Create app.py

app.py

@@ -0,0 +1,51 @@
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import re
+
+# Modell laden
+model = AutoModelForSequenceClassification.from_pretrained("pepegiallo/flan-t5-base_ner")
+tokenizer = AutoTokenizer.from_pretrained("pepegiallo/flan-t5-base_ner")
+model.eval()
+
+id2label = {0: "LOC", 1: "ORG", 2: "PER", 3: "O"}
+
+# Hilfsfunktionen
+def custom_tokenize(text):
+    return re.findall(r"\w+|[^\w\s]", text, re.UNICODE)
+
+def custom_detokenize(tokens):
+    text = ""
+    for i, token in enumerate(tokens):
+        if i > 0 and re.match(r"\w", token):
+            text += " "
+        text += token
+    return text
+
+def classify_tokens(text):
+    tokens = custom_tokenize(text)
+    results = []
+
+    for i in range(len(tokens)):
+        wrapped = tokens[:i] + ["<TSTART>", tokens[i], "<TEND>"] + tokens[i+1:]
+        prompt = "classify token in: " + custom_detokenize(wrapped)
+
+        inputs = tokenizer(prompt, return_tensors="pt", padding="max_length", truncation=True, max_length=128)
+        with torch.no_grad():
+            logits = model(**inputs).logits
+            pred_id = torch.argmax(logits, dim=-1).item()
+            label = id2label[pred_id]
+
+        results.append((tokens[i], label))
+    return results
+
+# Gradio-UI definieren
+demo = gr.Interface(
+    fn=classify_tokens,
+    inputs=gr.Textbox(lines=3, placeholder="Enter a sentence..."),
+    outputs=gr.HighlightedText(label="Token Classification Output"),
+    title="Flan-T5 Token Classification (NER)",
+    description="Classifies each token in the input text as LOC, ORG, PER, or O."
+)
+
+demo.launch()