Clone the repository and get into project folder with git clone https://github.com/alpertunga-bile/prompt-tools-webui.git && cd prompt-tools-webui command.
Start the application with python start.py command.
Tabs
Parser Tab
Choose prompt files to parse.
Choose if you want to translate your prompts to English.
Click Parse button to parse.
If your files are under the prompts folder, you can parse all of them by just clicking Parse All Files In Prompts Folder button.
Wait for Done !!! text to appear in Info label. Files are going to save in prompts folder.
Civitai Tab
Files can be already exists in dataset folder. In this condition, new datasets are added to files. Otherwise files are created under dataset folder.
Choose filename for saved datasets. Datasets are going to be saved in dataset folder.
Choose image per page.
Choose page number to start and end.
Write wanted and unwanted prompts seperated with comma(","). The example is given.
Choose sort, period and NSFW options with dropdowns.
Click Enhance button. Wait for the progress bar in Info label to finish.
You can check the datasets with Show Dataset Folder button. It opens the dataset folder in the project.
Train Tab
Write model name of the text generator. You can choose the models from this site. You have to right full name e.g. bigscience/bloom-560m.
Write a model folder name to be saved in dataset folder.
Choose epochs and batch size variables.
Choose the dataset you want to use for training.
Click Train button. Check the progress from the terminal. When finished in Train Info label will show Done !!! text and in Evaluate Info label will show loss value of the trained model.
Generate Tab
Give the full model name and model folder name. You can check the model folder name with Show Dataset Folder button. It will open the dataset folder so you can get the model folder name easily.
Select minimum and maximum lengths for the generated tokens for the text.
Choose do sample and early stop variables. Do sample picks words based on their conditional probability. If early stop is selected, generation finishes if the EOS token is reached.
Choose recursive level and check if you want the self recursive feature.
Enter your seed and click Generate button.
Wait until the text shows up in the Generated Text textbox.
How Recursive Works?
Let's say we give a, as seed and recursive level is 1. I am going to use the same outputs for this example to understand the functionality more accurately.
With self recursive, let's say generator's output is b. So next seed is going to be b and generator's output is c. Final output is a, c. It can be used for generating random outputs.
Without self recursive, let's say generator's output is b. So next seed is going to be a, b and generator's output is a, b, c. Final output is a, b, c. It can be used for more accurate prompts.
Upscale Tab
Put your images under upscaleInput folder in the project. The output is going to be under upscaleOutput folder.
Select Real-ESRGAN model to upscale.
Choose scale factor and face enhancement feature.
Choose FP32 feature if the outputs are black.
Click Upscale button. Wait until the outputs are appeared in the Outputs label.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent