change

app.py

@@ -1,64 +1,68 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from unsloth import FastLanguageModel
+from transformers import TextStreamer
 
-"""
-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("michailroussos/model_llama_8d")
+# Load the model and tokenizer locally
+max_seq_length = 2048
+dtype = None
+model_name_or_path = "michailroussos/model_llama_8d"
 
+# Load model and tokenizer using unsloth
+model, tokenizer = FastLanguageModel.from_pretrained(
+    model_name=model_name_or_path,
+    max_seq_length=max_seq_length,
+    dtype=dtype,
+    load_in_4bit=True,
+)
+FastLanguageModel.for_inference(model)  # Enable optimized inference
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
+# Define the response function
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    # Build the chat message history
     messages = [{"role": "system", "content": system_message}]
-
     for val in history:
-        if val[0]:
+        if val[0]:  # User message
             messages.append({"role": "user", "content": val[0]})
-        if val[1]:
+        if val[1]:  # Assistant message
             messages.append({"role": "assistant", "content": val[1]})
-
     messages.append({"role": "user", "content": message})
+    
+    # Tokenize the input messages
+    inputs = tokenizer.apply_chat_template(
+        messages,
+        tokenize=True,
+        add_generation_prompt=True,  # Required for generation
+        return_tensors="pt",
+    ).to("cuda")
 
-    response = ""
+    # Initialize a TextStreamer for streaming output
+    text_streamer = TextStreamer(tokenizer, skip_prompt=True)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
+    # Generate the model's response
+    response = ""
+    for output in model.generate(
+        input_ids=inputs,
+        streamer=text_streamer,
+        max_new_tokens=max_tokens,
+        use_cache=True,
         temperature=temperature,
         top_p=top_p,
     ):
-        token = message.choices[0].delta.content
-
+        token = tokenizer.decode(output, skip_special_tokens=True)
         response += token
         yield response
 
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# Define the Gradio interface
 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)",
-        ),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
     ],
 )
 
-
 if __name__ == "__main__":
     demo.launch()