Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
3 |
+
|
4 |
+
# LLaMA 3.3 8B Modell und Tokenizer laden
|
5 |
+
model_name = "meta-llama/Llama-3.3-8B"
|
6 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
7 |
+
model = AutoModelForCausalLM.from_pretrained(
|
8 |
+
model_name,
|
9 |
+
device_map=None, # Keine GPU-Zuweisung
|
10 |
+
torch_dtype="float32" # Float32 für CPU
|
11 |
+
)
|
12 |
+
|
13 |
+
# Funktion für die Textgenerierung
|
14 |
+
def generate_response(prompt):
|
15 |
+
inputs = tokenizer(prompt, return_tensors="pt", truncation=True)
|
16 |
+
outputs = model.generate(inputs["input_ids"], max_length=200, num_beams=5, early_stopping=True)
|
17 |
+
return tokenizer.decode(outputs[0], skip_special_tokens=True)
|
18 |
+
|
19 |
+
# Gradio-Interface erstellen
|
20 |
+
interface = gr.Interface(
|
21 |
+
fn=generate_response,
|
22 |
+
inputs="text",
|
23 |
+
outputs="text",
|
24 |
+
title="LLaMA 3.3 8B Text Generator (CPU)",
|
25 |
+
description="Gib einen Text ein, und LLaMA 3.3 8B generiert eine Antwort."
|
26 |
+
)
|
27 |
+
|
28 |
+
# App starten
|
29 |
+
if __name__ == "__main__":
|
30 |
+
interface.launch()
|