Update app.py

app.py

@@ -1,96 +1,230 @@
+# import gradio as gr
+# from huggingface_hub import InferenceClient
+
+# # Initialize the client with your desired model
+# client = InferenceClient("Bhaskar2611/Capstone")
+
+# # Define the system prompt as an AI Dermatologist
+# def format_prompt(message, history):
+#     prompt = "<s>"
+#     # Start the conversation with a system message
+#     prompt += "[INST] You are an AI Dermatologist chatbot designed to assist users with skin by only providing text and if user information is not provided related to skin then ask what they want to know related to skin.[/INST]"
+#     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
+
+# # Function to generate responses with the AI Dermatologist context
+# 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)
+
+#     generate_kwargs = dict(
+#         temperature=temperature,
+#         max_new_tokens=max_new_tokens,
+#         top_p=top_p,
+#         repetition_penalty=repetition_penalty,
+#         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
+
+# # Customizable input controls for the chatbot interface
+# Settings = [
+#     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=256,
+#         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",
+#     )
+# ]
+# # Define the chatbot interface with the starting system message as AI Dermatologist
+# gr.ChatInterface(
+#     fn=generate,
+#     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, layout="panel"),
+#     additional_inputs = Settings,
+#     title="Skin Bot"
+# ).launch(show_api=False)
+
+# # Load your model after launching the interface
+# # gr.load("models/Bhaskar2611/Capstone").launch()
+
+# import gradio as gr
+# from huggingface_hub import InferenceClient
+
+# # Initialize the client with your Hugging Face token
+# client = InferenceClient(
+#     model="HuggingFaceH4/zephyr-7b-beta",
+#     hf_token = os.getenv("HF_TOKEN")
+# )
+
+# # Define the system prompt as an AI Dermatologist
+# def format_prompt(message, history):
+#     prompt = "<s>"
+#     prompt += "[INST] You are an AI Dermatologist chatbot designed to assist users with skin by only providing text and if user information is not provided related to skin then ask what they want to know related to skin.[/INST]"
+#     for user_prompt, bot_response in history:
+#         prompt += f"[INST] {user_prompt} [/INST] {bot_response}</s> "
+#     prompt += f"[INST] {message} [/INST]"
+#     return prompt
+
+# # Function to generate responses
+# 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)
+
+#     generate_kwargs = dict(
+#         temperature=temperature,
+#         max_new_tokens=max_new_tokens,
+#         top_p=top_p,
+#         repetition_penalty=repetition_penalty,
+#         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
+
+# # Sliders for customization
+# Settings = [
+#     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=256, minimum=0, maximum=1048, step=64, interactive=True, info="The maximum number 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")
+# ]
+
+# # Chat interface
+# gr.ChatInterface(
+#     fn=generate,
+#     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, layout="panel"),
+#     additional_inputs=Settings,
+#     title="Skin Bot"
+# ).launch(show_api=False)
+
+# # Load any additional models if needed
+# # gr.load("models/Bhaskar2611/Capstone").launch()
+
+import torch
+import transformers
+from transformers import AutoTokenizer, AutoModelForCausalLM, TextStreamer
 import gradio as gr
-from huggingface_hub import InferenceClient
 
-# Initialize the client with your desired model
-client = InferenceClient("Bhaskar2611/Capstone")
+# Load tokenizer and model from Hugging Face (your fine-tuned model)
+model_id = "Bhaskar2611/Capstone"
 
-# Define the system prompt as an AI Dermatologist
+tokenizer = AutoTokenizer.from_pretrained(model_id, use_auth_token=True)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype=torch.float16,
+    device_map="auto",
+    use_auth_token=True
+)
+
+streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
+
+# Function to format prompt with Zephyr-style and system prompt
 def format_prompt(message, history):
-    prompt = "<s>"
-    # Start the conversation with a system message
-    prompt += "[INST] You are an AI Dermatologist chatbot designed to assist users with skin by only providing text and if user information is not provided related to skin then ask what they want to know related to skin.[/INST]"
-    for user_prompt, bot_response in history:
-        prompt += f"[INST] {user_prompt} [/INST]"
-        prompt += f" {bot_response}</s> "
-    prompt += f"[INST] {message} [/INST]"
+    system_prompt = (
+        "You are a helpful medical assistant specialized in skin diseases. "
+        "Always provide accurate, responsible, and actionable information. "
+        "If needed, recommend seeing a dermatologist."
+    )
+    prompt = f"<|user|>\n{system_prompt}\n<|assistant|>\nSure, I'm here to help you.</s>\n"
+    for user_msg, bot_msg in history:
+        prompt += f"<|user|>\n{user_msg}\n<|assistant|>\n{bot_msg}</s>\n"
+    prompt += f"<|user|>\n{message}\n<|assistant|>\n"
     return prompt
 
-# Function to generate responses with the AI Dermatologist context
-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)
-
-    generate_kwargs = dict(
-        temperature=temperature,
-        max_new_tokens=max_new_tokens,
-        top_p=top_p,
-        repetition_penalty=repetition_penalty,
+# Function to generate model response
+def respond(message, history):
+    prompt = format_prompt(message, history)
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    output = model.generate(
+        **inputs,
+        max_new_tokens=2048,
+        pad_token_id=tokenizer.eos_token_id,
         do_sample=True,
-        seed=42,
+        temperature=0.7,
+        top_p=0.9,
+        streamer=streamer
     )
+    decoded_output = tokenizer.decode(output[0], skip_special_tokens=True)
+    # Extract the final assistant message only
+    final_response = decoded_output.split("<|assistant|>")[-1].strip().split("</s>")[0].strip()
+    return final_response
 
-    formatted_prompt = format_prompt(prompt, history)
+# Create Gradio interface
+chat = gr.ChatInterface(
+    fn=respond,
+    title="Skin & Hair Disease Assistant",
+    chatbot=gr.Chatbot(height=400),
+    textbox=gr.Textbox(placeholder="Describe your symptoms or ask a question...", container=False, scale=7),
+    description="Ask about skin conditions, hair issues, or treatment guidance. Powered by a custom fine-tuned Zephyr model.",
+    theme="soft",
+    examples=["I have red patches on my skin.", "What causes hair loss?", "Is dandruff a fungal infection?"],
+    cache_examples=False,
+    retry_btn="🔁 Retry",
+    undo_btn="↩️ Undo",
+    clear_btn="🧹 Clear"
+)
+
+# Launch app
+if __name__ == "__main__":
+    chat.launch()
 
-    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
-
-# Customizable input controls for the chatbot interface
-Settings = [
-    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=256,
-        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",
-    )
-]
-# Define the chatbot interface with the starting system message as AI Dermatologist
-gr.ChatInterface(
-    fn=generate,
-    chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, layout="panel"),
-    additional_inputs = Settings,
-    title="Skin Bot"
-).launch(show_api=False)
-
-# Load your model after launching the interface
-# gr.load("models/Bhaskar2611/Capstone").launch()