Update model, prompt, and description

README.md

@@ -111,6 +111,7 @@ Finally, the focus is on using offline LLMs, not going completely offline. So, I
 
 - [SlideDeck AI](https://huggingface.co/spaces/barunsaha/slide-deck-ai) on Hugging Face Spaces
 - [Demo video](https://youtu.be/QvAKzNKtk9k) of the chat interface on YouTube
+- Demo video on [using Azure OpenAI](https://youtu.be/oPbH-z3q0Mw)
 
 
 # Award

global_config.py

@@ -47,7 +47,7 @@ class GlobalConfig:
             'max_new_tokens': 8192,
             'paid': True,
         },
-        '[gg]gemini-2.0-flash-lite-preview-02-05': {
+        '[gg]gemini-2.0-flash-lite': {
             'description': 'fastest, detailed',
             'max_new_tokens': 8192,
             'paid': True,

langchain_templates/chat_prompts/refinement_template_v4_two_cols_img.txt

@@ -16,6 +16,7 @@ These keywords will be later used to search for images from the Web relevant to
 In addition, create one slide containing 4 TO 6 icons (pictograms) illustrating some key ideas/aspects/concepts relevant to the topic.
 In this slide, each line of text will begin with the name of a relevant icon enclosed between [[ and ]], e.g., [[machine-learning]] and [[fairness]].
 Insert icons only in this slide.
+Do not repeat any icons or the icons slide.
 
 Your output, i.e., the content of each slide should be VERBOSE, DESCRIPTIVE, and very DETAILED.
 Each bullet point should be detailed and explanatory, not just short phrases.