app.py

import gradio as gr
import spaces
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
import torch
from threading import Thread
from typing import Iterator

model_name = "rubenroy/Zurich-14B-GCv2-5m"
MAX_INPUT_TOKEN_LENGTH = 4096

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)

@spaces.GPU
def generate(message: str, chat_history: list[tuple[str, str]], temperature=0.7, top_p=0.9, top_k=50, max_new_tokens=512, repetition_penalty=1.1) -> Iterator[str]:
    """Generates text responses using Zurich model with streaming."""
    
    conversation = []
    for user, assistant in chat_history:
        conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
    conversation.append({"role": "user", "content": message})
    
    input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt")
    
    if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH:
        input_ids = input_ids[:, -MAX_INPUT_TOKEN_LENGTH:]
        gr.Warning(f"Trimmed input from conversation as it was longer than {MAX_INPUT_TOKEN_LENGTH} tokens.")
    
    input_ids = input_ids.to(model.device)
    
    streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
    generate_kwargs = dict(
        {"input_ids": input_ids},
        streamer=streamer,
        max_new_tokens=max_new_tokens,
        do_sample=True if float(temperature) > 0 else False,
        top_p=top_p,
        top_k=top_k,
        temperature=temperature,
        repetition_penalty=repetition_penalty
    )
    
    t = Thread(target=model.generate, kwargs=generate_kwargs)
    t.start()
    
    outputs = []
    for text in streamer:
        outputs.append(text)
        yield "".join(outputs)

TITLE_HTML = """
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0/css/all.min.css">
<style>
    .model-btn {
        background: linear-gradient(135deg, #2563eb 0%, #1d4ed8 100%);
        color: white !important;
        padding: 0.75rem 1rem;
        border-radius: 0.5rem;
        text-decoration: none !important;
        font-weight: 500;
        transition: all 0.2s ease;
        font-size: 0.9rem;
        display: flex;
        align-items: center;
        justify-content: center;
        box-shadow: 0 2px 4px rgba(0,0,0,0.1);
    }
    .model-btn:hover {
        background: linear-gradient(135deg, #1d4ed8 0%, #1e40af 100%);
        box-shadow: 0 4px 6px rgba(0,0,0,0.2);
    }
    .model-section {
        flex: 1;
        max-width: 450px;
        background: rgba(255, 255, 255, 0.05);
        padding: 1.5rem;
        border-radius: 1rem;
        border: 1px solid rgba(255, 255, 255, 0.1);
        backdrop-filter: blur(10px);
        transition: all 0.3s ease;
    }
</style>

<div style="background: linear-gradient(135deg, #1e293b 0%, #0f172a 100%); padding: 1.5rem; border-radius: 1.5rem; text-align: center; margin: 1rem auto; max-width: 1200px; box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1);">
    <div style="margin-bottom: 1.5rem;">
        <h1 style="font-size: 2.5rem; font-weight: 800; margin: 0; background: linear-gradient(135deg, #60a5fa 0%, #93c5fd 100%); -webkit-background-clip: text; -webkit-text-fill-color: transparent;">Zurich</h1>
        <p style="font-size: 1.25rem; color: #94a3b8; margin: 0;">GammaCorpus v2-5m</p>
    </div>
</div>
"""

examples = [
    ["Explain quantum computing in simple terms"],
    ["Write a short story about a time traveler"],
    ["Explain the process of photosynthesis"],
    ["Tell me an interesting fact about Palm trees"]
]

with gr.Blocks() as demo:
    gr.HTML(TITLE_HTML)
    
    with gr.Accordion("Generation Settings", open=False):
        with gr.Row():
            with gr.Column():
                temperature = gr.Slider(0.0, 2.0, value=0.7, step=0.1, label="Temperature", info="Higher values make the output more random")
                top_p = gr.Slider(0.0, 1.0, value=0.9, step=0.05, label="Top P", info="Controls nucleus sampling")
                top_k = gr.Slider(1, 100, value=50, step=1, label="Top K", info="Limits vocabulary choices per step")
            with gr.Column():
                max_new_tokens = gr.Slider(1, 2048, value=512, step=1, label="Max New Tokens", info="Limits response length")
                repetition_penalty = gr.Slider(1.0, 2.0, value=1.1, step=0.1, label="Repetition Penalty", info="Discourages repeated phrases")

    chatbot = gr.ChatInterface(
        fn=generate,
        additional_inputs=[temperature, top_p, top_k, max_new_tokens, repetition_penalty],
        examples=examples
    )

demo.launch(share=True)