
- Allows setting a custom keyframe interval
- Supports an optional offset for keyframe numbering
- Outputs formatted text suitable for animation schedules or keyframe-based systems
- Useful for preparing text inputs for animated or time-based ComfyUI workflows
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/7bf50966-6e6b-43cd-92ba-42aae291be8f)

### Example pasting unformatted text
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/18a8cee5-3b13-47c5-87b1-6171fe7cd093)

### Example scheduling of Interrogator output
![image](https://github.com/KewkLW/ComfyUI-kewky_tools/assets/57611539/27b54da6-6b96-4adc-a601-508e4d3179c5)

These custom nodes extend the functionality of ComfyUI, providing tools for text manipulation, VRAM management, tensor analysis, and animation scheduling. They can be particularly useful for users working with complex workflows, large models, or animation projects in ComfyUI.

## Getting Started

Expand Down

0 comments on commit 50290d8

Please sign in to comment.