Update app.py

app.py

@@ -1,64 +1,64 @@
+import os
+import torch
 import gradio as gr
-from huggingface_hub import InferenceClient
+from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
+from threading import Thread
 
-"""
-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")
+MODEL_NAME = "S1mp1eXXX/Nimi-1b-thinking"
 
+# Load once at startup (important)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+model = AutoModelForCausalLM.from_pretrained(
+    MODEL_NAME,
+    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
+    device_map="auto"
+)
+
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+
+    messages = system_message + "\n"
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    for h in history:
+        messages += f"{h['role']}: {h['content']}\n"
 
-    messages.append({"role": "user", "content": message})
+    messages += f"user: {message}\nassistant:"
 
-    response = ""
+    inputs = tokenizer(messages, return_tensors="pt").to(model.device)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
+    streamer = TextIteratorStreamer(
+        tokenizer,
+        skip_prompt=True,
+        skip_special_tokens=True
+    )
+
+    generation_kwargs = dict(
+        **inputs,
+        max_new_tokens=max_tokens,
         temperature=temperature,
         top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+        streamer=streamer
+    )
+
+    thread = Thread(target=model.generate, kwargs=generation_kwargs)
+    thread.start()
 
-        response += token
-        yield response
+    partial_output = ""
+    for new_token in streamer:
+        partial_output += new_token
+        yield partial_output
 
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
+chatbot = gr.ChatInterface(
     respond,
+    type="messages",
     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.Textbox(value="You are a helpful assistant.", label="System message"),
+        gr.Slider(1, 2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(0.1, 4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(0.1, 1.0, value=0.95, step=0.05, label="Top-p"),
     ],
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    chatbot.launch()