Merge pull request #2 from RobCaamano/milestone-2

README.md

@@ -1 +1,12 @@
-# Finetuning_Language_Models-Toxic_Tweets
+---
+title: Finetuning Language Models-Toxic Tweets
+emoji: 🌖
+colorFrom: red
+colorTo: indigo
+sdk: streamlit
+sdk_version: 1.17.0
+app_file: app.py
+pinned: false
+---
+
+# Finetuning_Language_Models-Toxic_Tweets

app.py

@@ -0,0 +1,27 @@
+from transformers import AutoTokenizer
+import streamlit as st
+from transformers import pipeline
+from transformers import (
+    TFAutoModelForSequenceClassification as AutoModelForSequenceClassification,
+)
+
+st.title("Detecting Toxic Tweets")
+
+demo = """I'm so proud of myself for accomplishing my goals today. #motivation #success"""
+
+text = st.text_area("Input text", demo, height=250)
+
+model_name = "distilbert-base-uncased-finetuned-sst-2-english"
+
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSequenceClassification.from_pretrained(model_name)
+clf = pipeline(
+    "toxicity-analysis", model=model, tokenizer=tokenizer, return_all_scores=True
+)
+
+input = tokenizer(text, return_tensors="tf")
+
+if st.button("Submit", type="primary"):
+    results = clf(text)[0]
+    classes = dict(d.values() for d in results)
+    st.bar_chart(classes)

requirements.txt

@@ -0,0 +1,5 @@
+streamlit
+opencv-python-headless
+numpy
+easyocr
+Pillow