Update app.py

app.py

@@ -1,69 +1,135 @@
+import os
+from threading import Thread
+
 import gradio as gr
-from huggingface_hub import InferenceClient
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
+
+
+MODEL_ID = os.getenv("MODEL_ID", "GenueAI/Matrix-Prime-8B")
+
+tokenizer = None
+model = None
+
+
+def load_model():
+    global tokenizer, model
+
+    if tokenizer is not None and model is not None:
+        return tokenizer, model
+
+    tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
+    if tokenizer.pad_token_id is None:
+        tokenizer.pad_token = tokenizer.eos_token
+
+    kwargs = {
+        "trust_remote_code": True,
+        "torch_dtype": torch.float16 if torch.cuda.is_available() else torch.float32,
+        "low_cpu_mem_usage": True,
+    }
+
+    if torch.cuda.is_available():
+        kwargs["device_map"] = "auto"
 
+    model = AutoModelForCausalLM.from_pretrained(MODEL_ID, **kwargs)
+    if not torch.cuda.is_available():
+        model = model.to("cpu")
 
-def respond(
-    message,
-    history: list[dict[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-    hf_token: gr.OAuthToken,
-):
-    """
-    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(token=hf_token.token, model="openai/gpt-oss-20b")
+    model.eval()
+    return tokenizer, model
 
-    messages = [{"role": "system", "content": system_message}]
 
-    messages.extend(history)
+def build_prompt(message, history, system_prompt):
+    messages = []
+    if system_prompt.strip():
+        messages.append({"role": "system", "content": system_prompt.strip()})
+
+    for user_message, assistant_message in history:
+        if user_message:
+            messages.append({"role": "user", "content": user_message})
+        if assistant_message:
+            messages.append({"role": "assistant", "content": assistant_message})
 
     messages.append({"role": "user", "content": message})
 
-    response = ""
+    tok, _ = load_model()
+    if hasattr(tok, "apply_chat_template") and tok.chat_template:
+        return tok.apply_chat_template(
+            messages,
+            tokenize=False,
+            add_generation_prompt=True,
+        )
+
+    prompt = ""
+    for item in messages:
+        role = item["role"].capitalize()
+        prompt += f"{role}: {item['content']}\n"
+    return prompt + "Assistant:"
+
+
+def chat(message, history, system_prompt, max_new_tokens, temperature, top_p, repetition_penalty):
+    if not message.strip():
+        yield ""
+        return
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        choices = message.choices
-        token = ""
-        if len(choices) and choices[0].delta.content:
-            token = choices[0].delta.content
+    tok, mdl = load_model()
+    prompt = build_prompt(message, history, system_prompt)
+    inputs = tok(prompt, return_tensors="pt").to(mdl.device)
 
+    streamer = TextIteratorStreamer(tok, skip_prompt=True, skip_special_tokens=True)
+    generation_kwargs = {
+        **inputs,
+        "streamer": streamer,
+        "max_new_tokens": int(max_new_tokens),
+        "temperature": float(temperature),
+        "top_p": float(top_p),
+        "repetition_penalty": float(repetition_penalty),
+        "do_sample": temperature > 0,
+        "pad_token_id": tok.pad_token_id,
+        "eos_token_id": tok.eos_token_id,
+    }
+
+    thread = Thread(target=mdl.generate, kwargs=generation_kwargs)
+    thread.start()
+
+    response = ""
+    for token in streamer:
         response += token
         yield response
 
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-chatbot = 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)",
-        ),
-    ],
-)
-
-with gr.Blocks() as demo:
-    with gr.Sidebar():
-        gr.LoginButton()
-    chatbot.render()
+with gr.Blocks(title="Matrix Prime 8B Chat") as demo:
+    gr.Markdown("# Matrix Prime 8B Chat")
+    gr.Markdown(f"Chat with `{MODEL_ID}` from Hugging Face.")
+
+    with gr.Row():
+        with gr.Column(scale=4):
+            chatbot = gr.ChatInterface(
+                fn=chat,
+                additional_inputs=[
+                    gr.Textbox(
+                        label="System prompt",
+                        value="You are a helpful assistant.",
+                        lines=3,
+                    ),
+                    gr.Slider(64, 4096, value=512, step=32, label="Max new tokens"),
+                    gr.Slider(0.0, 2.0, value=0.7, step=0.05, label="Temperature"),
+                    gr.Slider(0.05, 1.0, value=0.9, step=0.05, label="Top-p"),
+                    gr.Slider(1.0, 2.0, value=1.1, step=0.05, label="Repetition penalty"),
+                ],
+                textbox=gr.Textbox(
+                    placeholder="Ask Matrix Prime 8B anything...",
+                    container=False,
+                    scale=7,
+                ),
+                submit_btn="Send",
+                stop_btn="Stop",
+                retry_btn="Retry",
+                undo_btn="Undo",
+                clear_btn="Clear",
+            )
 
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.queue()
+    demo.launch()