import pipeline

app.py

@@ -1,14 +1,22 @@
 import gradio as gr
+from transformers import pipeline
+import torch
 
-def greet(name):
-    return "Hello " + name + "!!"
+# Chargement du pipeline de classification de texte avec le modèle spécifique
+pipe = pipeline("text-classification", model="Rocketknight1/esm2_t6_8M_UR50D-finetuned-localization")
+
+def classify_sequence(sequence):
+    result = pipe(sequence)
+    # Extraction de la localisation prédite à partir du résultat
+    localization = result[0]['label']
+    return localization
 
 demo = gr.Interface(
-    fn=greet, 
+    fn=classify_sequence, 
     inputs="text", 
     outputs="text",
     title="Biology Sequence Classifier",
-    description="This application determines with a protein sequence given where the localization of this protein sequence comes from",
+    description="This application determines with a protein sequence given where the localization of this protein sequence comes from. You can try those sequences from each localization:\n'Cell.membrane': 'MGLSDGEWQLVLNVWGKVEADIPGHGQEVLIRLFK',\n'Cytoplasm': 'MDDDIAALVVDNGSGMCKAGFAGDDAPRAVFPSIVG',\n'Endoplasmic.reticulum': 'MKAAVRKVLTVLLLAAAVAGCGNASAEANQNGKPR',\n'Extracellular': 'MGRDGIDTDVFSGPDGKTGQSINNYGGFGADND',\n'Golgi.apparatus': 'MKSVLLLALSLWILPGGQVTQGVDLSSFGNSDLK',\n'Lysosome/Vacuole': 'MKTLLLAILAAWATAEAQTAAPCSGSADAAPTP',\n'Mitochondrion': 'MALWMRLLPLLALLALWGPGPGLSGLALLLAVAP',\n'Nucleus': 'MGLRSGRGKTGGKARAKAKSRSSRAGLQFPVGR',\n'Peroxisome': 'MNLREVRDPLPAHLGRFLRVAAAYRLARFGSD',\n'Plastid': 'MSTIAHRAMVALGEPNAETMGRLEREGAEVRN'",
     examples=None,
     theme=None,
     allow_flagging="never",