initial commit

requirements.txt

@@ -0,0 +1,3 @@
+liqfit==1.0.0
+transformers==4.37.2
+SentencePiece

streamlit_app.py

@@ -0,0 +1,54 @@
+from liqfit.pipeline import ZeroShotClassificationPipeline
+from liqfit.models import T5ForZeroShotClassification
+from transformers import T5Tokenizer
+import streamlit as st
+import time
+
+model = T5ForZeroShotClassification.from_pretrained('knowledgator/comprehend_it-multilingual-t5-base')
+tokenizer = T5Tokenizer.from_pretrained('knowledgator/comprehend_it-multilingual-t5-base')
+classifier = ZeroShotClassificationPipeline(model=model, tokenizer=tokenizer,ypothesis_template = '{}', encoder_decoder = True)
+st.title('Natural Language Project')
+st.markdown('Hafizh Zaki Prasetyo Adi|[email protected]|https://www.linkedin.com/in/hafizhzpa/')
+part=st.sidebar.radio("project",["sentimen", "emosi", "label khusus"],captions = ["menentukan label sentimen", "menentukan label emosi", "klasifikasi berdasarkan label yang ditentukan"])
+text = st.text_input('text', 'Saya sudah menggunakan produk ini selama sebulan dan saya sangat puas dengan hasilnya')
+multiclass = st.checkbox('Izinkan multi label')
+if part=='label khusus':
+    start=time.time()
+    label = st.text_input('label', 'positive,negative,neutral')
+    if st.button('run'):
+        candidate_labels = label.split(',')
+        result=classifier(text, candidate_labels, multi_label=multiclass)
+        if not multiclass:
+            st.text(f"label:{result['labels'][0]}")
+            st.text(f"skor:{result['scores'][0]}")
+        else:
+            bool_score=[score>0.5 for score in result['scores']]
+            st.text(f"label:{','.join([label for i,label in enumerate(result['labels']) if bool_score[i]])}")
+            st.text(f"skor:{','.join([skor for i,skor in enumerate(result['scores']) if bool_score[i]])}")
+        st.text(f"waktu:{time.time()-start}")
+if part=='sentimen':
+    start=time.time()
+    if st.button('run'):
+        candidate_labels = ["positive','negative','neutral"]
+        result=classifier(text, candidate_labels, multi_label=multiclass)
+        if not multiclass:
+            st.text(f"label:{result['labels'][0]}")
+            st.text(f"skor:{result['scores'][0]}")
+        else:
+            bool_score=[score>0.5 for score in result['scores']]
+            st.text(f"label:{','.join([label for i,label in enumerate(result['labels']) if bool_score[i]])}")
+            st.text(f"skor:{','.join([skor for i,skor in enumerate(result['scores']) if bool_score[i]])}")
+        st.text(f"waktu:{time.time()-start}")
+if part=='emotion':
+    start=time.time()
+    if st.button('run'):
+        candidate_labels = ["bahagia", "sedih", "takut", "marah", "antisipasi", "terkejut", "jijik","percaya"]
+        result=classifier(text, candidate_labels, multi_label=multiclass)
+        if not multiclass:
+            st.text(f"label:{result['labels'][0]}")
+            st.text(f"skor:{result['scores'][0]}")
+        else:
+            bool_score=[score>0.5 for score in result['scores']]
+            st.text(f"label:{','.join([label for i,label in enumerate(result['labels']) if bool_score[i]])}")
+            st.text(f"skor:{','.join([skor for i,skor in enumerate(result['scores']) if bool_score[i]])}")
+        st.text(f"waktu:{time.time()-start}")