debug prompt

app.py

@@ -106,14 +106,14 @@ rag_prompt = PromptTemplate.from_template(RAG_PROMPT_TEMPLATE)
 hf_llm = HuggingFaceEndpoint(
     endpoint_url=HF_LLM_ENDPOINT,
     max_new_tokens=512,
-    top_k=10,
-    top_p=0.95,
-    typical_p=0.95,
-    temperature=0.01,
-    repetition_penalty=1.03,
+    top_k=50,  # Increase to allow more diverse sampling
+    top_p=0.9,  # Slightly decrease to balance diversity and coherence
+    temperature=0.8,  # Increase to add creativity and friendliness
+    repetition_penalty=1.01,  # Slightly lower to reduce repetition
     huggingfacehub_api_token=HF_TOKEN,
 )
 
+
 @cl.author_rename
 def rename(original_author: str):
     """
@@ -137,10 +137,7 @@ async def start_chat():
     """
 
     ### BUILD LCEL RAG CHAIN THAT ONLY RETURNS TEXT
-    lcel_rag_chain = (
-        {"context": itemgetter("query") | hf_retriever, "query": itemgetter("query")}
-        | rag_prompt | hf_llm
-        )
+    lcel_rag_chain = {"context": itemgetter("query") | hf_retriever, "query": itemgetter("query")}| rag_prompt | hf_llm
 
     cl.user_session.set("lcel_rag_chain", lcel_rag_chain)
 

chainlit.md

@@ -1,36 +1,31 @@
-### SF Sentinel: The Cutting-Edge AI Experience
+### 🚀 SF Sentinel: The Cutting-Edge AI Experience 🌉
 
-Welcome to **SF Sentinel**, your gateway to the future of intelligent information retrieval and augmented generation, inspired by the innovative spirit of San Francisco. Here’s why SF Sentinel is not just an app but a technological marvel:
+Welcome to **SF Sentinel**, your gateway to intelligent info retrieval inspired by San Francisco. Here's why SF Sentinel is a technological marvel:
 
 ---
 
-#### **Powered by State-of-the-Art Models**
+#### 🌟 **Powered by State-of-the-Art Models**
 
-1. **LLaMA 3: The Next Generation Language Model**
-   - **NousResearch/Meta-Llama-3-8B-Instruct**: At the heart of SF Sentinel is the LLaMA 3, a powerful language model designed to understand and generate human-like text. With 8 billion parameters, this model brings unparalleled accuracy and fluency to natural language processing, ensuring that every response is as insightful as a conversation with a San Francisco sage.
+1. **LLaMA 3: Next Gen Language Model**
+   - **NousResearch/Meta-Llama-3-8B-Instruct**: 8 billion parameters for unparalleled accuracy and fluency in natural language processing.
 
-2. **Arctic Embed: Precision Embeddings for Context-Aware Insights**
-   - **Snowflake/snowflake-arctic-embed-m**: Our embedding model, Arctic Embed, excels at capturing the essence of complex texts. By transforming textual data into high-dimensional vectors, it allows SF Sentinel to understand the nuanced relationships between different pieces of information, delivering precise and context-aware insights every time.
+2. **Arctic Embed: Precision Embeddings**
+   - **Snowflake/snowflake-arctic-embed-m**: Captures the essence of complex texts for context-aware insights.
 
 ---
 
-#### **Leveraging Hugging Face Inference Endpoints**
+#### ⚡ **Leveraging Hugging Face Inference Endpoints**
 
-- **Hugging Face Inference Endpoints**: The backbone of SF Sentinel's real-time processing capabilities, these endpoints enable seamless integration and deployment of cutting-edge models. By utilizing Hugging Face's robust infrastructure, we ensure that SF Sentinel can handle intensive computations with speed and reliability, providing instant responses to your queries.
+- **Real-Time Processing**: Instant responses with Hugging Face's robust infrastructure for seamless model integration.
 
 ---
 
-#### **Frameworks That Empower**
+#### 🔧 **Frameworks That Empower**
 
-1. **LangChain: The Ultimate Chain of Intelligence**
-   - **LangChain**: This powerful framework orchestrates the seamless interaction between different AI components. LangChain enables SF Sentinel to combine the strengths of LLaMA 3 and Arctic Embed, ensuring that data flows smoothly and insights are generated efficiently.
-
-2. **FAISS: High-Speed Similarity Search**
-   - **Facebook AI Similarity Search (FAISS)**: A critical component for managing and querying large-scale vector data, FAISS ensures that SF Sentinel can perform rapid and accurate similarity searches. This means you get the most relevant information faster than ever before.
-
-3. **Chainlit: Interactive AI Conversations**
-   - **Chainlit**: Our conversational framework, Chainlit, transforms SF Sentinel into an interactive assistant. With Chainlit, you can engage in dynamic, back-and-forth conversations, making the experience not just informative but also engaging and intuitive.
+1. **LangChain**: Orchestrates AI components for efficient data flow and insight generation.
+2. **FAISS**: High-speed similarity search for rapid and accurate info retrieval.
+3. **Chainlit**: Interactive AI conversations for engaging and intuitive user experiences.
 
 ---
 
-Embrace the future. Experience **SF Sentinel**.
+Experience the future. Discover **SF Sentinel** today! 🌉✨