Beginning from 1 November 2023, several changes have been implemented to enhance the experience of staff and students using PolyU GenAI application. Additional API connections have been established to Azure OpenAI services hosted at Data Centers in multiple regions to scale up concurrent chat conversations and performance. Students and staff can decide how to allocate available credits among models with fee to meet individual needs.
The credit required for GPT-4 is 20 to 30 times higher when compared to GPT-3.5 for input and completion respectively. You can exhaust the daily maximum Bing conversation first before using your credits on GPT-3 or GPT-4 chat. Alternately you can try a series of prompt input on GPT-3.5 before you use the refined prompt on GPT-4.
Always start a new conversation with the “sweep” icon 
if you switch topic. Clear previous chat conversation to conserve prompt input used tokens and allow the model to focus on your new ask and provide a more comprehensive response.
Always refer to examples in Hints and Tips on text to image service to plan for the right prompts to allow Stable Diffusion model to generate image meeting your expectations. If you would like to make minor modifications to the generated image, always input seed image number and include the required adjustments you would like to make in the prompt for the regeneration of next image.
As GenAI market landscape continues to evolve, we will monitor the latest development and deploy newer model released when they become available via Azure OpenAI API. Staff and students can leverage enhanced features or potentially benefit from lower usage fee of upcoming models.
Should you need further information or support in using PolyU GenAI application, please contact our IT HelpCentre (Tel: 2766 5900, WhatsApp/ WeChat: 6577 9669).
