Hi Great start. I am using the knowledge and embedding its working great even with a tiny embedding model, I want to use the embedding as knowledge and then upload a temporary pdf in the chat window for comparison and discussion with the knowledge as reference. so basically a temporary storage and hence the obligatory upload button in the chat. Following on it would be great when multimodal LLM are fully implemented for the pdf viewer to also see the images in the pdf etc etc . I have achieved this with scripts so can be a tool (in other applications not here yet)
many thanks
I am SELF hosting
I’m very pleased that you’re satisfied with TaskingAI and have provided this feedback. We understand your needs, and in fact, we are currently developing PDF file reading functionalities in chat. However, this will require certain capabilities from the underlying model. This has already been implemented in Retrieval, and it can be used through Retrieval at the current stage.
1 Like