Update app.py

app.py

@@ -1,64 +1,109 @@
 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")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    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
-
-        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=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)",
-        ),
-    ],
-)
-
-
-if __name__ == "__main__":
-    demo.launch()
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
+
+# --- Model Loading (Do this only once, outside the function) ---
+
+# Option 1: Pipeline (High-Level, Easier)
+use_pipeline = True  # Set to False to use the manual method
+
+if use_pipeline:
+    pipe = pipeline("text-generation", model="kakaocorp/kanana-nano-2.1b-base")
+else:
+    # Option 2: Manual Tokenizer and Model (More Control)
+    tokenizer = AutoTokenizer.from_pretrained("kakaocorp/kanana-nano-2.1b-base")
+    model = AutoModelForCausalLM.from_pretrained("kakaocorp/kanana-nano-2.1b-base")
+    # Move model to GPU if available
+    if model.device.type != 'cuda' and torch.cuda.is_available():
+        model = model.to("cuda")
+        print("Model moved to CUDA")
+
+# --- Generation Function ---
+
+def generate_text(prompt, max_length=50, temperature=1.0, top_k=50, top_p=1.0, no_repeat_ngram_size=0, num_return_sequences=1):
+    """Generates text based on the given prompt and parameters."""
+
+    if use_pipeline:
+        messages = [{"role": "user", "content": prompt}]  # Format for pipeline
+        try:
+            result = pipe(
+                messages,
+                max_length=max_length,
+                temperature=temperature,
+                top_k=top_k,
+                top_p=top_p,
+                no_repeat_ngram_size=no_repeat_ngram_size,
+                num_return_sequences=num_return_sequences,
+                return_full_text=False, # Important:  We only want generated text
+                pad_token_id=pipe.tokenizer.eos_token_id  # Prevent warning, pipeline knows the EOS token
+            )
+            # Pipeline returns a list of dictionaries, each with 'generated_text'
+            return "\n\n".join([res['generated_text'] for res in result])
+
+        except Exception as e:
+            return f"Error during generation: {e}"
+
+    else:  # Manual method
+        try:
+            inputs = tokenizer(prompt, return_tensors="pt")
+              # Move input tensors to the same device as the model
+            inputs = {k: v.to(model.device) for k, v in inputs.items()}
+            
+            outputs = model.generate(
+                **inputs,
+                max_length=max_length,
+                temperature=temperature,
+                top_k=top_k,
+                top_p=top_p,
+                no_repeat_ngram_size=no_repeat_ngram_size,
+                num_return_sequences=num_return_sequences,
+                pad_token_id=tokenizer.eos_token_id,  # Ensure padding is correct
+                do_sample=True # Ensure sampling happens.
+            )
+            
+            generated_texts = []
+            for i in range(outputs.shape[0]):
+                 generated_text = tokenizer.decode(outputs[i], skip_special_tokens=True)
+                 generated_texts.append(generated_text)
+           
+            return "\n\n".join(generated_texts)
+        except Exception as e:
+            return f"Error during generation: {e}"
+
+
+
+# --- Gradio Interface ---
+
+with gr.Blocks() as demo:
+    gr.Markdown("# Text Generation with kakaocorp/kanana-nano-2.1b-base")
+
+    with gr.Row():
+        with gr.Column():
+            prompt_input = gr.Textbox(label="Prompt", placeholder="Enter your prompt here...")
+            with gr.Accordion("Generation Parameters", open=False):
+                max_length_slider = gr.Slider(label="Max Length", minimum=10, maximum=512, value=50, step=1)
+                temperature_slider = gr.Slider(label="Temperature", minimum=0.1, maximum=2.0, value=1.0, step=0.1)
+                top_k_slider = gr.Slider(label="Top K", minimum=0, maximum=100, value=50, step=1)
+                top_p_slider = gr.Slider(label="Top P", minimum=0.0, maximum=1.0, value=1.0, step=0.05)
+                no_repeat_ngram_size_slider = gr.Slider(label="No Repeat N-gram Size", minimum=0, maximum=10, value=0, step=1)  # Add the slider
+                num_return_sequences_slider = gr.Slider(label="Number of Return Sequences", minimum=1, maximum=5, value=1, step=1)
+
+            generate_button = gr.Button("Generate")
+
+        with gr.Column():
+            output_text = gr.Textbox(label="Generated Text", readonly=True)
+
+    generate_button.click(
+        generate_text,
+        inputs=[
+            prompt_input,
+            max_length_slider,
+            temperature_slider,
+            top_k_slider,
+            top_p_slider,
+            no_repeat_ngram_size_slider,
+            num_return_sequences_slider
+        ],
+        outputs=output_text,
+    )
+
+demo.launch(share=True)