Update app.py

app.py

@@ -1,63 +1,49 @@
-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
+import spaces
+import gradio as gr  
+import torch  
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+title = """# Minitron-8B-Base"""
+description = """
+Minitron is a family of small language models (SLMs) obtained by pruning [NVIDIA's](https://huggingface.co/nvidia) Nemotron-4 15B model. We prune model embedding size, attention heads, and MLP intermediate dimension, following which, we perform continued training with distillation to arrive at the final models.
 """
-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]})
+# Load the tokenizer and model
+model_path = "nvidia/Minitron-8B-Base"
+tokenizer = AutoTokenizer.from_pretrained(model_path)
 
-    messages.append({"role": "user", "content": message})
+device='cuda'
+dtype=torch.bfloat16
+model = AutoModelForCausalLM.from_pretrained(model_path, torch_dtype=dtype, device_map=device)
 
-    response = ""
+# Define the prompt format  
+def create_prompt(instruction):  
+    PROMPT = '''Below is an instruction that describes a task.\n\nWrite a response that appropriately completes the request.\n\n### Instruction:\n{instruction}\n\n### Response:'''  
+    return PROMPT.format(instruction=instruction)  
 
-    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
+@spaces.GPU  
+def respond(message, history, system_message, max_tokens, temperature, top_p):  
+    prompt = create_prompt(message)  
+      
+    input_ids = tokenizer.encode(prompt, return_tensors="pt").to(model.device)
 
-        response += token
-        yield response
+    output_ids = model.generate(input_ids, max_length=50, num_return_sequences=1)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+    output_text = tokenizer.decode(output_ids[0], skip_special_tokens=True)
+      
+    return output_text  
+  
 demo = gr.ChatInterface(
-    respond,
+    title=gr.Markdown(title),
+    # description=gr.Markdown(description),
+    fn=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()
+        gr.Textbox(value="You are Minitron an AI assistant created by Tonic-AI", 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()