GPT-3.5 MODEL
The application uses the ChatGPT Language Model OpenAI API to generate text. The language model has limited knowledge about the world and events after 2021.
Based on the prompt entered by the user, the model generates a text completion (text completion) that tries to match the context that was given to it by the user. By default, the application is set to use the language model immediately in both single question and chat mode. However, it is recommended that you read the information on fine-tuning the model using the four scale ranges provided by clicking the i icon on the main page.
In the drop-down list of settings, you can select a role for the neural network. While the default value is the most generic, for a more deterministic conversation, the appropriate values can be used.
It is recommended to use the maximum value of the tokens responsible for the length of the answer (scale-4) in order to get the most complete answer. Scale-2 is the most interesting, allowing you to adjust the "creativity" of the model, the higher the value, the less deterministic the answer will be. Increasing the values in Scale-1 asks the model to look for an answer from different domains.
It is very important to correctly compose the prompt. It is necessary to clearly describe what you want, give clear instructions. According to the Content Policy and Terms, the user has full ownership of the results they create with ChatGPT.
DALL·E MODEL
DALL·E is a AI system that can create realistic images and art from a description in natural language. The generated images will be 1024x1024 pixels. The number of image for one prompt is four.
To download an image, click on it in the modal window that opens. A new browser page will open with the selected image. Press the right button (tap with a delay on a mobile device) and select the option "Download as ..."
