Update app

app.py

@@ -1,63 +1,67 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+from langchain.embeddings import SentenceTransformerEmbeddings
+from langchain.vectorstores import Chroma
+
+from transformers import T5Tokenizer, T5ForConditionalGeneration
+
+client = InferenceClient("google/flan-t5-large")
+
+embeddings = SentenceTransformerEmbeddings(model_name="msmarco-distilbert-base-v4")
+db = Chroma(persist_directory="embeddings", embedding_function=embeddings)
+
+tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
+model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large")
 
 
 def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
+        message,
+        history: list[tuple[str, str]],
+        max_tokens,
+        temperature,
+        repetition_penalty,
 ):
-    messages = [{"role": "system", "content": system_message}]
+    matching_docs = db.similarity_search(message)
+
+    context = ""
+    current_length = 0
+    for i, doc in enumerate(matching_docs):
+        doc_text = f"Document {i + 1}:\n{doc.page_content}\n\n"
+        doc_length = len(doc_text.split())
+        context += doc_text
+        current_length += doc_length
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    prompt = (
+        f"You are an expert in summarizing and answering questions based on given documents. "
+        f"Please provide a detailed and well-explained answer to the following query in 4-6 sentences:\n\n"
+        f"Query: {message}\n\n"
+        f"Based on the following documents:\n{context}\n\n"
+        f"Answer:"
+    )
 
-    messages.append({"role": "user", "content": message})
+    input_ids = tokenizer(prompt, return_tensors="pt").input_ids
 
-    response = ""
+    outputs = model.generate(input_ids,
+                             do_sample=True,
+                             max_new_tokens=max_tokens,
+                             temperature=temperature,
+                             repetition_penalty=repetition_penalty)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+    return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
-        response += token
-        yield response
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
 demo = gr.ChatInterface(
     respond,
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=1, maximum=2048, value=1024, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
+        gr.Slider(minimum=0.1, maximum=10, value=1.5, step=0.1, label="Repetition penalty"),
     ],
+    retry_btn=None,
+    undo_btn=None,
+    clear_btn=None,
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()

requirements.txt

@@ -3,4 +3,5 @@ langchain
 langchain-community
 chromadb
 sentence_transformers
-unstructured
+unstructured
+sentencepiece