Add application file

app.py

@@ -0,0 +1,47 @@
+import gradio as gr
+import torch
+import torch.nn.functional as F
+from transformers import RobertaTokenizerFast, RobertaForSequenceClassification
+from sklearn.preprocessing import LabelEncoder
+import numpy as np
+
+# Load tokenizer dan model
+# Load model directly
+tokenizer = RobertaTokenizerFast.from_pretrained("FadQ/results")
+model = RobertaForSequenceClassification.from_pretrained("FadQ/results")
+model.eval()
+
+# Label Ekman (urutan harus cocok dengan urutan training)
+ekman_labels = ['anger', 'disgust', 'fear', 'joy', 'sadness', 'surprise', 'neutral']
+label_encoder = LabelEncoder()
+label_encoder.fit(ekman_labels)
+
+# Fungsi prediksi rata-rata emosi per baris
+def predict_emotion_distribution(text):
+    lines = [line.strip() for line in text.split('\n') if line.strip()]
+    all_probs = []
+
+    for line in lines:
+        inputs = tokenizer(line, return_tensors="pt", truncation=True, padding=True, max_length=128)
+        with torch.no_grad():
+            logits = model(**inputs).logits
+            probs = F.softmax(logits, dim=-1).squeeze().cpu().numpy()
+            all_probs.append(probs)
+
+    if not all_probs:
+        return {label: 0.0 for label in ekman_labels}
+
+    avg_probs = np.mean(all_probs, axis=0)
+    result = {label: float(np.round(prob, 4)) for label, prob in zip(ekman_labels, avg_probs)}
+    return result
+
+# Gradio Interface
+interface = gr.Interface(
+    fn=predict_emotion_distribution,
+    inputs=gr.Textbox(lines=10, placeholder="Tulis diary-mu. Setiap baris = 1 kalimat...", label="Catatan Harian"),
+    outputs=gr.Label(num_top_classes=7, label="Distribusi Emosi"),
+    title="Prediksi Emosi dari Diary Harian",
+    description="Model mendeteksi emosi (7 label Ekman) dari teks harian. Input dipisah per baris. Output adalah rata-rata probabilitas per emosi."
+)
+
+interface.launch()