Update app.py

app.py

@@ -1,23 +1,136 @@
 import gradio as gr
-from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
+from datasets import load_dataset
+
+import os
 import spaces
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer, BitsAndBytesConfig
+import torch
+from threading import Thread
+from sentence_transformers import SentenceTransformer
+from datasets import load_dataset
+import time
+
+token = os.environ["HF_TOKEN"]
+ST = SentenceTransformer("mixedbread-ai/mxbai-embed-large-v1")
+
+dataset = load_dataset("not-lain/wikipedia", "Yoxas/statistical_literacy",revision = "embedded")
+
+data = dataset["train", "10kstats"]
+data = data.add_faiss_index("embeddings") # column name that has the embeddings of the dataset
+
+
+model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
+
+# use quantization to lower GPU usage
+bnb_config = BitsAndBytesConfig(
+    load_in_4bit=True, bnb_4bit_use_double_quant=True, bnb_4bit_quant_type="nf4", bnb_4bit_compute_dtype=torch.bfloat16
+)
+
+tokenizer = AutoTokenizer.from_pretrained(model_id,token=token)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype=torch.bfloat16,
+    device_map="auto",
+    quantization_config=bnb_config,
+    token=token
+)
+terminators = [
+    tokenizer.eos_token_id,
+    tokenizer.convert_tokens_to_ids("<|eot_id|>")
+]
+
+SYS_PROMPT = """You are an assistant for answering questions.
+You are given the extracted parts of a long document and a question. Provide a conversational answer.
+If you don't know the answer, just say "I do not know." Don't make up an answer."""
+
+
+
+def search(query: str, k: int = 3 ):
+    """a function that embeds a new query and returns the most probable results"""
+    embedded_query = ST.encode(query) # embed new query
+    scores, retrieved_examples = data.get_nearest_examples( # retrieve results
+        "embeddings", embedded_query, # compare our new embedded query with the dataset embeddings
+        k=k # get only top k results
+    )
+    return scores, retrieved_examples
+
+def format_prompt(prompt,retrieved_documents,k):
+    """using the retrieved documents we will prompt the model to generate our responses"""
+    PROMPT = f"Question:{prompt}\nContext:"
+    for idx in range(k) :
+        PROMPT+= f"{retrieved_documents['text'][idx]}\n"
+    return PROMPT
+
+
+@spaces.GPU(duration=150)
+def talk(prompt,history):
+    k = 1 # number of retrieved documents
+    scores , retrieved_documents = search(prompt, k)
+    formatted_prompt = format_prompt(prompt,retrieved_documents,k)
+    formatted_prompt = formatted_prompt[:2000] # to avoid GPU OOM
+    messages = [{"role":"system","content":SYS_PROMPT},{"role":"user","content":formatted_prompt}]
+    # tell the model to generate
+    input_ids = tokenizer.apply_chat_template(
+      messages,
+      add_generation_prompt=True,
+      return_tensors="pt"
+    ).to(model.device)
+    outputs = model.generate(
+      input_ids,
+      max_new_tokens=1024,
+      eos_token_id=terminators,
+      do_sample=True,
+      temperature=0.6,
+      top_p=0.9,
+    )
+    streamer = TextIteratorStreamer(
+            tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True
+        )
+    generate_kwargs = dict(
+        input_ids= input_ids,
+        streamer=streamer,
+        max_new_tokens=1024,
+        do_sample=True,
+        top_p=0.95,
+        temperature=0.75,
+        eos_token_id=terminators,
+    )
+    t = Thread(target=model.generate, kwargs=generate_kwargs)
+    t.start()
 
-# Load the tokenizer and model
+    outputs = []
+    for text in streamer:
+        outputs.append(text)
+        print(outputs)
+        yield "".join(outputs)
 
-tokenizer = AutoTokenizer.from_pretrained("Yoxas/autotrain-gpt2-statistical1")
-model = AutoModelForCausalLM.from_pretrained("Yoxas/autotrain-gpt2-statistical1")
 
-# Use a pipeline as a high-level helper
-pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
+TITLE = "# RAG"
 
-# Define the chatbot function
-@spaces.GPU(duration=120)
-def chatbot(input_text):
-    response = pipe(input_text, max_length=150, num_return_sequences=1)
-    return response[0]['generated_text']
+DESCRIPTION = """
+A rag pipeline with a chatbot feature
+Resources used to build this project :
+* embedding model : https://huggingface.co/mixedbread-ai/mxbai-embed-large-v1
+* dataset : https://huggingface.co/datasets/not-lain/wikipedia
+* faiss docs : https://huggingface.co/docs/datasets/v2.18.0/en/package_reference/main_classes#datasets.Dataset.add_faiss_index 
+* chatbot : https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct
+"""
 
-# Create the Gradio interface
-interface = gr.Interface(fn=chatbot, inputs="text", outputs="text", title="Research Paper Abstract Chatbot")
 
-# Launch the Gradio app
-interface.launch()
+demo = gr.ChatInterface(
+    fn=talk,
+    chatbot=gr.Chatbot(
+        show_label=True,
+        show_share_button=True,
+        show_copy_button=True,
+        likeable=True,
+        layout="bubble",
+        bubble_full_width=False,
+    ),
+    theme="Soft",
+    examples=[["what's anarchy ? "]],
+    title=TITLE,
+    description=DESCRIPTION,
+    
+)
+demo.launch(debug=True)