Update app.py

app.py

@@ -1,18 +1,10 @@
-import gradio as gr
+"""import gradio as gr
 import torch
 import time
 from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
 from threading import Thread
-import time
 import pytz
 from datetime import datetime
-import gradio as gr
-import torch
-import time
-import pytz
-from datetime import datetime
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
-from threading import Thread
 print("Loading model and tokenizer...")
 model_name = "large-traversaal/Phi-4-Hindi"
 tokenizer = AutoTokenizer.from_pretrained(model_name)
@@ -121,5 +113,59 @@ with gr.Blocks() as demo:
         inputs=None,
         outputs=[input_text, output_text]
     )
+if __name__ == "__main__":
+    demo.queue().launch()
+"""
+import gradio as gr
+import torch
+import time
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
+from threading import Thread
+import pytz
+from datetime import datetime
+print("Loading model and tokenizer...")
+model_name = "large-traversaal/Phi-4-Hindi"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
+print("Model and tokenizer loaded successfully!")
+option_mapping = {"translation": "### TRANSLATION ###", "mcq": "### MCQ ###", "nli": "### NLI ###", "summarization": "### SUMMARIZATION ###",
+    "long response": "### LONG RESPONSE ###", "direct response": "### DIRECT RESPONSE ###", "paraphrase": "### PARAPHRASE ###", "code": "### CODE ###"}
+def generate_response(message, temperature, max_new_tokens, top_p, task):
+    append_text = option_mapping.get(task, "")
+    prompt = f"INPUT : {message} {append_text} ### RESPONSE : "
+    print(f"Prompt: {prompt}")
+    start_time = time.time()
+    inputs = tokenizer.encode(prompt, return_tensors="pt").to(model.device)
+    outputs = model.generate(input_ids=inputs, max_new_tokens=max_new_tokens, use_cache=True, temperature=temperature, min_p=top_p, pad_token_id=tokenizer.eos_token_id)
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    processed_response = response.split("### RESPONSE :assistant")[-1].strip()
+    end_time = time.time()
+    time_taken = end_time - start_time
+    print(f"Output: {processed_response}")
+    print(f"Time taken: {time_taken:.2f} seconds")
+    pst_timezone = pytz.timezone('America/Los_Angeles')
+    current_time_pst = datetime.now(pst_timezone).strftime("%Y-%m-%d %H:%M:%S %Z%z")
+    print(f"Current timestamp (PST): {current_time_pst}")
+    return processed_response
+with gr.Blocks() as demo:
+    gr.Markdown("# Phi-4-Hindi Demo")
+    with gr.Row():
+        with gr.Column():
+            input_text = gr.Textbox(label="Input", placeholder="Enter your text here...", lines=5)
+            task_dropdown = gr.Dropdown(choices=["translation", "mcq", "nli", "summarization", "long response", "direct response", "paraphrase", "code"], value="long response", label="Task")
+            with gr.Row():
+                with gr.Column():
+                    temperature = gr.Slider(minimum=0.0, maximum=1.0, value=0.1, step=0.01, label="Temperature")
+                with gr.Column():
+                    max_new_tokens = gr.Slider(minimum=10, maximum=1000, value=10, step=10, label="Max New Tokens")
+                with gr.Column():
+                    top_p = gr.Slider(minimum=0.0, maximum=1.0, value=0.1, step=0.01, label="Top P")
+            with gr.Row():
+                clear_btn = gr.Button("Clear")
+                send_btn = gr.Button("Send", variant="primary")
+        with gr.Column():
+            output_text = gr.Textbox(label="Output", lines=15)
+    send_btn.click(fn=generate_response, inputs=[input_text, temperature, max_new_tokens, top_p, task_dropdown], outputs=output_text)
+    clear_btn.click(fn=lambda: ("", ""), inputs=None, outputs=[input_text, output_text])
 if __name__ == "__main__":
     demo.queue().launch()