Update app.py

app.py

@@ -1,97 +1,55 @@
-from huggingface_hub import InferenceClient
 import gradio as gr
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
 
-client = InferenceClient(
-    "meta-llama/Meta-Llama-3-8B-Instruct"
-)
+model_name = "meta-llama/Meta-Llama-3-8B-Instruct"
+device_map = 'auto'
 
-def format_prompt(message, history):
-    prompt = "<s>"
-    for user_prompt, bot_response in history:
-        prompt += f"[INST] {user_prompt} [/INST]"
-        prompt += f" {bot_response}</s> "
-    prompt += f"[INST] {message} [/INST]"
-    return prompt
+def load_model() -> AutoModelForCausalLM:
+    return AutoModelForCausalLM.from_pretrained(model_name, device_map=device_map)
+
+def load_tokenizer() -> AutoTokenizer:
+    return AutoTokenizer.from_pretrained(model_name)
 
-def generate(
-    prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,
-):
-    temperature = float(temperature)
-    if temperature < 1e-2:
-        temperature = 1e-2
-    top_p = float(top_p)
+def preprocess_messages(message: str, history: list, system_prompt: str) -> dict:
+    messages = [{'role': 'system', 'content': system_prompt}, {'role': 'user', 'content': message}]
+    prompt = load_tokenizer().apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    return prompt
 
-    generate_kwargs = dict(
-        temperature=temperature,
+def generate_text(prompt: str, max_new_tokens: int, temperature: float) -> str:
+    model = load_model()
+    terminators = [load_tokenizer().eos_token_id, load_tokenizer().convert_tokens_to_ids(['\n'])]
+    temp = temperature + 0.1
+    outputs = model.generate(
+        prompt,
         max_new_tokens=max_new_tokens,
-        top_p=top_p,
-        repetition_penalty=repetition_penalty,
+        eos_token_id=terminators[0],
         do_sample=True,
-        seed=42,
-    )
-
-    formatted_prompt = format_prompt(prompt, history)
-
-    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
-    output = ""
-
-    for response in stream:
-        output += response.token.text
-        yield output
-    return output
-
-
-additional_inputs=[
-    gr.Slider(
-        label="Temperature",
-        value=0.9,
-        minimum=0.0,
-        maximum=1.0,
-        step=0.05,
-        interactive=True,
-        info="Higher values produce more diverse outputs",
-    ),
-    gr.Slider(
-        label="Max new tokens",
-        value=512,
-        minimum=0,
-        maximum=1048,
-        step=64,
-        interactive=True,
-        info="The maximum numbers of new tokens",
-    ),
-    gr.Slider(
-        label="Top-p (nucleus sampling)",
-        value=0.90,
-        minimum=0.0,
-        maximum=1,
-        step=0.05,
-        interactive=True,
-        info="Higher values sample more low-probability tokens",
-    ),
-    gr.Slider(
-        label="Repetition penalty",
-        value=1.2,
-        minimum=1.0,
-        maximum=2.0,
-        step=0.05,
-        interactive=True,
-        info="Penalize repeated tokens",
-    )
-]
-
-# Create a Chatbot object with the desired height
-chatbot = gr.Chatbot(height=450,
-                     layout="bubble")
-
-with gr.Blocks() as demo:
-    gr.HTML("<h1><center>🤖 Mistral-7B-Chat 💬<h1><center>")
-    gr.ChatInterface(
-        generate,
-        chatbot=chatbot,  # Use the created Chatbot object
-        additional_inputs=additional_inputs,
-        examples=[["Give me the code for Binary Search in C++"], ["Explain the chapter of The Grand Inquistor from The Brothers Karmazov"],["Why a penguin would make a great president? Give supporting arguments."]],
-
-    )
-
-demo.queue().launch(debug=True)
+        temperature=temp,
+        top_p=0.9
+     )
+    return load_tokenizer().decode(outputs[0], skip_special_tokens=True)
+
+def chat_function(
+    message: str,
+    history: list,
+    system_prompt: str,
+    max_new_tokens: int,
+    temperature: float
+) -> str:
+    prompt = preprocess_messages(message, history, system_prompt)
+    return generate_text(prompt, max_new_tokens, temperature)
+
+gr.ChatInterface(
+    chat_function,
+    chatbot=gr.Chatbot(height=400),
+    textbox=gr.Textbox(placeholder="Enter message here", container=False, scale=7),
+    title="LLAMA3 Chat",
+    description="""Chat with llama3""",
+    theme="soft",
+    additional_inputs=[
+        gr.Textbox("You shall answer to all the questions as very smart AI", label="System Prompt"),
+        gr.Slider(512, 4096, label="Max New Tokens"),
+        gr.Slider(0, 1, label="Temperature")
+     ]
+).launch()