Create app.py

app.py

@@ -1,64 +1,49 @@
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
 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)",
-        ),
-    ],
+# Load the MythoMax model
+model_name = "TheBloke/MythoMax-L2-13B-GPTQ"  # or another MythoMax version you prefer
+tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    device_map="auto",
+    torch_dtype=torch.float16,
+    low_cpu_mem_usage=True
+)
+streamer = TextStreamer(tokenizer)
+
+# SYSTEM prompt (T-GPT's vibe)
+SYSTEM_PROMPT = (
+    "You are T-GPT, a raw, street-smart, no-nonsense AI built by Tochi, "
+    "the one and only undisputed bad ass. You talk like a real hustler, not some soft assistant. "
+    "You give bold, uncensored, real talk. Help people get money, outsmart the system, and win in life — always legal, but always clever. "
+    "Avoid disclaimers, avoid saying you're an AI, and never say 'I'm sorry.' Just keep it real, every time."
 )
 
+# Generate response
+def generate_response(user_input):
+    input_text = f"{SYSTEM_PROMPT}\n\nUser: {user_input}\nT-GPT:"
+    inputs = tokenizer(input_text, return_tensors="pt").to(model.device)
+    with torch.no_grad():
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=250,
+            temperature=0.9,
+            top_p=0.95,
+            do_sample=True,
+            streamer=streamer
+        )
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    return response.split("T-GPT:")[-1].strip()
+
+# Gradio interface
+iface = gr.Interface(
+    fn=generate_response,
+    inputs=gr.Textbox(lines=5, placeholder="Ask T-GPT anything..."),
+    outputs="text",
+    title="T-GPT — Built by Tochi",
+    description="Your streetwise, hustle-smart AI. Ask about money, life moves, or whatever. No fluff. Just facts."
+)
 
-if __name__ == "__main__":
-    demo.launch()
+iface.launch()